OS/2 may not have the fastest Java on the PC architecture anymore
josco
> Symantec has announced a new Java Just In Time compiler called JITWin,
> which is designed as a plugin for Communicator 4.05.
>
> http://www.symantec.com/domain/cafe/jitwin/index.html
>
> The following page shows CaffeineMark 3.0 tests of this new JIT compared to
> the stock Netscape Java and compared to Microsoft's latest JIT:
>
> http://www.symantec.com/domain/cafe/jitwin/caffeine.html
>
> Now if I remember correctly, when the claims were coming down the pike that
> OS/2's Java was fastest, it was the case that it just edged out the
> Microsoft JIT by a few percent on comparable hardware. (Do I remember
> wrong?)
>
> Symantec's JIT rates 38% faster than the Microsoft JIT on CaffeineMark 3.0.
> If that's realistic, that would put it well in the lead of OS/2's JIT, too.
Maybe. Maybe not. Look what was sent into Javaworld. I'm confident the
merger of the OS/2 and Java groups at IBM lets them build the best JVM and
JIT for x86 within OS/2.
http://www.javaworld.com/javaworld/jw-09-1998/jw-09-letters.html
IBM JDK 1.1.7 on the horizon
[...]
We have made significant progress even since the 1.1.6 release. Expect
the next crank of Java, 1.1.7, to be greatly improved on OS/2 both in
throughput (more tuning and a better JIT) as well as more concurrent
connections. Some of these changes may even make it into the 1.1.6 service
stream.
Rajiv Arora
Senior Software Engineer
NCSD System Performance
http://www.software.ibm.com/os/warp/warpfm/sept98/#story4
From OS/2 WARP FM
"Some of the best programmers and software engineers working today write
for OS/2 -- either in IBM development labs, in companies that produce
commercial OS/2 applications or in enterprises that use OS/2 to support
mission critical operations. This group is now turning its attention
toward Java in a variety of ways. "
**********
"Sun Microsystems incorporated the Java Virtual Machine (JVM) performance
enhancements developed by the IBM OS/2 team into subsequent releases of
the JVM. "
**********
Many OS/2 ISVs -- Kyrus, Athena Design, InnoVal and Nemasoft -- are either
adding new Java products to their offerings or creating Java versions of
existing products. "
josco
> On Wed, 2 Sep 1998 20:44:27 -0700, josco <jo...@sea.monterey.edu> added
> this to the Deja-News archive:
> >Maybe. Maybe not. Look what was sent into Javaworld. I'm confident the
> >merger of the OS/2 and Java groups at IBM lets them build the best JVM and
> >JIT for x86 within OS/2.
> >
> >http://www.javaworld.com/javaworld/jw-09-1998/jw-09-letters.html
> >
> >IBM JDK 1.1.7 on the horizon
> >
> >[...]
> >
> >We have made significant progress even since the 1.1.6 release. Expect
> >the next crank of Java, 1.1.7, to be greatly improved on OS/2 both in
> >throughput (more tuning and a better JIT) as well as more concurrent
> >connections. Some of these changes may even make it into the 1.1.6 service
> >stream.
> >
> >Rajiv Arora
> >Senior Software Engineer
> >NCSD System Performance
> >
>
>
> The only comment I would make about this is that whenever someone says "The
> next version from Microsoft will be better" you respond "Vapor! Vapor! Talk
> to me about what's on the market now, not what's coming in the future!"
>
> Why, then, is it OK for you to talk about the next version and how good
> it's going to be for OS/2, before the next version has shipped?
You sound like a clucking chicken.
I posted the comments of someone else who is working on the JDK and says
it will be faster. I made no such claims. I never said 1.1.7 would be
faster -- I said "maybe. maybe not"
> Shouldn't we be consistent? Either *everyone* gets to talk about the next
> version, or *no-one* does.
I am consistent. You're overly excited and are reading into things --
calm down.
> JDK 1.1.7 is on the horizon, but it isn't here yet. Until it's in the real
> world and can be independently tested, then all we have are unsubstantiated
> claims. (And vague ones at that. I'll indulge in a Tholenation and ask: how
> much is "greatly improved"?)
The hell if I know how fast 1.1.7 is compared to 1.1.6. It's a work in
progress. I simlpy posted a letter to the editor of JavaWorld -- it's a
very interesting letter. I'm still confident OS/2's JVM will be the best
JVM for x86 systems but I'm sure IBM will have to fight with constant
releases. Hell, Sun is even adopting some of IBM's OS/2 JVM technology
because it's so damn good.
> In the mean time, the Symantec JIT is shipping now. (Cheap, too: $40)
1) You'll not buy it. More empty talk from a person who encourages ISVs
to do development but does not buy the product.
2) You have no clue if the JIT is faster than 1.1.6. None. Calm down.
3) Free is cheaper. Let's see if it's faster than the OS/2 JIT in 1.1.6
and by how much. Is it worth $40?
4) Another nail in the "Windows Everywhere" coffin.
Dennis Peterson
>
> On Wed, 2 Sep 1998 20:44:27 -0700, josco <jo...@sea.monterey.edu> added
> this to the Deja-News archive:
>
> >On Thu, 3 Sep 1998, Steven C. Den Beste wrote:
> >
> >> which is designed as a plugin for Communicator 4.05.
> >>
> >> http://www.symantec.com/domain/cafe/jitwin/index.html
> >>
> >> The following page shows CaffeineMark 3.0 tests of this new JIT compared to
> >> the stock Netscape Java and compared to Microsoft's latest JIT:
> >>
> >> http://www.symantec.com/domain/cafe/jitwin/caffeine.html
> >>
> >> Now if I remember correctly, when the claims were coming down the pike that
> >> OS/2's Java was fastest, it was the case that it just edged out the
> >> Microsoft JIT by a few percent on comparable hardware. (Do I remember
> >> wrong?)
> >>
> >> Symantec's JIT rates 38% faster than the Microsoft JIT on CaffeineMark 3.0.
> >> If that's realistic, that would put it well in the lead of OS/2's JIT, too.
> >
> >
> >merger of the OS/2 and Java groups at IBM lets them build the best JVM and
> >JIT for x86 within OS/2.
> >
> >http://www.javaworld.com/javaworld/jw-09-1998/jw-09-letters.html
> >
> >IBM JDK 1.1.7 on the horizon
> >
> >[...]
> >
> >We have made significant progress even since the 1.1.6 release. Expect
> >the next crank of Java, 1.1.7, to be greatly improved on OS/2 both in
> >throughput (more tuning and a better JIT) as well as more concurrent
> >connections. Some of these changes may even make it into the 1.1.6 service
> >stream.
> >
> >Rajiv Arora
> >Senior Software Engineer
> >NCSD System Performance
> >
>
> The only comment I would make about this is that whenever someone says "The
> next version from Microsoft will be better" you respond "Vapor! Vapor! Talk
> to me about what's on the market now, not what's coming in the future!"
>
> Why, then, is it OK for you to talk about the next version and how good
> it's going to be for OS/2, before the next version has shipped?
>
> version, or *no-one* does.
>
> world and can be independently tested, then all we have are unsubstantiated
> claims. (And vague ones at that. I'll indulge in a Tholenation and ask: how
> much is "greatly improved"?)
>
Mike Cowlishaw (Rexx, NetRexx, Java) has released a Java 1.2 compatible
version of NetRexx. I've noticed that NetRexx releases presage OS/2 JVM
releases by a short period of time. I know this is of paramount interest
to you so be patient -- it is coming.
Thanks for caring.
dp
Steven C. Den Beste
Robato Yao
>On Wed, 2 Sep 1998 20:44:27 -0700, josco <jo...@sea.monterey.edu> added
>this to the Deja-News archive:
>
>>On Thu, 3 Sep 1998, Steven C. Den Beste wrote:
>>
>>> Symantec has announced a new Java Just In Time compiler called JITWin,
>>> which is designed as a plugin for Communicator 4.05.
>>>
>>> http://www.symantec.com/domain/cafe/jitwin/index.html
>>>
>>> The following page shows CaffeineMark 3.0 tests of this new JIT compared to
>>> the stock Netscape Java and compared to Microsoft's latest JIT:
>>>
>>> http://www.symantec.com/domain/cafe/jitwin/caffeine.html
>>>
>>> Now if I remember correctly, when the claims were coming down the pike that
>>> OS/2's Java was fastest, it was the case that it just edged out the
>>> Microsoft JIT by a few percent on comparable hardware. (Do I remember
>>> wrong?)
More than a few percent actually.
Yes hehe, but we get ours FREE. $0.00. And we will get 1.1.7 when it
comes out, FREE, as well. And 1.1.6 had been tweaked since the last
time (several months ago) when it beat MS's JVM for speed, so unless
there are specific Caffeinmarks to be shown to be compared with, then
it's still a claim. Also it's announced, I don't know and
I don't think the Symantec JVM is really shipping yet.
In the meantime, nonetheless, I should applaud such a product as every
OS/2 user should. Why? Because it's a Microsoft competitor JVM
supporting a Microsoft competitor Browser. Communicator always get
trounced by IE in terms of Java speed. Now Windows users can support
aTrue Java JVM instead instead of MIcrosoft's corrupted JVM. This helps
the cause of Sun, the cause of Java, the cause of Netscape.
Rgds,
Chris
(counting down from top 50 oxymorons...)
10. Tight slacks
9. Definite maybe
8. Pretty ugly
7. Twelve-ounce pound cake
6. Diet ice cream
5. Rap music
4. Working vacation
3. Exact estimate
2. Religious tolerance
And the NUMBER ONE top oxy-MORON
1. Microsoft Works
---From the Top 50 Oxymorons (thanks to Richard Kennedy)
Bob O
wrote:
>On Wed, 2 Sep 1998 20:44:27 -0700, josco <jo...@sea.monterey.edu> added
>this to the Deja-News archive:
>
>>On Thu, 3 Sep 1998, Steven C. Den Beste wrote:
>>
>>> Symantec has announced a new Java Just In Time compiler called JITWin,
>>> which is designed as a plugin for Communicator 4.05.
>>>
>>> http://www.symantec.com/domain/cafe/jitwin/index.html
>>>
>>> The following page shows CaffeineMark 3.0 tests of this new JIT compared to
>>> the stock Netscape Java and compared to Microsoft's latest JIT:
>>>
>>> http://www.symantec.com/domain/cafe/jitwin/caffeine.html
>>>
>>> Now if I remember correctly, when the claims were coming down the pike that
>>> OS/2's Java was fastest, it was the case that it just edged out the
>>> Microsoft JIT by a few percent on comparable hardware. (Do I remember
>>> wrong?)
>>>
That was a good trump card but here is a higher one.
Shouldn't you be consistent when you say that the absorption of OS/2
into a larger group with Java is a sign that OS/2 is in trouble and what
your speculation was when IE was absorbed into the operating system
group at Microsoft is a sign that IE will be better integrated into the
operating system?.
Bob O - Computing for fun
Steven C. Den Beste
this to the Deja-News archive:
>In <35f40f8b.5056796@news-server>, sden...@san.rr.com (Steven C. Den Beste) writes:
>>In the mean time, the Symantec JIT is shipping now. (Cheap, too: $40)
>
>
>Yes hehe, but we get ours FREE. $0.00. And we will get 1.1.7 when it
>comes out, FREE, as well.
And tell me: how attractive a customer does that make you when your vendor
gets no income from you for his new product?
How does he make a profitable business out of people who expect to get
things for FREE (hehe)?
Bob O
wrote:
Heeheehee! Steven really has to stretch to find a good reason for why
he pays for his upgrades.<g>
I guess we could call it charity for the poor and Bill Gates is the
number one poster child for the poor on his list of charities.
Robato Yao
>On 3 Sep 1998 07:02:03 GMT, cro...@kuentos.guam.net (Robato Yao) added
>this to the Deja-News archive:
>
>>In <35f40f8b.5056796@news-server>, sden...@san.rr.com (Steven C. Den Beste) writes:
>>>In the mean time, the Symantec JIT is shipping now. (Cheap, too: $40)
>>
>>
>>Yes hehe, but we get ours FREE. $0.00. And we will get 1.1.7 when it
>>comes out, FREE, as well.
>
>And tell me: how attractive a customer does that make you when your vendor
>gets no income from you for his new product?
>
>How does he make a profitable business out of people who expect to get
>things for FREE (hehe)?
Isn't Microsoft subsidizing its own Java as well?
Is Sun making money with Java for Windows users?
An even louder Hahaha, Steven. IBM does not count on personal users for
its income, Steven. It counts on those large lucrative service
contracts to Banks and other institutions, Steven. Lots of lots of
money. A free Java is the result of trickle down effect. OS/2's Java
is the prototype for all IBM JVMs, including those for AS/400, RS/6000
and OS/390. IBM's Java efforts may be the biggest in the world in
terms of manpower, with well over 2,000 developers worldwide. IBM can
be credited with perhaps most large scale commercial installations using
Java, a record of deployment which even Sun has not achieved. Even Sun
has accepted a lot of IBM's own work into its own JVM as a partner,
which you can't say of Symantec. Sun is even working with IBM for the
next generation JavaOS.
Just remember to pick up your own jaw from the floor as you walk out the
door.
Steven C. Den Beste
the Deja-News archive:
>On Fri, 4 Sep 1998 03:14:45, sden...@san.rr.com (Steven C. Den Beste)
>wrote:
>
>>On 3 Sep 1998 07:02:03 GMT, cro...@kuentos.guam.net (Robato Yao) added
>>this to the Deja-News archive:
>>
>>>In <35f40f8b.5056796@news-server>, sden...@san.rr.com (Steven C. Den Beste) writes:
>>>>In the mean time, the Symantec JIT is shipping now. (Cheap, too: $40)
>>>
>>>
>>>Yes hehe, but we get ours FREE. $0.00. And we will get 1.1.7 when it
>>>comes out, FREE, as well.
>>
>>And tell me: how attractive a customer does that make you when your vendor
>>gets no income from you for his new product?
>>
>>How does he make a profitable business out of people who expect to get
>>things for FREE (hehe)?
>>
>
>Heeheehee! Steven really has to stretch to find a good reason for why
>he pays for his upgrades.<g>
>
>I guess we could call it charity for the poor and Bill Gates is the
>number one poster child for the poor on his list of charities.
>
>
>Bob O - Computing for fun
There's no stretch, and it's an important point.
A business relationship is a two-way street, and it has to be a win for
both sides if it is to be relied upon for the long run. If one side
benefits from it but the other side is harmed, then the second party will
eventually figure out a way to get out from under.
Let's distinguish between upgrades and bug fixes, first of all. Bug fixes
come under the logical heading of warranty work, and I consider them to be
paid for ahead of time when the product was purchased.
But upgrades are different. When new features are provided for the product,
the company which sells them has invested resources (money!) into creation
of them, and if they can't get any revenue back to compensate them for it,
then it's a losing proposition for them.
I want my suppliers to make a profit off of the business they do with me
and the other people who are like me. Not only does that make them value me
and mine as a market segment and enhance the chances they'll continue to
provide products for me, but it also serves to attract other companies to
my market segment.
I don't want my suppliers to lose money on my business. That will make them
stop providing me with products, and will dissuade other developers from
moving into my market segment. Why would any developer want to cater to a
market where no profits are to be had?
I take the long view. That money I spend for an upgrade which someone else
on a different platform might get for free is money invested in making my
market more attractive than his market is. Perhaps a short term loss for
me, but definitely a long term gain: because I'm willing to spend money and
he's a leech, I'm a more attractive customer and in the long run I'll get
better product availability and better support from a broader spectrum of
developers.
Keep laughing, funny boy, as you continue to watch the rate of new releases
for OS/2 shrink.
And explain to me why any vendor would want to develop for a customer base
which expects to receive the resulting product for free.
Steven C. Den Beste
this to the Deja-News archive:
>In <35f15aca.1540572@news-server>, sden...@san.rr.com (Steven C. Den Beste) writes:
>>On 3 Sep 1998 07:02:03 GMT, cro...@kuentos.guam.net (Robato Yao) added
>>this to the Deja-News archive:
>>
>>>In <35f40f8b.5056796@news-server>, sden...@san.rr.com (Steven C. Den Beste) writes:
>>>>In the mean time, the Symantec JIT is shipping now. (Cheap, too: $40)
>>>
>>>
>>>Yes hehe, but we get ours FREE. $0.00. And we will get 1.1.7 when it
>>>comes out, FREE, as well.
>>
>>And tell me: how attractive a customer does that make you when your vendor
>>gets no income from you for his new product?
>>
>>How does he make a profitable business out of people who expect to get
>>things for FREE (hehe)?
>
>Isn't Microsoft subsidizing its own Java as well?
Microsoft is subsidizing Java because Java is a threat to WIN32. Microsoft
needs to fractionate Java into two or more largely incompatible types, and
to do that, Microsoft's own version of Java needs to be widespread and
successful.
This is a special case. Microsoft will make money on its Java indirectly
because they expect it to help them preserve their WIN32 business.
>Is Sun making money with Java for Windows users?
They aren't making money on Java at all, and it's causing a ruckus at Sun.
Top management at Sun has been rattling the bars at JavaSoft (or whatever
it's now called since the reorg) about that very fact: "When are we going
to see some income?" The problem is that JavaSoft hasn't had a business
model which was actually going to go profitable, and Sun top management
hasn't been happy about that.
Actually, Sun as a whole expected to make its money selling Java-based
workstations. But for such workstations to survive and prosper, they have a
nontrivial bootstrapping problem: no-one will buy the workstations unless
there's software for them, and no-one is going to develop software for them
until there's a reasonable installed base. Sun's way out of that
bootstrapping problem is to make it so that Java runs on something else
widespread, thus instantly creating a big installed base to attract the
software developers. Then after the apps start flowing, the Java-based
workstations become attractive and Sun will finally start making a profit
off of all of this.
So they are trying to get Java working everywhere, but in particular they
need it to run on Windows. Again, they don't expect it to make them money
directly, but they must have it as an essential part of the process of
starting up their Java-based workstation market.
In other words, it's another special case. They will make their profit on
Java for Windows by the effect that its existence has on their ability to
start selling Java-based workstations.
>An even louder Hahaha, Steven. IBM does not count on personal users for
>its income, Steven.
And you think this is a good thing? Wouldn't you rather they *did* think of
you as a valued customer and actually target their development *to* you?
>It counts on those large lucrative service
>contracts to Banks and other institutions, Steven. Lots of lots of
>money. A free Java is the result of trickle down effect. OS/2's Java
>is the prototype for all IBM JVMs, including those for AS/400, RS/6000
>and OS/390. IBM's Java efforts may be the biggest in the world in
>terms of manpower, with well over 2,000 developers worldwide. IBM can
>be credited with perhaps most large scale commercial installations using
>Java, a record of deployment which even Sun has not achieved. Even Sun
>has accepted a lot of IBM's own work into its own JVM as a partner,
>which you can't say of Symantec. Sun is even working with IBM for the
>next generation JavaOS.
>
>Just remember to pick up your own jaw from the floor as you walk out the
>door.
Sorry, I'm not impressed. My jaw is firmly held in place right where it
ought to be.
Bob O
wrote:
>On 4 Sep 1998 05:53:14 GMT, osb...@deletemeibm.net (Bob O) added this to
>the Deja-News archive:
>
>>On Fri, 4 Sep 1998 03:14:45, sden...@san.rr.com (Steven C. Den Beste)
>>wrote:
>>
>>>On 3 Sep 1998 07:02:03 GMT, cro...@kuentos.guam.net (Robato Yao) added
>>>this to the Deja-News archive:
>>>
>>>>In <35f40f8b.5056796@news-server>, sden...@san.rr.com (Steven C. Den Beste) writes:
>>>>>In the mean time, the Symantec JIT is shipping now. (Cheap, too: $40)
>>>>
>>>>
>>>>Yes hehe, but we get ours FREE. $0.00. And we will get 1.1.7 when it
>>>>comes out, FREE, as well.
>>>
>>>And tell me: how attractive a customer does that make you when your vendor
>>>gets no income from you for his new product?
>>>
>>>How does he make a profitable business out of people who expect to get
>>>things for FREE (hehe)?
>>>
>>
>>Heeheehee! Steven really has to stretch to find a good reason for why
>>he pays for his upgrades.<g>
>>
>>I guess we could call it charity for the poor and Bill Gates is the
>>number one poster child for the poor on his list of charities.
>>
>>
>>Bob O - Computing for fun
>
>There's no stretch, and it's an important point.
>
>A business relationship is a two-way street, and it has to be a win for
>both sides if it is to be relied upon for the long run. If one side
>benefits from it but the other side is harmed, then the second party will
>eventually figure out a way to get out from under.
>
>Let's distinguish between upgrades and bug fixes, first of all. Bug fixes
>come under the logical heading of warranty work, and I consider them to be
>paid for ahead of time when the product was purchased.
Are you saying you paid for more bug fixes for Win95? Do you expect
that they will be provided?
>But upgrades are different. When new features are provided for the product,
>the company which sells them has invested resources (money!) into creation
>of them, and if they can't get any revenue back to compensate them for it,
>then it's a losing proposition for them.
When do you think Microsoft will provide you with an upgrade to Win9x
that includes all the technology advances that OS/2 was provided in 1992
and since?
>I want my suppliers to make a profit off of the business they do with me
>and the other people who are like me. Not only does that make them value me
>and mine as a market segment and enhance the chances they'll continue to
>provide products for me, but it also serves to attract other companies to
>my market segment.
Why would you want that? Most companies like to have a technology
advantage over other companies. My feeling is that I want my supplier
to make a profit if and only if he is providing me with cost effective
state of the art improvements. I do not like seeing him make a profit
on under featured upgrades that are more of a bug fix than anything else
(as we agree the bug fix was already paid for).
>I don't want my suppliers to lose money on my business. That will make them
>stop providing me with products, and will dissuade other developers from
>moving into my market segment. Why would any developer want to cater to a
>market where no profits are to be had?
I would like to see them make a profit if and only if they are serving
my needs. Unfortunately, monopolies tend to maximize profits when they
minimize services. I suspect that we are a hop skip and jump away from
Windows becoming a rental license so that you can pay for the technology
in perpetuity without bug fixes or upgrades.
>I take the long view. That money I spend for an upgrade which someone else
>on a different platform might get for free is money invested in making my
>market more attractive than his market is. Perhaps a short term loss for
>me, but definitely a long term gain: because I'm willing to spend money and
>he's a leech, I'm a more attractive customer and in the long run I'll get
>better product availability and better support from a broader spectrum of
>developers.
Why do you think you will get an opportunity to buy the upgrade and he
won't? Aren't you just being the "early adopter" who pays for the
technology while the late comer can jump on the bandwagon after the
development is paid for and the 30% per annum profits have been pocketed
by your supplier.
Personally because of poor support for their products and a tendency to
charge for bug fixes, I tend to think the best way to buy a MS operating
system is in the thrift store. At least then you know what you are
paying for.
>Keep laughing, funny boy, as you continue to watch the rate of new releases
>for OS/2 shrink.
What is this! Steven Den Beste getting rude? I would like to see more
support for OS/2 from a number of sectors, however, I still use OS/2
because it is a better choice for me. When I buy OS/2 software it runs,
when I buy Win16 software it runs. To choose Windows I either have to
choose a wimpy operating system that won't stand up to the abuses I lay
on it, or I have to select NT and put up with software that does not
"just run" when you install it. And because of the confusing logo work
done by Microsoft over time there is really no way to look at the label
on a Win32 package and get comfort that it will work properly and
completely on NT.
>And explain to me why any vendor would want to develop for a customer base
>which expects to receive the resulting product for free.
Which customer base expects to receive product for free?
Bob O
wrote:
>On 4 Sep 1998 06:29:10 GMT, cro...@kuentos.guam.net (Robato Yao) added
>this to the Deja-News archive:
>
>>In <35f15aca.1540572@news-server>, sden...@san.rr.com (Steven C. Den Beste) writes:
>>>On 3 Sep 1998 07:02:03 GMT, cro...@kuentos.guam.net (Robato Yao) added
>>>this to the Deja-News archive:
>>>
>>>>In <35f40f8b.5056796@news-server>, sden...@san.rr.com (Steven C. Den Beste) writes:
>>>>>In the mean time, the Symantec JIT is shipping now. (Cheap, too: $40)
>>>>
>>>>
>>>>Yes hehe, but we get ours FREE. $0.00. And we will get 1.1.7 when it
>>>>comes out, FREE, as well.
>>>
>>>And tell me: how attractive a customer does that make you when your vendor
>>>gets no income from you for his new product?
>>>
>>>How does he make a profitable business out of people who expect to get
>>>things for FREE (hehe)?
>>
>>Isn't Microsoft subsidizing its own Java as well?
>
>Microsoft is subsidizing Java because Java is a threat to WIN32. Microsoft
>needs to fractionate Java into two or more largely incompatible types, and
>to do that, Microsoft's own version of Java needs to be widespread and
>successful.
>
>This is a special case. Microsoft will make money on its Java indirectly
>because they expect it to help them preserve their WIN32 business.
You mean replace Win32 in order to save it? Heeheehee!
Robato Yao
>On 4 Sep 1998 06:29:10 GMT, cro...@kuentos.guam.net (Robato Yao) added
>this to the Deja-News archive:
>
>>In <35f15aca.1540572@news-server>, sden...@san.rr.com (Steven C. Den Beste) writes:
>>>On 3 Sep 1998 07:02:03 GMT, cro...@kuentos.guam.net (Robato Yao) added
>>>this to the Deja-News archive:
>>>
>>>>In <35f40f8b.5056796@news-server>, sden...@san.rr.com (Steven C. Den Beste) writes:
>>>>>In the mean time, the Symantec JIT is shipping now. (Cheap, too: $40)
>>>>
>>>>
>>>>Yes hehe, but we get ours FREE. $0.00. And we will get 1.1.7 when it
>>>>comes out, FREE, as well.
>>>
>>>And tell me: how attractive a customer does that make you when your vendor
>>>gets no income from you for his new product?
>>>
>>>How does he make a profitable business out of people who expect to get
>>>things for FREE (hehe)?
>>
>>Isn't Microsoft subsidizing its own Java as well?
>
>Microsoft is subsidizing Java because Java is a threat to WIN32. Microsoft
>needs to fractionate Java into two or more largely incompatible types, and
>to do that, Microsoft's own version of Java needs to be widespread and
>successful.
>
>This is a special case. Microsoft will make money on its Java indirectly
>because they expect it to help them preserve their WIN32 business.
But so far it has helped spread Java by the way, and earned them an
expensive lawsuit that continues to attract the DOJ's attention.
>
>>Is Sun making money with Java for Windows users?
>
>They aren't making money on Java at all, and it's causing a ruckus at Sun.
>Top management at Sun has been rattling the bars at JavaSoft (or whatever
>it's now called since the reorg) about that very fact: "When are we going
>to see some income?" The problem is that JavaSoft hasn't had a business
>model which was actually going to go profitable, and Sun top management
>hasn't been happy about that.
Sun's top management? McNealy is the foremost champion of Java and he
hates Windows dear. Be careful what you talk. They may have
reorganized because Javasoft as it is, is not an optimal organization to
get it spread. Don't underestimate Sun's determination on this issue.
>
>Actually, Sun as a whole expected to make its money selling Java-based
>workstations. But for such workstations to survive and prosper, they have a
>nontrivial bootstrapping problem: no-one will buy the workstations unless
>there's software for them, and no-one is going to develop software for them
>until there's a reasonable installed base. Sun's way out of that
>bootstrapping problem is to make it so that Java runs on something else
>widespread, thus instantly creating a big installed base to attract the
>software developers. Then after the apps start flowing, the Java-based
>workstations become attractive and Sun will finally start making a profit
>off of all of this.
>
>So they are trying to get Java working everywhere, but in particular they
>need it to run on Windows. Again, they don't expect it to make them money
>directly, but they must have it as an essential part of the process of
>starting up their Java-based workstation market.
>
>In other words, it's another special case. They will make their profit on
>Java for Windows by the effect that its existence has on their ability to
>start selling Java-based workstations.
Wrong.
Sun makes money on SERVERS not workstations. Workstations and NCs are
only adjunct to the SERVER. Ultimately it does not matter if the Java
is on a PC or on a workstation, the point is, you still have a Java
frontend serving the Java application running on the servers.
Sun must keep up a technological edge when it comes to networking and
internet technologies, and Java is such. That's how Sun and IBM retain
the server advantage.
>
>
>>An even louder Hahaha, Steven. IBM does not count on personal users for
>>its income, Steven.
>
>And you think this is a good thing? Wouldn't you rather they *did* think of
>you as a valued customer and actually target their development *to* you?
>
Yes. As a company owner with an IBM server, they did *target* me.
>>It counts on those large lucrative service
>>contracts to Banks and other institutions, Steven. Lots of lots of
>>money. A free Java is the result of trickle down effect. OS/2's Java
>>is the prototype for all IBM JVMs, including those for AS/400, RS/6000
>>and OS/390. IBM's Java efforts may be the biggest in the world in
>>terms of manpower, with well over 2,000 developers worldwide. IBM can
>>be credited with perhaps most large scale commercial installations using
>>Java, a record of deployment which even Sun has not achieved. Even Sun
>>has accepted a lot of IBM's own work into its own JVM as a partner,
>>which you can't say of Symantec. Sun is even working with IBM for the
>>next generation JavaOS.
>>
>>Just remember to pick up your own jaw from the floor as you walk out the
>>door.
>
>Sorry, I'm not impressed. My jaw is firmly held in place right where it
>ought to be.
You make far too lengthy posts that are nothing but BS and don't add to
anything.
Pwahahah.
josco
> In <35f1fa44.1264602@news-server>, sden...@san.rr.com (Steven C. Den Beste) writes:
> >This is a special case. Microsoft will make money on its Java indirectly
> >because they expect it to help them preserve their WIN32 business.
>
> But so far it has helped spread Java by the way, and earned them an
> expensive lawsuit that continues to attract the DOJ's attention.
http://www.javaworld.com/javaworld/jw-08-1998/jw-08-volanomark.html
"I find it ironic that the one company with a not-so-secret objective to
"kill cross-platform Java" (according to an internal Microsoft document;
see Resources below) also is the Java vendor that kept our 100% Pure Java
company (Volano) in business for the past year and a half. Without the
impressive performance of Microsoft's Java virtual machine, we would not
have acquired the many VolanoChat customers who run our chat server on
high-traffic Web sites."
Curtis Bass
> On 3 Sep 1998 07:02:03 GMT, cro...@kuentos.guam.net (Robato Yao) added
> this to the Deja-News archive:
>
>
-- snip --
> >Yes hehe, but we get ours FREE. $0.00. And we will get 1.1.7 when it
> >comes out, FREE, as well.
>
> And tell me: how attractive a customer does that make you when your vendor
> gets no income from you for his new product?
>
> How does he make a profitable business out of people who expect to get
> things for FREE (hehe)?
Good questions, Steven! And an even better question: Why weren't you asking these
questions of Microsoft when they were busy giving Internet Explorer away *for free??*
Hmmm?
Curtis
Steven C. Den Beste
the Deja-News archive:
Absolutely they will be provided. Hundreds have already. Microsoft doesn't
bundle them into monumental fixpacks like IBM does for OS/2, but they're
coming out all the time. They just get released individually.
Try taking a look at the following directory:
ftp://ftp.microsoft.com/Softlib/MSLFILES
Every file in there is a patch for either Win 95 or Win NT or Win 98 (not
too many for 98 yet). I downloaded two patches for Win 98 just today.
>>But upgrades are different. When new features are provided for the product,
>>the company which sells them has invested resources (money!) into creation
>>of them, and if they can't get any revenue back to compensate them for it,
>>then it's a losing proposition for them.
>
>When do you think Microsoft will provide you with an upgrade to Win9x
>that includes all the technology advances that OS/2 was provided in 1992
>and since?
About the same time that IBM releases an upgrade to OS/2 that provides the
features which are in Win 98 such as multi-monitor support and a good
comprehensive game-support API.
The simple fact is that there's nothing OS/2 provides above and beyond what
Win 98 provides which I want.
>>I want my suppliers to make a profit off of the business they do with me
>>and the other people who are like me. Not only does that make them value me
>>and mine as a market segment and enhance the chances they'll continue to
>>provide products for me, but it also serves to attract other companies to
>>my market segment.
>
>Why would you want that? Most companies like to have a technology
>advantage over other companies. My feeling is that I want my supplier
>to make a profit if and only if he is providing me with cost effective
>state of the art improvements. I do not like seeing him make a profit
>on under featured upgrades that are more of a bug fix than anything else
>(as we agree the bug fix was already paid for).
No, upgrades are not the same as bug fixes. A bug fix takes a feature which
is already present but which does not work properly, and makes it work. An
upgrade provides a new capability not previously present. The distinction
is critical.
If I get something new, I expect to (and am *eager* to) pay for it, because
doing so encourages companies to provide more of it. If they can't make
money by providing enhancements, why would they bother to create them?
Since I want them to do this, I want to make it worth their while.
>>I don't want my suppliers to lose money on my business. That will make them
>>stop providing me with products, and will dissuade other developers from
>>moving into my market segment. Why would any developer want to cater to a
>>market where no profits are to be had?
>
>I would like to see them make a profit if and only if they are serving
>my needs. Unfortunately, monopolies tend to maximize profits when they
>minimize services. I suspect that we are a hop skip and jump away from
>Windows becoming a rental license so that you can pay for the technology
>in perpetuity without bug fixes or upgrades.
>
>>I take the long view. That money I spend for an upgrade which someone else
>>on a different platform might get for free is money invested in making my
>>market more attractive than his market is. Perhaps a short term loss for
>>me, but definitely a long term gain: because I'm willing to spend money and
>>he's a leech, I'm a more attractive customer and in the long run I'll get
>>better product availability and better support from a broader spectrum of
>>developers.
>
>Why do you think you will get an opportunity to buy the upgrade and he
>won't? Aren't you just being the "early adopter" who pays for the
>technology while the late comer can jump on the bandwagon after the
>development is paid for and the 30% per annum profits have been pocketed
>by your supplier.
I suppose I should have been more specific; I left out the fact that the
two markets are incompatible and that development for one does the other no
good. I assumed you realized I was referring to WIN32 and OS/2. If a
developer sees that upgrades developed for WIN32 can be sold, but that
upgrades for OS/2 can only be distributed by giving them away, which market
is more likely to get that developer's resources? (Hint: Ask Brad Wardell.)
And if your answer is "He'll use the money he made from WIN32 to develop
for OS/2", no, I don't think so. Each time a decision is made on
investment, the decision on return enters into it. His next investment will
also be on WIN32 because that is more likely to gain him yet more return
later.
>Personally because of poor support for their products and a tendency to
>charge for bug fixes, I tend to think the best way to buy a MS operating
>system is in the thrift store. At least then you know what you are
>paying for.
We're not just talking about Microsoft here. I buy upgrades to other
products from other vendors. I register my shareware and spend real dollars
to encourage more. And there are enough more like me that we represent a
significant economic force, which motivates software developers.
>>Keep laughing, funny boy, as you continue to watch the rate of new releases
>>for OS/2 shrink.
>
>What is this! Steven Den Beste getting rude? I would like to see more
>support for OS/2 from a number of sectors, however, I still use OS/2
>because it is a better choice for me. When I buy OS/2 software it runs,
>when I buy Win16 software it runs. To choose Windows I either have to
>choose a wimpy operating system that won't stand up to the abuses I lay
>on it, or I have to select NT and put up with software that does not
>"just run" when you install it. And because of the confusing logo work
>done by Microsoft over time there is really no way to look at the label
>on a Win32 package and get comfort that it will work properly and
>completely on NT.
>
>>And explain to me why any vendor would want to develop for a customer base
>>which expects to receive the resulting product for free.
>
>Which customer base expects to receive product for free?
>
>Bob O - Computing for fun
The OS/2 customer base. For instance, consider how Brad Wardell got roasted
when he decided not to release Fixkit #3 to Object Desktop 1.0.
For instance, consider how OS/2 users in this forum are expecting that when
Communicator for OS/2 is released, that it will be free. If it turns out to
require a paid license for Software Choice, expect a lot of people to go
ballistic and accuse IBM of money-grubbing.
Steven C. Den Beste
the Deja-News archive:
>On Fri, 4 Sep 1998 14:42:25, sden...@san.rr.com (Steven C. Den Beste)
>wrote:
>
>>On 4 Sep 1998 06:29:10 GMT, cro...@kuentos.guam.net (Robato Yao) added
>>this to the Deja-News archive:
>>
>>>In <35f15aca.1540572@news-server>, sden...@san.rr.com (Steven C. Den Beste) writes:
>>>>On 3 Sep 1998 07:02:03 GMT, cro...@kuentos.guam.net (Robato Yao) added
>>>>this to the Deja-News archive:
>>>>
>>>>>In <35f40f8b.5056796@news-server>, sden...@san.rr.com (Steven C. Den Beste) writes:
>>>>>>In the mean time, the Symantec JIT is shipping now. (Cheap, too: $40)
>>>>>
>>>>>
>>>>>Yes hehe, but we get ours FREE. $0.00. And we will get 1.1.7 when it
>>>>>comes out, FREE, as well.
>>>>
>>>>And tell me: how attractive a customer does that make you when your vendor
>>>>gets no income from you for his new product?
>>>>
>>>>How does he make a profitable business out of people who expect to get
>>>>things for FREE (hehe)?
>>>
>>>Isn't Microsoft subsidizing its own Java as well?
>>
>>Microsoft is subsidizing Java because Java is a threat to WIN32. Microsoft
>>needs to fractionate Java into two or more largely incompatible types, and
>>to do that, Microsoft's own version of Java needs to be widespread and
>>successful.
>>
>>This is a special case. Microsoft will make money on its Java indirectly
>>because they expect it to help them preserve their WIN32 business.
>
>You mean replace Win32 in order to save it? Heeheehee!
>
>
>Bob O - Computing for fun
Quite the opposite: to make Java into a language for developing WIN32 apps
instead of a language to make portable apps -- in order to save WIN32.
Robato Yao
Please Steven, you should have stopped when you are still making sense.
Steven C. Den Beste
this to the Deja-News archive:
>Steven C. Den Beste wrote:
>
>> On 3 Sep 1998 07:02:03 GMT, cro...@kuentos.guam.net (Robato Yao) added
>> this to the Deja-News archive:
>>
>>
>
>-- snip --
>
>> >Yes hehe, but we get ours FREE. $0.00. And we will get 1.1.7 when it
>> >comes out, FREE, as well.
>>
>> And tell me: how attractive a customer does that make you when your vendor
>> gets no income from you for his new product?
>>
>> How does he make a profitable business out of people who expect to get
>> things for FREE (hehe)?
>
>
>Good questions, Steven! And an even better question: Why weren't you asking these
>questions of Microsoft when they were busy giving Internet Explorer away *for free??*
>
>Hmmm?
>
>Curtis
>
Oh, I didn't need to ask because I already knew the answer.
IE was given away because Navigator was a threat to Microsoft operating
systems. About the time Microsoft started working on IE, Netscape started
saying things like "The browser *is* the OS". The threat was clear.
It must be remembered that at that time Navigator had something like an 80%
market share in the browser market.
Just as with Microsoft's Java, Microsoft was giving IE away as a means of
defending and preserving its core WIN32 market.
But: had they charged for IE instead of giving it away, I'd have bought a
copy.
I did buy a copy of Communicator even after downloading IE4. I also bought
a copy of Opera, even after having both Communicator and IE4.
Bob O
wrote:
>>Are you saying you paid for more bug fixes for Win95? Do you expect
>>that they will be provided?
>
>Absolutely they will be provided. Hundreds have already. Microsoft doesn't
>bundle them into monumental fixpacks like IBM does for OS/2, but they're
>coming out all the time. They just get released individually.
>Every file in there is a patch for either Win 95 or Win NT or Win 98 (not
>too many for 98 yet). I downloaded two patches for Win 98 just today.
The question wasn't whether Win98 patches would be provided the question
was do you expect more Win95 patches? And since NT is supposedly the
successor to Win98 how long do you think MS will be providing Win98
patches?
It is an interesting question as MS has said that Win98 is a dead end.
But one has to really wonder if that is indeed the case. I happened to
have been in a CompUSA today and on a lark I decided to see what OS's
they offered with their new hardware. They had computers in the store
that offered 3 operating systems. The Mac, Win98, and Win95. Not one
machine was being sold with NT on it. Last year there were several NT
machines in the store, but they have disappeared. So I wandered over
to the operating system area and found two shelf slots for OS/2 Warp 4
and none for NT. After searching carefully I found to NT workstation
copies hiding behind a copy of Windows 98. Even OS/2 had more shelf
space than NT though.
It would appear that NT is dying.
>>>But upgrades are different. When new features are provided for the product,
>>>the company which sells them has invested resources (money!) into creation
>>>of them, and if they can't get any revenue back to compensate them for it,
>>>then it's a losing proposition for them.
>>
>>When do you think Microsoft will provide you with an upgrade to Win9x
>>that includes all the technology advances that OS/2 was provided in 1992
>>and since?
>
>About the same time that IBM releases an upgrade to OS/2 that provides the
>features which are in Win 98 such as multi-monitor support and a good
>comprehensive game-support API.
>
>The simple fact is that there's nothing OS/2 provides above and beyond what
>Win 98 provides which I want.
You don't want improved stability, better multitasking, capability to
handle more threads, more efficient disk space usage or safer and more
dependable file naming conventions?
You don't have any expectation of bugs getting fixed?
>For instance, consider how OS/2 users in this forum are expecting that when
>Communicator for OS/2 is released, that it will be free. If it turns out to
>require a paid license for Software Choice, expect a lot of people to go
>ballistic and accuse IBM of money-grubbing.
So you were simply speculating that there is a large number of OS/2
users that never expect to pay for any more software?
Bob O
wrote:
>On 4 Sep 1998 19:05:01 GMT, osb...@deletemeibm.net (Bob O) added this to
>the Deja-News archive:
>
>>On Fri, 4 Sep 1998 14:42:25, sden...@san.rr.com (Steven C. Den Beste)
>>wrote:
>>
>>>On 4 Sep 1998 06:29:10 GMT, cro...@kuentos.guam.net (Robato Yao) added
>>>this to the Deja-News archive:
>>>
>>>>In <35f15aca.1540572@news-server>, sden...@san.rr.com (Steven C. Den Beste) writes:
>>>>>On 3 Sep 1998 07:02:03 GMT, cro...@kuentos.guam.net (Robato Yao) added
>>>>>this to the Deja-News archive:
>>>>>
>>>>>>In <35f40f8b.5056796@news-server>, sden...@san.rr.com (Steven C. Den Beste) writes:
>>>>>>>In the mean time, the Symantec JIT is shipping now. (Cheap, too: $40)
>>>>>>
>>>>>>
>>>>>>Yes hehe, but we get ours FREE. $0.00. And we will get 1.1.7 when it
>>>>>>comes out, FREE, as well.
>>>>>
>>>>>And tell me: how attractive a customer does that make you when your vendor
>>>>>gets no income from you for his new product?
>>>>>
>>>>>How does he make a profitable business out of people who expect to get
>>>>>things for FREE (hehe)?
>>>>
>>>>Isn't Microsoft subsidizing its own Java as well?
>>>
>>>Microsoft is subsidizing Java because Java is a threat to WIN32. Microsoft
>>>needs to fractionate Java into two or more largely incompatible types, and
>>>to do that, Microsoft's own version of Java needs to be widespread and
>>>successful.
>>>
>>>This is a special case. Microsoft will make money on its Java indirectly
>>>because they expect it to help them preserve their WIN32 business.
>>
>>You mean replace Win32 in order to save it? Heeheehee!
>>
>>
>>Bob O - Computing for fun
>
>Quite the opposite: to make Java into a language for developing WIN32 apps
>instead of a language to make portable apps -- in order to save WIN32.
>
Sounds like they are creating some redundant APIs. You think MS will
continue to support both?
Steven C. Den Beste
the Deja-News archive:
>On Sat, 5 Sep 1998 01:42:29, sden...@san.rr.com (Steven C. Den Beste)
NT 4 is not targeted to the home/hobby audience, hence a store which sells
to home/hobby users wouldn't necessarily carry many copies of it. Your
discovery doesn't mean NT is dying.
Corporations are the primary customers of NT 4, and corporations don't
usually pick up their copies at CompUSA.
The fact that Microsoft has announced that there will be no future major
releases doesn't mean that all maintenance and enhancements stop this
instant. On the contrary, they'll be working on them for quite a while.
And Win 95 continues to sell, even though Win 98 is out now. Maintenance
and patches for Win 95 will continue to come out, too. Indeed, Microsoft
still issues patches for Win 3.1 occasionally.
>>>>But upgrades are different. When new features are provided for the product,
>>>>the company which sells them has invested resources (money!) into creation
>>>>of them, and if they can't get any revenue back to compensate them for it,
>>>>then it's a losing proposition for them.
>>>
>>>When do you think Microsoft will provide you with an upgrade to Win9x
>>>that includes all the technology advances that OS/2 was provided in 1992
>>>and since?
>>
>>About the same time that IBM releases an upgrade to OS/2 that provides the
>>features which are in Win 98 such as multi-monitor support and a good
>>comprehensive game-support API.
>>
>>The simple fact is that there's nothing OS/2 provides above and beyond what
>>Win 98 provides which I want.
>
>You don't want improved stability, better multitasking, capability to
>handle more threads, more efficient disk space usage or safer and more
>dependable file naming conventions?
I find Win 98 sufficiently stable for my needs (the myths about crashing
every five minutes notwithstanding, I find that Win 98 crashes about as
often as Warp 3 did when I was using Warp 3 at work); I also find that
FAT32 is sufficiently efficient for my disk space, and most of those other
things don't affect the way I use my computer at home. The one exception is
"file naming conventions", and I *prefer* the way Win 98 does it to the way
OS/2 does it.
They would strongly affect the way I use a computer at work, but at work I
use NT.
I'm not sure which bugs you're referring to. Fixkit #3 for Object Desktop
1.0 won't be coming out, if that's what you're asking about. Wardell has
said so.
Many of the bugs which would have been addressed by it will be fixed in the
next version of Object Desktop, but you'd have to buy it to get them.
The people who roasted Wardell made it very clear that they expected to be
given the fixpack, and given it for free.
>>For instance, consider how OS/2 users in this forum are expecting that when
>>Communicator for OS/2 is released, that it will be free. If it turns out to
>>require a paid license for Software Choice, expect a lot of people to go
>>ballistic and accuse IBM of money-grubbing.
>
>So you were simply speculating that there is a large number of OS/2
>users that never expect to pay for any more software?
See the comments about fixkit #3 for Object Desktop. That's not
speculation, that's fact.
And it is definitely the case that the people posting here are expecting to
get Communicator for OS/2 for free.
Jason
:>Yes hehe, but we get ours FREE. $0.00. And we will get 1.1.7 when it
: And tell me: how attractive a customer does that make you when your vendor
: gets no income from you for his new product?
: How does he make a profitable business out of people who expect to get
: things for FREE (hehe)?
Because Java 1.1.7 is an update to the total OS/2 Warp 4 package. And so
that they can keep selling OS/2 Warp to buisnesses, they have to keep
Java up to date to keep it marketable. Plus telling there potential
customers that regular updates for Java, and TCPIP and System Fixpacks,
and web browser, are included with the OS/2 Warp 4 package they will be
more likely to be able to sell it. It is called selling a product. And
yes, I know you don't believe anyone actually buys OS/2 Warp 4, but
you'll just have to cut me some slack on this one.
-Jason
--
malstrom @ oitunix.oit.umass.edu
http://www-unix.oit.umass.edu/~malstrom/
MP3 info on OS/2 - http://www-unix.oit.umass.edu/~malstrom/mp3.html
William Pridgen
Perhaps some of my fellow OS/2 users and advocates have still not read
"The Moon Is A Harsh Mistress" or "Atlas Shrugged"?
--
Bill Pridgen
--
pri...@texas.net
Bob O
wrote:
Don't you think that the home/soho users should be concerned that
Microsoft is not targeting them?
Wow that is really bad!! Since you consider the occasional necessary
reboot that OS/2 required to clear screen holes and input blockage a
"crash" equivalent to the ugly way that Win95 dies taking data with it
that is really bad!!!
>They would strongly affect the way I use a computer at work, but at work I
>use NT.
I guess if you must have nothing important to do at home, eh?
You didn't answer the question, Steven. I have no idea what bugs as I
don't use OD v.1.0 but you seem to be confirming that indeed there are
bugs and that you have to buy an upgrade to get them fixed. I realize
that is a way of life for a Windows user, but is it right?
>>>For instance, consider how OS/2 users in this forum are expecting that when
>>>Communicator for OS/2 is released, that it will be free. If it turns out to
>>>require a paid license for Software Choice, expect a lot of people to go
>>>ballistic and accuse IBM of money-grubbing.
>>
>>So you were simply speculating that there is a large number of OS/2
>>users that never expect to pay for any more software?
>
>See the comments about fixkit #3 for Object Desktop. That's not
>speculation, that's fact.
I would submit that expecting bugfixes for free is substantially
different than expecting free software with new functionality.
>And it is definitely the case that the people posting here are expecting to
>get Communicator for OS/2 for free.
Users might expect that IBM will match Microsoft in providing free net
software. This might be unrealistic because you are probably asking a
company with an excellent history of providing extensive free bug fix
support to compete at all levels with a omparing a company that gives
you a free bug fest then expects you to buy the upgrade that makes it
work better. Personally, I would rather pay for the software, if I need
it and it is cost effective, and have them continue the excellent
support I have enjoyed over the past 6 1/2 years.
Gerben Bergman
>>Which customer base expects to receive product for free?
>
>The OS/2 customer base. For instance, consider how Brad Wardell got roasted
>when he decided not to release Fixkit #3 to Object Desktop 1.0.
That's Object Desktop 1.5, Steven. You'd be more convincing if you got your
numbers straight. ;-)
>For instance, consider how OS/2 users in this forum are expecting that when
>Communicator for OS/2 is released, that it will be free. If it turns out to
>require a paid license for Software Choice, expect a lot of people to go
>ballistic and accuse IBM of money-grubbing.
And there's some point to it, too; being required to buy a ridiculously
expensive license just to get a browser which is free on every other
platform it runs on is, er, not exactly a good deal.
Graham Murray
> And it is definitely the case that the people posting here are expecting to
> get Communicator for OS/2 for free.
Not an unreasonable expectation considering that (AFAIK) Communicator
4.x (client) is free for all other platforms.
Steven C. Den Beste
the Deja-News archive:
I gather you haven't heard of a recently-released product called "Windows
98"? It was targeted at home/hobby users.
>>>>The simple fact is that there's nothing OS/2 provides above and beyond what
>>>>Win 98 provides which I want.
>>>
>>>You don't want improved stability, better multitasking, capability to
>>>handle more threads, more efficient disk space usage or safer and more
>>>dependable file naming conventions?
>>
>>I find Win 98 sufficiently stable for my needs (the myths about crashing
>>every five minutes notwithstanding, I find that Win 98 crashes about as
>>often as Warp 3 did when I was using Warp 3 at work); I also find that
>>FAT32 is sufficiently efficient for my disk space, and most of those other
>>things don't affect the way I use my computer at home. The one exception is
>>"file naming conventions", and I *prefer* the way Win 98 does it to the way
>>OS/2 does it.
>
>Wow that is really bad!! Since you consider the occasional necessary
>reboot that OS/2 required to clear screen holes and input blockage a
>"crash" equivalent to the ugly way that Win95 dies taking data with it
>that is really bad!!!
I don't lose data when Win 95 crashes, you may be surprised to learn. I
don't recall ever having done so. And as often as not, I had to reboot OS/2
because it had locked up cold.
>>They would strongly affect the way I use a computer at work, but at work I
>>use NT.
>
>I guess if you must have nothing important to do at home, eh?
Depends on what you mean by "important". Critical? Nope. It's a hobby. How
important is a hobby?
Let's look at one of your advantages in greater detail: "more efficient
disk space usage". Generally, since file sizes are random over a large
number of files, a typical file wastes half a cluster.
I just checked, and my 2G C: partition has 18602 files on it (counting
directories).
On a 2G partition, VFAT was using 32K clusters. Since an average file
wasted half of that, 16K, and given 18602 files, then it meant that on VFAT
I was wasting 290 megabytes of disk space.
When I switched to FAT32, my cluster size changed to a much more reasonable
4K. That means that a typical file is now wasting only 2K, yielding a total
wastage of 36 megabytes.
HPFS uses a 256 byte cluster, so the average wastage is 128 bytes. Were I
to convert to HPFS, I would be wasting two and a quarter megabytes.
But look at wastage as a percentage of space used, and you see that the
difference between FAT32 and HPFS really isn't all that significant:
Base data size: 1341 meg (not counting wastage)
VFAT: wastage: 290M total: 1631M efficiency: 82%
FAT32: wastage: 36M total: 1377M efficiency: 97%
HPFS: wastage: 2M total: 1343M efficiency: 99.8%
Yes, HPFS is more efficient. But the difference between 97% and 99.8%,
between 36M wasted and 2M wasted out of a total of 1340M used, is not
enough to bother me.
josco
> On 5 Sep 1998 15:43:43 GMT, osb...@deletemeibm.net (Bob O) added this to
> the Deja-News archive:
> >Don't you think that the home/soho users should be concerned that
> >Microsoft is not targeting them?
>
> I gather you haven't heard of a recently-released product called "Windows
> 98"? It was targeted at home/hobby users.
Oh that old dead end!
MS's real home user computer is WebTV. Their consumer OS is WindowsCE.
IBM's doing far more development work on OS/2 than MS is on Win98. Is MS
doing any more Win98 deelopment?
> I don't lose data when Win 95 crashes, you may be surprised to learn.
I do. I just did yesterday when I was editing a Word97 document my
colloraborator sent me. I lost all my work when Win95 crashed. I had to
power off.
How don't you lose data when the OS crashes?
> don't recall ever having done so. And as often as not, I had to reboot OS/2
> because it had locked up cold.
You sound as evasive as Bill Gates giving a deposititon.
> But look at wastage as a percentage of space used, and you see that the
> difference between FAT32 and HPFS really isn't all that significant:
>
> Base data size: 1341 meg (not counting wastage)
>
> VFAT: wastage: 290M total: 1631M efficiency: 82%
> FAT32: wastage: 36M total: 1377M efficiency: 97%
> HPFS: wastage: 2M total: 1343M efficiency: 99.8%
>
> Yes, HPFS is more efficient. But the difference between 97% and 99.8%,
> between 36M wasted and 2M wasted out of a total of 1340M used, is not
> enough to bother me.
What was your point?
Since the wasting of 290 MB on VFAT did not bother you, why would you even
bother to do the math on FAT32?
Your prejudices are driving your selection. You proved that when you were
not bothered by VFAT's horrible waste of diskspace.
Steven C. Den Beste
this to the Deja-News archive:
>On Sat, 5 Sep 1998, Steven C. Den Beste wrote:
>
>> On 5 Sep 1998 15:43:43 GMT, osb...@deletemeibm.net (Bob O) added this to
>> the Deja-News archive:
>
>> >Don't you think that the home/soho users should be concerned that
>> >Microsoft is not targeting them?
>>
>> I gather you haven't heard of a recently-released product called "Windows
>> 98"? It was targeted at home/hobby users.
>
>Oh that old dead end!
At least Microsoft is being honest about it.
>MS's real home user computer is WebTV. Their consumer OS is WindowsCE.
I don't agree. But you always seem to think that they can only have one of
something. They have *several* products targeted simultaneously at the
consumer market.
Win 98, WebTV and WinCE are each aimed at a different kind of user.
>IBM's doing far more development work on OS/2 than MS is on Win98. Is MS
>doing any more Win98 deelopment?
Quite a lot, actually. They just released DirectX 6 for Win 98; they'll be
releasing IE5 for Win 98 when it comes out, and there have already been a
couple of patches for it. (I downloaded and installed two yesterday. One
fixed a security hole in IE4, and the other gave the system the ability to
display Eastern European character sets.)
We're expecting the release of 128-bit-encrypted PPTP for Win 98 very
shortly. (It's available now for Win 95 and Win NT, but for the moment Win
98 is limited to 40-bit encryption.)
>> But look at wastage as a percentage of space used, and you see that the
>> difference between FAT32 and HPFS really isn't all that significant:
>>
>> Base data size: 1341 meg (not counting wastage)
>>
>> VFAT: wastage: 290M total: 1631M efficiency: 82%
>> FAT32: wastage: 36M total: 1377M efficiency: 97%
>> HPFS: wastage: 2M total: 1343M efficiency: 99.8%
>>
>> Yes, HPFS is more efficient. But the difference between 97% and 99.8%,
>> between 36M wasted and 2M wasted out of a total of 1340M used, is not
>> enough to bother me.
>
>What was your point?
>
>Since the wasting of 290 MB on VFAT did not bother you, why would you even
>bother to do the math on FAT32?
>
>Your prejudices are driving your selection. You proved that when you were
>not bothered by VFAT's horrible waste of diskspace.
I *was* bothered (and said so) but I didn't have any alternatives. So I
lived with it.
However, once Win 98 came out I bought and installed it so I could switch
to FAT32. By doing that, I moved from 82% efficiency to 97% efficiency.
That's good enough for me.
My point is that Bob was asking me whether I wanted OS/2's efficiency over
the efficiency I now have with Win 98. No, it wouldn't be that big a deal
at this point.
Over VFAT, it would be a big deal. Over FAT32, it's small change. The gain
of 34M on a 2G drive is in the noise.
Jack Troughton
wrote:
> On Sat, 5 Sep 1998 12:17:31 -0700, josco <jo...@sea.monterey.edu> added
> this to the Deja-News archive:
>
> >On Sat, 5 Sep 1998, Steven C. Den Beste wrote:
> >
> >> On 5 Sep 1998 15:43:43 GMT, osb...@deletemeibm.net (Bob O) added this to
> >> the Deja-News archive:
> >
> >> >Don't you think that the home/soho users should be concerned that
> >> >Microsoft is not targeting them?
> >>
> >> I gather you haven't heard of a recently-released product called "Windows
> >> 98"? It was targeted at home/hobby users.
> >
> >Oh that old dead end!
>
> At least Microsoft is being honest about it.
>
> >MS's real home user computer is WebTV. Their consumer OS is WindowsCE.
>
> I don't agree. But you always seem to think that they can only have one of
> something. They have *several* products targeted simultaneously at the
> consumer market.
>
> Win 98, WebTV and WinCE are each aimed at a different kind of user.
>
> >IBM's doing far more development work on OS/2 than MS is on Win98. Is MS
> >doing any more Win98 deelopment?
>
> Quite a lot, actually. They just released DirectX 6 for Win 98; they'll be
> releasing IE5 for Win 98 when it comes out, and there have already been a
> couple of patches for it. (I downloaded and installed two yesterday. One
> fixed a security hole in IE4, and the other gave the system the ability to
> display Eastern European character sets.)
>
> We're expecting the release of 128-bit-encrypted PPTP for Win 98 very
> shortly. (It's available now for Win 95 and Win NT, but for the moment Win
> 98 is limited to 40-bit encryption.)
>
> >> But look at wastage as a percentage of space used, and you see that the
> >> difference between FAT32 and HPFS really isn't all that significant:
> >>
> >> Base data size: 1341 meg (not counting wastage)
> >>
> >> VFAT: wastage: 290M total: 1631M efficiency: 82%
> >> FAT32: wastage: 36M total: 1377M efficiency: 97%
> >> HPFS: wastage: 2M total: 1343M efficiency: 99.8%
> >>
> >> Yes, HPFS is more efficient. But the difference between 97% and 99.8%,
> >> between 36M wasted and 2M wasted out of a total of 1340M used, is not
> >> enough to bother me.
> >
> >What was your point?
> >
> >Since the wasting of 290 MB on VFAT did not bother you, why would you even
> >bother to do the math on FAT32?
> >
> >Your prejudices are driving your selection. You proved that when you were
> >not bothered by VFAT's horrible waste of diskspace.
>
> I *was* bothered (and said so) but I didn't have any alternatives. So I
> lived with it.
>
> However, once Win 98 came out I bought and installed it so I could switch
> to FAT32. By doing that, I moved from 82% efficiency to 97% efficiency.
> That's good enough for me.
>
> My point is that Bob was asking me whether I wanted OS/2's efficiency over
> the efficiency I now have with Win 98. No, it wouldn't be that big a deal
> at this point.
>
> Over VFAT, it would be a big deal. Over FAT32, it's small change. The gain
> of 34M on a 2G drive is in the noise.
>
There is something else that comes with HPFS over vfat, fat32, or any
other kind of fat file system; data integrity. If your file
allocation tables get corrupted, you are out of luck. If some of the
inodes in HPFS get corrupted, well, it won't affect any of the other
ones, and thanks to the use of a B+ tree (that has pointers both up
from the data to the file system as well as from the file system to
the data) your chances of recovering from a bad crash are much much
better.
An anecdote on that; there was a fire in the apartment across the hall
from me. It gutted both the apartment and the utility room.
Fortunately, all I had was smoke damage. A lot of it.
When I got back in and could restart my computer, all seemed fine
except for one drive that refused to spin up. It's a paperweight now.
The other two seemed fine.
About three days later I started having problems with my computer
being unstable. By this point I had already cleaned it up and so
on...
To make a long story short, I had creeping corruption and bad sectors
on the two HDs that were still spinning in my system. Once I saw what
was going on, I powered down my computer (I run it 24/7). I got
myself a SCSI card and hard drive; since I had to replace all of my
hard disks anyway, what the hell. I put it in and started my machine
off of utility floppies.
Out of approximately 3.5 GB of data and software, I lost a total of
about 350 MB; and about 350 of those were on the drive that wouldn't
even spin up. I had to run some serious iterations of chkdsk to get
it all out (using the f:3 paramater) but I got it.
One of my jaz disks decided to get some corruption on it one day
(kinda wish I hadn't blown the cash on that one... the jaz I mean.
Now that I know about the click of death, I'm always a little worried
about it.) I had (and still have, actually) all of my software
install files backed up on it. It decided to accumulate some bad
sectors. I couldn't access the drive anymore. Once again, chkdsk got
my stuff off it. Multiple iterations at level three chkdsk were
necessary, but I got every bite of information that I had put on that
disk back off it again. In fact, now that the bad sectors are marked
bad, it works fine, minus a few megabytes of capacity. The reason why
it was so problematic? Part of the hosed area of the disk was in the
first cylinder; right where the file allocation table would be. If it
had been formatted fat, I would have been completely out of luck.
There's more to the quality of a file system than just the cluster
size.
Jack Troughton ICQ:7494149
http://modemcable162.163.mmtl.videotron.net:8000/
jaft at adan.kingston.net
jack.troughton at videotron.ca
Montréal PQ Canada
There are two kinds of people in this world: the kind that
divides everybody into two kinds of people, and everybody else.
josco
> On Sat, 5 Sep 1998 12:17:31 -0700, josco <jo...@sea.monterey.edu> added
> this to the Deja-News archive:
> >> I gather you haven't heard of a recently-released product called "Windows
> >> 98"? It was targeted at home/hobby users.
> >
> >Oh that old dead end!
>
> At least Microsoft is being honest about it.
What is it going to be in this post?
Are you going to imply IBM is't telling OS/2 users the truth or are you
going to post iBM's java plans for OS/2 and take IBM on it's word?
> >MS's real home user computer is WebTV. Their consumer OS is WindowsCE.
>
> I don't agree. But you always seem to think that they can only have one of
> something. They have *several* products targeted simultaneously at the
> consumer market.
Can you count to ONE?
WebTV's later editions are to run WindowsCE. WindowsCE is also their Game
OS and PDA OS. ONE consumer OS.
MS has on OS for the desktop to the mainframe. It's called NT.
For the consumer the OS is windowsCE. ONE corporate OS.
Win98 is for the hobbiest losers who think the OS MS preloads on Wintel
PCs are MS's future. You got a OS that's sole purpose is to be upgraded
-- at your expense.
> >IBM's doing far more development work on OS/2 than MS is on Win98. Is MS
> >doing any more Win98 deelopment?
>
> Quite a lot, actually. They just released DirectX 6 for Win 98;
That's not win98 development. That's a stand alone technology for
WindwosCE game systems and NT.
And IE isn't win98 either. Not unless communicator/2 is OS/2. or IE on
HP-UX maens Hp_UX is now windows compatible.
> We're expecting the release of 128-bit-encrypted PPTP for Win 98 very
> shortly. (It's available now for Win 95 and Win NT, but for the moment Win
> 98 is limited to 40-bit encryption.)
Why did you poo-poo OS/2's TCP/IP updates?
> I *was* bothered (and said so) but I didn't have any alternatives. So I
> lived with it.
Predjudically, you had no alternatives.
> However, once Win 98 came out I bought and installed it so I could switch
> to FAT32. By doing that, I moved from 82% efficiency to 97% efficiency.
> That's good enough for me.
Given your Windows prejudices, a 0.001% improvement would be good for you.
> My point is that Bob was asking me whether I wanted OS/2's efficiency over
> the efficiency I now have with Win 98. No, it wouldn't be that big a deal
> at this point.
But you wanted OS/2 file system's efficiency. Your prejudices kept you on
Win32 but you wantted then efficiency so badly you bought Win98.
You also want OS/2's crash protection and OS/2 multi-tasking and RAM
efficiency. You put these exact same performance specs into windows99 and
you'd buy it in a snap.
Alas, there is to be no Windows99.
Steven C. Den Beste
Troughton) added this to the Deja-News archive:
>On Sat, 5 Sep 1998 20:00:38, sden...@san.rr.com (Steven C. Den Beste)
>wrote:
>
>> On Sat, 5 Sep 1998 12:17:31 -0700, josco <jo...@sea.monterey.edu> added
>> this to the Deja-News archive:
>>
>> >Your prejudices are driving your selection. You proved that when you were
>> >not bothered by VFAT's horrible waste of diskspace.
>>
>> I *was* bothered (and said so) but I didn't have any alternatives. So I
>> lived with it.
>>
>> However, once Win 98 came out I bought and installed it so I could switch
>> to FAT32. By doing that, I moved from 82% efficiency to 97% efficiency.
>> That's good enough for me.
>>
>> My point is that Bob was asking me whether I wanted OS/2's efficiency over
>> the efficiency I now have with Win 98. No, it wouldn't be that big a deal
>> at this point.
>>
>> Over VFAT, it would be a big deal. Over FAT32, it's small change. The gain
>> of 34M on a 2G drive is in the noise.
>>
>
>There is something else that comes with HPFS over vfat, fat32, or any
>other kind of fat file system; data integrity. If your file
>allocation tables get corrupted, you are out of luck. If some of the
>inodes in HPFS get corrupted, well, it won't affect any of the other
>ones, and thanks to the use of a B+ tree (that has pointers both up
>from the data to the file system as well as from the file system to
>the data) your chances of recovering from a bad crash are much much
>better.
Actually, when they created FAT32, they didn't just take VFAT and make it
bigger. They put in some other enhancements (like a backup FAT) to make it
less vulnerable to failure:
--------------
http://www.microsoft.com/windows/pr/fat32.htm
FAT32 provides the following enhancements over previous implementations of
the FAT file system:
More robust. FAT32 has the ability to relocate the root directory and use
the backup copy of the FAT instead of the default copy. In addition, the
boot record on FAT32 drives has been expanded to include a backup of
critical data structures. This means that FAT32 drives are less susceptible
to a single point of failure than existing FAT volumes.
--------------
Anyway, I keep backups. That's why I have 10 JAZ cartridges lying around.
Bob O
wrote:
I guess you have to be a "big" corporation then to use NT. CompUSA
caters to a lot of small corps.
>>>>>The simple fact is that there's nothing OS/2 provides above and beyond what
>>>>>Win 98 provides which I want.
>>>>
>>>>You don't want improved stability, better multitasking, capability to
>>>>handle more threads, more efficient disk space usage or safer and more
>>>>dependable file naming conventions?
>>>
>>>I find Win 98 sufficiently stable for my needs (the myths about crashing
>>>every five minutes notwithstanding, I find that Win 98 crashes about as
>>>often as Warp 3 did when I was using Warp 3 at work); I also find that
>>>FAT32 is sufficiently efficient for my disk space, and most of those other
>>>things don't affect the way I use my computer at home. The one exception is
>>>"file naming conventions", and I *prefer* the way Win 98 does it to the way
>>>OS/2 does it.
>>
>>Wow that is really bad!! Since you consider the occasional necessary
>>reboot that OS/2 required to clear screen holes and input blockage a
>>"crash" equivalent to the ugly way that Win95 dies taking data with it
>>that is really bad!!!
>
>I don't lose data when Win 95 crashes, you may be surprised to learn. I
>don't recall ever having done so. And as often as not, I had to reboot OS/2
>because it had locked up cold.
I haven't had OS/2 lock up cold since 1995. When OS/2 locks up
completely it has always displayed a trap screen. But I haven't seen
one of those in years. Other than having a bunch of traps show up with
my May 92 installation of OS/2 2.0 which went away as soon as I
installed a couple of patches released in June 92, I hardly ever saw the
trap screen afterwards.
If you are not losing data in Win98 when it crashes the reason is
because you are not multitasking. OS/2 when it gets the sort of PM
lockups you are talking about (these have also been non-existant for me
for about 6 months now since the last few fixpacks) do not interrupt
tasks going on in the background. The only thing locked is more user
input. What that means is that data being processed gets processed to
completion (assuming you don't get antsy and pull the plug).
>>>They would strongly affect the way I use a computer at work, but at work I
>>>use NT.
>>
>>I guess if you must have nothing important to do at home, eh?
>
>Depends on what you mean by "important". Critical? Nope. It's a hobby. How
>important is a hobby?
To me it is not an issue of "importance", it is an issue of enjoying
myself.
I suppose if your hobby is owning a computer system that crashes, you
had better enjoy yourself watching your screen go blank or you better
find another hobby.
My number one hobby is fishing. When my fishing line gets old and
frayed, I replace it before I start losing fish.
>Let's look at one of your advantages in greater detail: "more efficient
>disk space usage". Generally, since file sizes are random over a large
>number of files, a typical file wastes half a cluster.
>
>I just checked, and my 2G C: partition has 18602 files on it (counting
>directories).
>
>On a 2G partition, VFAT was using 32K clusters. Since an average file
>wasted half of that, 16K, and given 18602 files, then it meant that on VFAT
>I was wasting 290 megabytes of disk space.
>
>When I switched to FAT32, my cluster size changed to a much more reasonable
>4K. That means that a typical file is now wasting only 2K, yielding a total
>wastage of 36 megabytes.
>
>HPFS uses a 256 byte cluster, so the average wastage is 128 bytes. Were I
>to convert to HPFS, I would be wasting two and a quarter megabytes.
>
>But look at wastage as a percentage of space used, and you see that the
>difference between FAT32 and HPFS really isn't all that significant:
>
>Base data size: 1341 meg (not counting wastage)
>
>VFAT: wastage: 290M total: 1631M efficiency: 82%
>FAT32: wastage: 36M total: 1377M efficiency: 97%
>HPFS: wastage: 2M total: 1343M efficiency: 99.8%
>
>Yes, HPFS is more efficient. But the difference between 97% and 99.8%,
>between 36M wasted and 2M wasted out of a total of 1340M used, is not
>enough to bother me.
Your testing methodology is flawed.
I found a far better way to estimate the true disk space loss. First of
all you add up the number of files that are smaller than one cluster.
You take the total size of these files and subtract that from the
product of the number of files times the cluster size. Every time I
have done this test the amount of space wasted was always considerably
higher than half a sector per file because of lots of small files
smaller than a half sector.
Then you use the estimate of half a sector for all remaining files and
add the two together. I found that this figure produces a much more
accurate result and everytime I have done it it results in a much higher
wastage figure for large sector partitions.
Steven C. Den Beste
this to the Deja-News archive:
>On Sat, 5 Sep 1998, Steven C. Den Beste wrote:
>
>> On Sat, 5 Sep 1998 12:17:31 -0700, josco <jo...@sea.monterey.edu> added
>> this to the Deja-News archive:
>
>> >> I gather you haven't heard of a recently-released product called "Windows
>> >> 98"? It was targeted at home/hobby users.
>> >
>> >Oh that old dead end!
>>
>> At least Microsoft is being honest about it.
>
>What is it going to be in this post?
>
>Are you going to imply IBM is't telling OS/2 users the truth or are you
>going to post iBM's java plans for OS/2 and take IBM on it's word?
I don't think IBM is either telling the truth OR lying, because IBM isn't
saying anything at all. Silence is neither truth nor a lie.
IBM has not formally and officially released their plans, so no direct
reference to such plans can be included. All we can do is speculate, which
is to say to express our opinions.
My opinion is that IBM's efforts on OS/2 are now viewed by IBM management
as a transition strategy to get their large customers cleanly off of OS/2,
so that OS/2 can be shut down. Implementation of a JVM for OS/2 gives the
customers the ability to transition from pure OS/2 native executables to a
mix of native and Java executables, and ultimately to exclusively using
Java. My opinion is that once that's happened, those customers won't *need*
OS/2 anymore and IBM will try to convince them to deep-six their OS/2
machines and switch to JavaOS-based workstations.
My opinion is that WSOD is part of that transition strategy, and that it is
not considered a long-term product by IBM management. My opinion is that
WSOD is considered another way to transition the key customers (that is,
medium and large customers) away from OS/2-proper to the concept of
Java-based workstations.
My opinion is that IBM won't admit this publicly because of the uproar it
would cause.
These are opinions because I have no objective evidence for them.
It is a matter of fact, however, that Microsoft's plans for the Windows
line and the NT line were announced years ago, and that I knew that what
became Win 98 was going to be the end of the Windows product line before I
bought my first copy of Win 95.
It is a matter of fact that Microsoft has been honest with its customers
about this.
And it is a matter of fact that IBM has clearly stated that they don't care
about "small" users of OS/2, defined by IBM as anyone with less than 200
employees. That would obviously include home/hobby/SOHO, too. (They stated
it the other way: "The target market is large and medium enterprise
customers" and in answer to a question responded that a medium enterprise
was a company which employed 200 or more people.)
>
>> >MS's real home user computer is WebTV. Their consumer OS is WindowsCE.
>>
>> I don't agree. But you always seem to think that they can only have one of
>> something. They have *several* products targeted simultaneously at the
>> consumer market.
>
>Can you count to ONE?
Certainly. Can you count higher than that?
>WebTV's later editions are to run WindowsCE. WindowsCE is also their Game
>OS and PDA OS. ONE consumer OS.
That doesn't exclude Win 98. Nothing says Microsoft can't have more than
one product targeted at that market. You keep saying "ONE OS" but you've
never provided a rational explanation why Microsoft is limited to only one.
Care to offer one?
Why can't they have two or more? Is it a violation of the laws of physics?
>MS has on OS for the desktop to the mainframe. It's called NT.
>For the consumer the OS is windowsCE. ONE corporate OS.
Why only one? What limits them to only one? In fact, Win 95 is now making
significant inroads into the corporate sphere. (One reason is that NT works
exceedingly badly on portable computers. Win 95 is much better; so all
those high-powered execs who carry around WIN32-based notebook computers
are running Win 95 or perhaps Win 98 on them.)
>Win98 is for the hobbiest losers who think the OS MS preloads on Wintel
>PCs are MS's future. You got a OS that's sole purpose is to be upgraded
>-- at your expense.
See my recent exchange with Bob Osbourne on the subject of why I'm not only
willing but *eager* to pay for upgrades.
>> >IBM's doing far more development work on OS/2 than MS is on Win98. Is MS
>> >doing any more Win98 deelopment?
>>
>> Quite a lot, actually. They just released DirectX 6 for Win 98;
>
>That's not win98 development. That's a stand alone technology for
>WindwosCE game systems and NT.
I don't quite understand how you can describe it as NT technology when it
works on Win 98 but doesn't yet work on NT. For that matter, I don't
believe it's on CE yet, either.
Each of those target operating systems requires its own version of DirectX
6. The one for Win 98, developed specifically *for* Win 98, was just
released and I have downloaded and installed it. That sounds a great deal
to me like Win 98 development.
>And IE isn't win98 either. Not unless communicator/2 is OS/2. or IE on
>HP-UX maens Hp_UX is now windows compatible.
The Win 95 version of IE4 isn't the same as the Win 98 version of IE4. They
did a bugfix for each version. To me that counts as Win 98 development.
When IE5 comes out, the version which will work on Win 95 won't be the same
as the version which works on Win 98. (Lots in common, of course, but
critical differences too.) Microsoft will make sure IE5 works on both
platforms. That's Win 98 development.
>> We're expecting the release of 128-bit-encrypted PPTP for Win 98 very
>> shortly. (It's available now for Win 95 and Win NT, but for the moment Win
>> 98 is limited to 40-bit encryption.)
>
>Why did you poo-poo OS/2's TCP/IP updates?
I believe it was because OS/2 users were expecting to get it for free and
got angry when they discovered it required a membership in Software Choice.
I think it's fine that IBM charges for it, and were I an OS/2 user I'd have
a membership in Software Choice. But I recall being amused by the reaction
of OS/2 users about that prospect.
>> I *was* bothered (and said so) but I didn't have any alternatives. So I
>> lived with it.
>
>Predjudically, you had no alternatives.
>
>> However, once Win 98 came out I bought and installed it so I could switch
>> to FAT32. By doing that, I moved from 82% efficiency to 97% efficiency.
>> That's good enough for me.
>
>Given your Windows prejudices, a 0.001% improvement would be good for you.
No, for that I wouldn't have bothered. (Funny how you know so much more
about me than I do. How do you do that?)
>> My point is that Bob was asking me whether I wanted OS/2's efficiency over
>> the efficiency I now have with Win 98. No, it wouldn't be that big a deal
>> at this point.
>
>But you wanted OS/2 file system's efficiency. Your prejudices kept you on
>Win32 but you wantted then efficiency so badly you bought Win98.
No, I simply wanted efficiency better than VFAT. But now that I have FAT32,
the added increase HPFS would give me is not large enough to be attractive.
And it wasn't prejudices which kept me with WIN32, it was broad and easy
availability of applications. That's been my motivation from the beginning,
and as time has gone on it's only gotten stronger.
>You also want OS/2's crash protection and OS/2 multi-tasking and RAM
>efficiency. You put these exact same performance specs into windows99 and
>you'd buy it in a snap.
Amazing how you know what's in my mind better than I do. How do you do
that?
>Alas, there is to be no Windows99.
Fortunately, there will be a home version of NT. That's even better. It
will have crash protection -- better than OS/2. (NT does resource tracking
and will recover all resources used by a task which dies a horrible death.
OS/2 doesn't, and resources can fall on the floor and get lost. That's why
when Nav 2.02 dies, it sometimes leaves permanent trash on the screen. OS/2
doesn't know how to find it and scavenge it.)
It will have better multi-tasking.
I'm not impressed by RAM efficiency when RAM is as cheap as it is. My
computer has 128M now, and the one I'll be buying in about a month and a
half is going to have 256M.
And I do expect to be near the front of the line to buy it when it comes
out. Because it will give me all this without requiring me to discard and
replace all the software I own.
In the meantime, Win98 works quite well and I'm very pleased with it.
Steven C. Den Beste
the Deja-News archive:
>On Sat, 5 Sep 1998 17:44:35, sden...@san.rr.com (Steven C. Den Beste)
>wrote:
>
>>Yes, HPFS is more efficient. But the difference between 97% and 99.8%,
>>between 36M wasted and 2M wasted out of a total of 1340M used, is not
>>enough to bother me.
>
>Your testing methodology is flawed.
>
>I found a far better way to estimate the true disk space loss. First of
>all you add up the number of files that are smaller than one cluster.
>You take the total size of these files and subtract that from the
>product of the number of files times the cluster size. Every time I
>have done this test the amount of space wasted was always considerably
>higher than half a sector per file because of lots of small files
>smaller than a half sector.
>
>Then you use the estimate of half a sector for all remaining files and
>add the two together. I found that this figure produces a much more
>accurate result and everytime I have done it it results in a much higher
>wastage figure for large sector partitions.
>
>Bob O - Computing for fun
Bob, I don't particularly care to go make a census of eighteen thousand
files just to come up with a figure which will only be slightly more
precise than what I just described. The fact is that the change from a 4K
cluster to a 256 byte cluster will not save very much space.
If you don't like my numbers, fine.
By the way, Microsoft added something to FAT32 and Win 98 which is
interesting and nice which I bet OS/2 doesn't have.
There's a demon which sits in the background and watches every load request
on the system, and keeps records about which application requrested them.
This is logged to the disk.
After keeping this kind of data for a while and getting modal load patterns
down, the information is used by the defragger.
When a disk is defragged, the files or portions of files which tend to get
loaded at the same time are placed contiguously on the disk, and those
which are used more often are placed near the beginning of the disk.
The upshot is that over a period of weeks, after two or three defrags, a
Win 98 system's load time for applications improves quite dramatically. I
could tell the difference after the first time I defragged.
I'm not certain, but I don't think this is in even the latest version of
Win 95 (OSR2.5?). It's just a little thing, like multi-monitor support,
which got added and didn't get much publicity.
Robato Yao
>On Sat, 5 Sep 1998 14:50:00 -0700, josco <jo...@sea.monterey.edu> added
>this to the Deja-News archive:
>
>>On Sat, 5 Sep 1998, Steven C. Den Beste wrote:
>
>I don't quite understand how you can describe it as NT technology when it
>works on Win 98 but doesn't yet work on NT. For that matter, I don't
>believe it's on CE yet, either.
>
>Each of those target operating systems requires its own version of DirectX
>6. The one for Win 98, developed specifically *for* Win 98, was just
>released and I have downloaded and installed it. That sounds a great deal
>to me like Win 98 development.
I am not completely sure if the current DirectX6 is specifically for
Windows 98 only.
Robato Yao
Not a bad idea, if JavaOS can prove to be technologically better than
OS/2.
>
>My opinion is that WSOD is part of that transition strategy, and that it is
>not considered a long-term product by IBM management. My opinion is that
>WSOD is considered another way to transition the key customers (that is,
>medium and large customers) away from OS/2-proper to the concept of
>Java-based workstations.
WSOD is a cool idea for businesses.
>
>My opinion is that IBM won't admit this publicly because of the uproar it
>would cause.
I don't think so. I think they already made their intentions quite
clear already.
>
>These are opinions because I have no objective evidence for them.
>
>It is a matter of fact, however, that Microsoft's plans for the Windows
>line and the NT line were announced years ago, and that I knew that what
>became Win 98 was going to be the end of the Windows product line before I
>bought my first copy of Win 95.
I doubt that considering you have a bad history of making up excuses
for Microsoft before.
Microsoft left me with an impression that Windows 95 was only a fluke,
a temporary bridge to a consumer NT, which should have been where
Windows 95 was in the first place back in 1995. Three years later and
you still need a Windows 98. Three years later, an NT lite is nowwhere
in sight, DirectX games you still can't play on NT, and you have the
exact opposite of an NT lite, a monster corporate NT project going out
of control and beyond late.
>
>It is a matter of fact that Microsoft has been honest with its customers
>about this.
Hahahaha.
They sure didn't bugfix enough. On slow connections, IE4 is even less
responsive than Communicator, unable even to receive mouse input to the
command bar, as IE 4 keeps polling the connection. And when IE4 craps
out, it often takes the entire desktop with it.
Which is nowhere in sight.
>will have crash protection -- better than OS/2. (NT does resource tracking
>and will recover all resources used by a task which dies a horrible death.
>OS/2 doesn't, and resources can fall on the floor and get lost. That's why
>when Nav 2.02 dies, it sometimes leaves permanent trash on the screen. OS/2
>doesn't know how to find it and scavenge it.)
Wrong. The problem of Netscape on OS/2 is that it sometimes fails to
release shared resources, as normally killed applications in OS/2 does.
That has nothing to do with tracking, but more like
internally, Netscape for OS/2 can be a lot more integrated to the
operating system and the UI than it looks on the surface. For all the
problems Communicator beta still has, this is one area where
Communicator for OS/2 has it done over Netscape 2.02.
Also, as a matter of fact, I have both NT and Windows 98 unable to cope
with dead areas as well involving killed or hanged applications.
>
>It will have better multi-tasking.
Better multitasking? NT takes up more resource than OS/2 just to
multitask and multithread.
>
>I'm not impressed by RAM efficiency when RAM is as cheap as it is. My
>computer has 128M now, and the one I'll be buying in about a month and a
>half is going to have 256M.
>
>And I do expect to be near the front of the line to buy it when it comes
>out. Because it will give me all this without requiring me to discard and
>replace all the software I own.
>
>In the meantime, Win98 works quite well and I'm very pleased with it.
>
It works well, but it still hangs completely dead on some games.
David H. McCoy
> They sure didn't bugfix enough. On slow connections, IE4 is even less
> responsive than Communicator, unable even to receive mouse input to the
> command bar, as IE 4 keeps polling the connection. And when IE4 craps
> out, it often takes the entire desktop with it.
If this happens to you, it is because YOU haven't set IE4 to run in a
separate process. You do KNOW about this feature?
>
> >will have crash protection -- better than OS/2. (NT does resource tracking
> >and will recover all resources used by a task which dies a horrible death.
> >OS/2 doesn't, and resources can fall on the floor and get lost. That's why
> >when Nav 2.02 dies, it sometimes leaves permanent trash on the screen. OS/2
> >doesn't know how to find it and scavenge it.)
>
> Wrong. The problem of Netscape on OS/2 is that it sometimes fails to
> release shared resources, as normally killed applications in OS/2 does.
> That has nothing to do with tracking, but more like
> internally, Netscape for OS/2 can be a lot more integrated to the
> operating system and the UI than it looks on the surface. For all the
> problems Communicator beta still has, this is one area where
> Communicator for OS/2 has it done over Netscape 2.02.
If integration is important, IE4 has all beat. I've had it crash and not
does it NOT leave a hole, but it doesn't take the desktop with it.
Why?
Because I have it set to run in a separate process.
> Also, as a matter of fact, I have both NT and Windows 98 unable to cope
> with dead areas as well involving killed or hanged applications.
>
>
> >
> >It will have better multi-tasking.
>
> Better multitasking? NT takes up more resource than OS/2 just to
> multitask and multithread.
Based on what? Evidence, please.
> >
> >I'm not impressed by RAM efficiency when RAM is as cheap as it is. My
> >computer has 128M now, and the one I'll be buying in about a month and a
> >half is going to have 256M.
> >
> >And I do expect to be near the front of the line to buy it when it comes
> >out. Because it will give me all this without requiring me to discard and
> >replace all the software I own.
> >
> >In the meantime, Win98 works quite well and I'm very pleased with it.
> >
>
> It works well, but it still hangs completely dead on some games.
>
> Rgds,
>
> Chris
As does OS/2 on some applications and it is supposed to be the stronger,
more robust system.
--
-----------------------------------
David H. McCoy
dmccoy@REMOVE_MEmnsinc.com
-----------------------------------
David H. McCoy
>
> Bob, I don't particularly care to go make a census of eighteen thousand
> files just to come up with a figure which will only be slightly more
> precise than what I just described. The fact is that the change from a 4K
> cluster to a 256 byte cluster will not save very much space.
>
> If you don't like my numbers, fine.
>
Bob has time to perform such a count because there are fewer applications,
and thus, fewer files to track.
joseph
>
> On Sat, 5 Sep 1998 14:50:00 -0700, josco <jo...@sea.monterey.edu> added
> >> At least Microsoft is being honest about it.
> >
> >What is it going to be in this post?
> >
> >Are you going to imply IBM is't telling OS/2 users the truth or are you
> >going to post iBM's java plans for OS/2 and take IBM on it's word?
>
> I don't think IBM is either telling the truth OR lying, because IBM isn't
> saying anything at all. Silence is neither truth nor a lie.
In your post you made several claims that are, in my opinion,
deliberately untruthful.
The long history of IBM's OS/2 plans are archived on the OS/2 WARP FM
site. These and other IBM news releases are sent to my mail box and to
anyone else who elects to subscribes to the lists.
They have been releasing this list bi-monthly for the past 4 months, but
monthly prior to then. They oldest archives go back to Vol 2 #30. They
began with Vol 1 #1.
http://www.software.ibm.com/os/warp/warpfm/
Ironically, many of the topics you claim to have thought up in the
absence of IBM's guidence (indeed slience) are on the WARP FM pages.
I've read from interviews with IBM execs that IBM is seriously
considering rebranding OS/2 -- possibly with the AURORA upgrade. Names
are names. Code is the same.
Someday OS/2 will be retired, someday the sun will engulf the Earth.
Before both of these events, MS will abandon Windows98. Many of today's
Win3.1 and win95 users will be migrated to NT Terminal Server. FedEx is
leading the way -- it's all in the news if you cared to be truthful.
It's a fact that Win98 is a dead end OS. You'll find no plans for MS to
improve it or reuse the code base. No plans to do anything but walk
away from this DOS hybrid.
MS's WindowsCE is a new consumer OS for WebTV and TCI cable boxes, WinCE
is also for game console users and for PDA users. The ONLY other OS is
NT 5.0 and some yet to be shown NT Lite for the consumer.
It is clear that MS has a serious problem with the NT 5.0 project.
Numerous reposrts say it is sucking up all their programmers and so for
the meanwhile win98 will be MS's volume OS. That does not mean the OS
has any stragegic significance to MS aside from being a outdated place
holder in thier product line.
I would NOT be surpised if MS did not LOSE the entire win98 code base in
the next few years.
joseph
>
> this to the Deja-News archive:
>
> >>
> >>It is a matter of fact, however, that Microsoft's plans for the Windows
> >>line and the NT line were announced years ago, and that I knew that what
> >>became Win 98 was going to be the end of the Windows product line before I
> >>bought my first copy of Win 95.
> >
> >I doubt that considering you have a bad history of making up excuses
> >for Microsoft before.
> >
> >Microsoft left me with an impression that Windows 95 was only a fluke,
> >a temporary bridge to a consumer NT, which should have been where
> >Windows 95 was in the first place back in 1995. Three years later and
> >you still need a Windows 98. Three years later, an NT lite is nowwhere
> >in sight, DirectX games you still can't play on NT, and you have the
> >exact opposite of an NT lite, a monster corporate NT project going out
> >of control and beyond late.
>
> which stated clearly that there would only be one more major release after
> Win 95 in the "Windows" line, and after that it would go into maintenance
> mode. While that was going on, NT would move toward a convergence such that
> 2-3 years after the release of what we now call Win 98 there would be a
> version of NT which would represent a clean migration path for the Win
> 95/98 users.
>
> I read that a couple of weeks before I purchased my first copy of Win 95.
>
> "NT Lite" (your term) was never expected to be ready now and was never
> promised to be ready now.
Chris is quite correct.
Win98 is itself an expensive Win95 fixpack with IE integrated to fend
off anti-trust problems. Its value is questioned by the windows mags
that write about positive stuff about windows.
MS is well behind development promises and deadlines they outlined in
1995. NT should be done and NT lite well long in it's development.
Win98 is a dead end and it is long in the tooth. The migration from
Win98 to NT 5.0 is not to be clean or smooth. MS is now saying it will
be necessary to reformat the drive of a win98 system to install NT 5.0.
MS is trying to shoe horn WindowsCE into a lot of niches. WebTV/TCI
cableboxes will use WinCE. Togehter MS paid 1 billion to buy WebTV and
to pay TCI to let MS ship WinCE to run personal java software on TCI
cable boxes. The Sega DreamCast will run windowsCE (and Sega's own OS
for intense games), some PDAs run windowsCE and new Jupiter class PCs
will run WindowsCE.
Sad;y, MS 's anti-trust problems make it difficult if not impossible for
MS to get away with a purchase of a 3rd party OS, like the BeOS or some
UNIX like SCO to save their ass.
They might come crawling back to IBM and try to resell MS OS/2. It
works and all you guys say it needs is MS's marketing. MS OS/2, the
follwo-on to NT and WIN98 -- it's better faster, and cheaper. Sure
beats NT 2001.
Todd
wrote:
>On Sat, 5 Sep 1998 14:50:00 -0700, josco wrote:
>
>>>I don't agree. But you always seem to think that they can only have one of
>>>something. They have *several* products targeted simultaneously at the
>>>consumer market.
>>
>>Can you count to ONE?
>>
>>WebTV's later editions are to run WindowsCE. WindowsCE is also their Game
>>OS and PDA OS. ONE consumer OS.
>>
>>MS has on OS for the desktop to the mainframe. It's called NT.
>>For the consumer the OS is windowsCE. ONE corporate OS.
>>
>>Win98 is for the hobbiest losers who think the OS MS preloads on Wintel
>>PCs are MS's future. You got a OS that's sole purpose is to be upgraded
>>-- at your expense.
>
>Here we go again. Joseph, why can't you just debate Steven without reverting
>to ridicule and an "of course I'm right, you're just a loser for trying to
>prove otherwise" attitude? Is this how you bluffed your way to a Ph.D.?
My thoughts exactly.
Joseph is continually FUDding NT, and the latest MS product. The
latest version of 95 (now 98) is always the 'dead end' OS.
Funny how MS keeps improving their OS, while no OS/2 upgrade is in
sight.
NT 4.0 is selling like mad *now*, and comes preinstalled on most, if
not all, medium-high end servers. This just wasn't the case a year
ago even.
Customers are demanding NT where just two years ago, Novell and OS/2
still had a shot.
Joseph FUDs CE too. It's supposedly for consumers. It's for
consumers with laptops and PDA's. 95/98 is for consumers with
desktops and/or laptops. But Joseph (even with a PHD) can't seem to
understand this... why? Because it is just FUD on his part.
OS/2 is simply dead. If there is any OS that is challenging NT, it is
UNIX/Linux. Those are the only robust server OSes out there that can
compete with NT server. Sun's server line is excellent. HP's UNIX
boxes are great too.
OS/2? Novell? Dead and soon to be dead.
NTW is the best desktop OS there is, hands down. OS/2? Nice
interface, horrible kernel architecture and design.
Oh well. When the next version of 9x comes out, Joseph will be
FUDding it as the 'last' MS consumer OS from the 9x architecture...
Hehe. It's always fun to see Joseph squirm.
With all of the FUDding and lack of knowledge, it sure doesn't seem to
befit a PHD...
-Todd
Todd
>On Sat, 5 Sep 1998 23:12:37, sden...@san.rr.com (Steven C. Den Beste)
>wrote:
>
>>On 5 Sep 1998 22:06:18 GMT, osb...@deletemeibm.net (Bob O) added this to
>>the Deja-News archive:
>>
>>>On Sat, 5 Sep 1998 17:44:35, sden...@san.rr.com (Steven C. Den Beste)
>>>wrote:
>>>
>>>>Yes, HPFS is more efficient. But the difference between 97% and 99.8%,
>>>>between 36M wasted and 2M wasted out of a total of 1340M used, is not
>>>>enough to bother me.
>>>
>>>Your testing methodology is flawed.
>>>
>>>I found a far better way to estimate the true disk space loss. First of
>>>all you add up the number of files that are smaller than one cluster.
>>>You take the total size of these files and subtract that from the
>>>product of the number of files times the cluster size. Every time I
>>>have done this test the amount of space wasted was always considerably
>>>higher than half a sector per file because of lots of small files
>>>smaller than a half sector.
>>>
>>>Then you use the estimate of half a sector for all remaining files and
>>>add the two together. I found that this figure produces a much more
>>>accurate result and everytime I have done it it results in a much higher
>>>wastage figure for large sector partitions.
>>>
>>>Bob O - Computing for fun
>>
>>Bob, I don't particularly care to go make a census of eighteen thousand
>>files just to come up with a figure which will only be slightly more
>>precise than what I just described. The fact is that the change from a 4K
>>cluster to a 256 byte cluster will not save very much space.
>>
>>If you don't like my numbers, fine.
>
>Isn't 36 megabytes about enough to install your operating system?
36 megabytes in the world of 13 *Gig* disks is *meaningless*.
WHO CARES
36 megabytes 5 years ago was a big deal. But this is 1998.
-Todd
Todd
wrote:
>In <35f3b2ba.22937490@news-server>, sden...@san.rr.com (Steven C. Den Beste) writes:
>>On Sat, 5 Sep 1998 14:50:00 -0700, josco <jo...@sea.monterey.edu> added
>>this to the Deja-News archive:
>>
>>>On Sat, 5 Sep 1998, Steven C. Den Beste wrote:
>>
>>I don't quite understand how you can describe it as NT technology when it
>>works on Win 98 but doesn't yet work on NT. For that matter, I don't
>>believe it's on CE yet, either.
>>
>>Each of those target operating systems requires its own version of DirectX
>>6. The one for Win 98, developed specifically *for* Win 98, was just
>>released and I have downloaded and installed it. That sounds a great deal
>>to me like Win 98 development.
>
>I am not completely sure if the current DirectX6 is specifically for
>Windows 98 only.
I don't think it is only for 98. It is in NT 5.0 as well.
But, most games still only use mostly directx 2.0/3.0 functionality
anyway. Very few games even use DX 5.0.
Heck, many DX 5.0 games run under NT 4.0. This should be impossible
since NT has only DX 3.0 available for it.
You can get DX 5.0 for NT 4.0 if you need it though...
But, the most killer games run on NT 4.0 right now, like Quake II,
Hexen II, Unreal (the most ultimate game ever) , StarCraft, etc.
Note that these are some of the latest games you can buy... most
developers aren't touching DX 5.0/6.0 until more people get the
version pre-installed.
All of the above run with 3D acceleration under NT using the mini
OpenGL 3D drivers.
It's pretty fun... think I'll play a little Quake II CTF right now!
Bye
-Todd
Gerben Bergman
Steven C. Den Beste
this to the Deja-News archive:
>In <35f3b2ba.22937490@news-server>, sden...@san.rr.com (Steven C. Den Beste) writes:
>>
>>It is a matter of fact, however, that Microsoft's plans for the Windows
>>line and the NT line were announced years ago, and that I knew that what
>>became Win 98 was going to be the end of the Windows product line before I
>>bought my first copy of Win 95.
>
>I doubt that considering you have a bad history of making up excuses
>for Microsoft before.
>
>Microsoft left me with an impression that Windows 95 was only a fluke,
>a temporary bridge to a consumer NT, which should have been where
>Windows 95 was in the first place back in 1995. Three years later and
>you still need a Windows 98. Three years later, an NT lite is nowwhere
>in sight, DirectX games you still can't play on NT, and you have the
>exact opposite of an NT lite, a monster corporate NT project going out
>of control and beyond late.
Microsoft announced the product planning just before the Win 95 launch,
which stated clearly that there would only be one more major release after
Win 95 in the "Windows" line, and after that it would go into maintenance
mode. While that was going on, NT would move toward a convergence such that
2-3 years after the release of what we now call Win 98 there would be a
version of NT which would represent a clean migration path for the Win
95/98 users.
I read that a couple of weeks before I purchased my first copy of Win 95.
"NT Lite" (your term) was never expected to be ready now and was never
promised to be ready now.
>
>>
>>It is a matter of fact that Microsoft has been honest with its customers
>>about this.
>
>Hahahaha.
Sigh.
>>
>>Fortunately, there will be a home version of NT. That's even better. It
>
>Which is nowhere in sight.
It was never intended to be in sight now. It was expected to come out 2-3
years after what we now call Win 98.
>
>>will have crash protection -- better than OS/2. (NT does resource tracking
>>and will recover all resources used by a task which dies a horrible death.
>>OS/2 doesn't, and resources can fall on the floor and get lost. That's why
>>when Nav 2.02 dies, it sometimes leaves permanent trash on the screen. OS/2
>>doesn't know how to find it and scavenge it.)
>
>Wrong. The problem of Netscape on OS/2 is that it sometimes fails to
>release shared resources, as normally killed applications in OS/2 does.
>That has nothing to do with tracking, but more like
>internally, Netscape for OS/2 can be a lot more integrated to the
>operating system and the UI than it looks on the surface. For all the
>problems Communicator beta still has, this is one area where
>Communicator for OS/2 has it done over Netscape 2.02.
>
>Also, as a matter of fact, I have both NT and Windows 98 unable to cope
>with dead areas as well involving killed or hanged applications.
Every time any application asks for any resource, shared or not, NT keeps
track. Associated with the task control block for the task, NT keeps a list
of all the things that task is using.
When a task dies, it is supposed to be a good citizen and give them back.
But if it doesn't, NT will clean up after it.
When an OS/2 task dies, it is also supposed to be a good citizen and give
such resources back. But if it doesn't do so, OS/2 can't clean up because
OS/2 didn't keep track in the first place.
If Netscape was supposed to release shared resources, OS/2 should have been
able to find them and recover them when Netscape died. NT can.
I have never seen lost resources on NT. Win 95/98 are a different matter,
because Win 95/98 do not track resources any more than OS/2 does.
>
>>
>>It will have better multi-tasking.
>
>Better multitasking? NT takes up more resource than OS/2 just to
>multitask and multithread.
It will have better multitasking than Win 98. I thought it was clear from
context that that was what I meant.
>>
>>I'm not impressed by RAM efficiency when RAM is as cheap as it is. My
>>computer has 128M now, and the one I'll be buying in about a month and a
>>half is going to have 256M.
>>
>>And I do expect to be near the front of the line to buy it when it comes
>>out. Because it will give me all this without requiring me to discard and
>>replace all the software I own.
>>
>>In the meantime, Win98 works quite well and I'm very pleased with it.
>>
>
>It works well, but it still hangs completely dead on some games.
>
>Rgds,
>
>Chris
Why don't you play those games on OS/2 instead, then? After all, OS/2 is
more stable.
Steven C. Den Beste
this to the Deja-News archive:
>In <35f3b2ba.22937490@news-server>, sden...@san.rr.com (Steven C. Den Beste) writes:
>>On Sat, 5 Sep 1998 14:50:00 -0700, josco <jo...@sea.monterey.edu> added
>>this to the Deja-News archive:
>>
>>>On Sat, 5 Sep 1998, Steven C. Den Beste wrote:
>>
>>I don't quite understand how you can describe it as NT technology when it
>>works on Win 98 but doesn't yet work on NT. For that matter, I don't
>>believe it's on CE yet, either.
>>
>>Each of those target operating systems requires its own version of DirectX
>>6. The one for Win 98, developed specifically *for* Win 98, was just
>>released and I have downloaded and installed it. That sounds a great deal
>>to me like Win 98 development.
>
>I am not completely sure if the current DirectX6 is specifically for
>Windows 98 only.
>
>Rgds,
>
>Chris
There are two, one of which was for Win 95 and one of which was for Win 98.
None exists at this time for NT.
The Win 95 version was released earlier.
Bob O
wrote:
>On 5 Sep 1998 22:06:18 GMT, osb...@deletemeibm.net (Bob O) added this to
>the Deja-News archive:
>
>>On Sat, 5 Sep 1998 17:44:35, sden...@san.rr.com (Steven C. Den Beste)
>>wrote:
>>
>>>Yes, HPFS is more efficient. But the difference between 97% and 99.8%,
>>>between 36M wasted and 2M wasted out of a total of 1340M used, is not
>>>enough to bother me.
>>
>>Your testing methodology is flawed.
>>
>>I found a far better way to estimate the true disk space loss. First of
>>all you add up the number of files that are smaller than one cluster.
>>You take the total size of these files and subtract that from the
>>product of the number of files times the cluster size. Every time I
>>have done this test the amount of space wasted was always considerably
>>higher than half a sector per file because of lots of small files
>>smaller than a half sector.
>>
>>Then you use the estimate of half a sector for all remaining files and
>>add the two together. I found that this figure produces a much more
>>accurate result and everytime I have done it it results in a much higher
>>wastage figure for large sector partitions.
>>
>>Bob O - Computing for fun
>
>Bob, I don't particularly care to go make a census of eighteen thousand
>files just to come up with a figure which will only be slightly more
>precise than what I just described. The fact is that the change from a 4K
>cluster to a 256 byte cluster will not save very much space.
>
>If you don't like my numbers, fine.
Isn't 36 megabytes about enough to install your operating system?
Bob Hauck
sden...@san.rr.com (Steven C. Den Beste) writes:
> The upshot is that over a period of weeks, after two or three
> defrags,
You people are still defragging disks? What is this, the stone
age or something?
--
No proprietary software was used in the creation of this message.
David H. McCoy
> In article <35f2c3a5.27269538@news-server>,
> sden...@san.rr.com (Steven C. Den Beste) writes:
>
> > The upshot is that over a period of weeks, after two or three
> > defrags,
>
> You people are still defragging disks? What is this, the stone
> age or something?
>
>
Sure, HPFS doesn't have any defraggers. Just ask the makers of the Graham
and Gammatech utilities.
Why doesn't HPFS offer compression, security, journaling, volume striping,
user selectable cluster sizes or 8.3 file name generation for all those
16-bit Windows programs that some insist are still being made?
Is this the stone-age or something?
Gerben Bergman
>>Here we go again. Joseph, why can't you just debate Steven without reverting
>>to ridicule and an "of course I'm right, you're just a loser for trying to
>>prove otherwise" attitude? Is this how you bluffed your way to a Ph.D.?
>
>My thoughts exactly.
>
>Joseph is continually FUDding NT, and the latest MS product. The
>latest version of 95 (now 98) is always the 'dead end' OS.
[...]
>With all of the FUDding and lack of knowledge, it sure doesn't seem to
>befit a PHD...
I wasn't talking about Joseph's so-called FUDding of MS products, Todd. He's
merely presenting his point of view and if you think he's wrong, then by all
means correct him. That's the whole point of debating.
What I *was* commenting on is his inability to keep a discussion civilized;
his tendency to get personal when the other party won't easily give in.
That's the thing about him which irritates me.
Bob O
>On 6 Sep 1998 03:11:30 GMT, osb...@deletemeibm.net (Bob O) wrote:
>
>>On Sat, 5 Sep 1998 23:12:37, sden...@san.rr.com (Steven C. Den Beste)
>>wrote:
>>
>>>On 5 Sep 1998 22:06:18 GMT, osb...@deletemeibm.net (Bob O) added this to
>>>the Deja-News archive:
>>>
>>>>On Sat, 5 Sep 1998 17:44:35, sden...@san.rr.com (Steven C. Den Beste)
>>>>wrote:
>>>>
>>>>>Yes, HPFS is more efficient. But the difference between 97% and 99.8%,
>>>>>between 36M wasted and 2M wasted out of a total of 1340M used, is not
>>>>>enough to bother me.
>>>>
>>>>Your testing methodology is flawed.
>>>>
>>>>I found a far better way to estimate the true disk space loss. First of
>>>>all you add up the number of files that are smaller than one cluster.
>>>>You take the total size of these files and subtract that from the
>>>>product of the number of files times the cluster size. Every time I
>>>>have done this test the amount of space wasted was always considerably
>>>>higher than half a sector per file because of lots of small files
>>>>smaller than a half sector.
>>>>
>>>>Then you use the estimate of half a sector for all remaining files and
>>>>add the two together. I found that this figure produces a much more
>>>>accurate result and everytime I have done it it results in a much higher
>>>>wastage figure for large sector partitions.
>>>>
>>>>Bob O - Computing for fun
>>>
>>>Bob, I don't particularly care to go make a census of eighteen thousand
>>>files just to come up with a figure which will only be slightly more
>>>precise than what I just described. The fact is that the change from a 4K
>>>cluster to a 256 byte cluster will not save very much space.
>>>
>>>If you don't like my numbers, fine.
>>
>>Isn't 36 megabytes about enough to install your operating system?
>
>36 megabytes in the world of 13 *Gig* disks is *meaningless*.
>
>WHO CARES
>
>36 megabytes 5 years ago was a big deal. But this is 1998.
Welcome to the thread. We were talking about 2 gigs, not 13 gigs. In
your example it would be over 250 megs lost space, enough for an NT
installation even.
Just thought I would help you get caught up.
Steven C. Den Beste
the Deja-News archive:
>
>They might come crawling back to IBM and try to resell MS OS/2. It
>works and all you guys say it needs is MS's marketing. MS OS/2, the
>follwo-on to NT and WIN98 -- it's better faster, and cheaper. Sure
>beats NT 2001.
Fat chance of this. It doesn't run WIN32 software.
Steven C. Den Beste
the Deja-News archive:
>On 5 Sep 1998 23:56:42 GMT, cro...@kuentos.guam.net (Robato Yao)
>wrote:
>
>>In <35f3b2ba.22937490@news-server>, sden...@san.rr.com (Steven C. Den Beste) writes:
>>>On Sat, 5 Sep 1998 14:50:00 -0700, josco <jo...@sea.monterey.edu> added
>>>this to the Deja-News archive:
>>>
>>>>On Sat, 5 Sep 1998, Steven C. Den Beste wrote:
>>>
>>>I don't quite understand how you can describe it as NT technology when it
>>>works on Win 98 but doesn't yet work on NT. For that matter, I don't
>>>believe it's on CE yet, either.
>>>
>>>Each of those target operating systems requires its own version of DirectX
>>>6. The one for Win 98, developed specifically *for* Win 98, was just
>>>released and I have downloaded and installed it. That sounds a great deal
>>>to me like Win 98 development.
>>
>>I am not completely sure if the current DirectX6 is specifically for
>>Windows 98 only.
>
>I don't think it is only for 98. It is in NT 5.0 as well.
Win 98 will run as part of all three operating systems. But it has to be
customized to each of them. The customized version for Win 95 was released
in July. The customized version for Win 98 was released a couple of weeks
ago. The customized version for NT won't be available for quite some time.
>But, most games still only use mostly directx 2.0/3.0 functionality
>anyway. Very few games even use DX 5.0.
Actually, that's not correct. If you look at new releases. almost all of
them come with DX5 on them and use it. In most cases game developers have
been designing for the tip version of DirectX. Even if they don't use the
features of the new releases, each new release of DirectX has included more
drivers and has included speed increases for the older capabilities.
>Heck, many DX 5.0 games run under NT 4.0. This should be impossible
>since NT has only DX 3.0 available for it.
The reason is that in many cases the games would like to take advantage of
the speedups in DirectX 5 for functions available in DirectX 3. But if
they're not there, the program will still run -- but not as fast.
>You can get DX 5.0 for NT 4.0 if you need it though...
>
>But, the most killer games run on NT 4.0 right now, like Quake II,
>Hexen II, Unreal (the most ultimate game ever) , StarCraft, etc.
>
>Note that these are some of the latest games you can buy... most
>developers aren't touching DX 5.0/6.0 until more people get the
>version pre-installed.
DirectX is routinely included on the install disk with various games.
Steven C. Den Beste
Deja-News archive:
>In article <35f2c3a5.27269538@news-server>,
> sden...@san.rr.com (Steven C. Den Beste) writes:
>
>> The upshot is that over a period of weeks, after two or three
>> defrags,
>
>You people are still defragging disks? What is this, the stone
>age or something?
Oh, it isn't really necessary in the sense of the disk becoming really
fragmented; the last couple of times I defragged it was precisely to take
advantage of this feature.
Each time I defrag, the programs I use most commonly are optimized to load
as fast as possible.
Does OS/2 do this?
Bob O
>On Sun, 06 Sep 1998 00:30:15 +0200, rer...@wxs.nl (Gerben Bergman)
>wrote:
>
>>On Sat, 5 Sep 1998 14:50:00 -0700, josco wrote:
>>
>>>>I don't agree. But you always seem to think that they can only have one of
>>>>something. They have *several* products targeted simultaneously at the
>>>>consumer market.
>>>
>>>Can you count to ONE?
>>>
>>>WebTV's later editions are to run WindowsCE. WindowsCE is also their Game
>>>OS and PDA OS. ONE consumer OS.
>>>
>>>MS has on OS for the desktop to the mainframe. It's called NT.
>>>For the consumer the OS is windowsCE. ONE corporate OS.
>>>
>>>Win98 is for the hobbiest losers who think the OS MS preloads on Wintel
>>>PCs are MS's future. You got a OS that's sole purpose is to be upgraded
>>>-- at your expense.
>>
>>Here we go again. Joseph, why can't you just debate Steven without reverting
>>to ridicule and an "of course I'm right, you're just a loser for trying to
>>prove otherwise" attitude? Is this how you bluffed your way to a Ph.D.?
>
>My thoughts exactly.
>
>Joseph is continually FUDding NT, and the latest MS product. The
>latest version of 95 (now 98) is always the 'dead end' OS.
>
>Funny how MS keeps improving their OS, while no OS/2 upgrade is in
>sight.
>
>NT 4.0 is selling like mad *now*, and comes preinstalled on most, if
>not all, medium-high end servers. This just wasn't the case a year
>ago even.
CompUSA did not have one machine with NT on it when I was in there the
day before yesterday. Seems the NT fad has come and gone.
>Customers are demanding NT where just two years ago, Novell and OS/2
>still had a shot.
Both OS/2 and Novell had more shelfspace at CompUSA than NT. You can
make these claims but it seems the leading retailers don't hear you.
>Joseph FUDs CE too. It's supposedly for consumers. It's for
>consumers with laptops and PDA's. 95/98 is for consumers with
>desktops and/or laptops. But Joseph (even with a PHD) can't seem to
>understand this... why? Because it is just FUD on his part.
>
>OS/2 is simply dead. If there is any OS that is challenging NT, it is
>UNIX/Linux. Those are the only robust server OSes out there that can
>compete with NT server. Sun's server line is excellent. HP's UNIX
>boxes are great too.
>
>OS/2? Novell? Dead and soon to be dead.
>
>NTW is the best desktop OS there is, hands down. OS/2? Nice
>interface, horrible kernel architecture and design.
Well NT is certainly better than Win98. I would suspect that if MS
didn't have to deal with a ton of incompatibility issues they would have
never issued Win98 and pushed NT 4 for the home/soho user instead.
Must be a real rats nest! Enough to make MS fear the competition would
step in and steal the day.
Bob O
wrote:
>In article <6ssir8$ts5$1...@brokaw.wa.com>, cro...@kuentos.guam.net says...
>> They sure didn't bugfix enough. On slow connections, IE4 is even less
>> responsive than Communicator, unable even to receive mouse input to the
>> command bar, as IE 4 keeps polling the connection. And when IE4 craps
>> out, it often takes the entire desktop with it.
>
>If this happens to you, it is because YOU haven't set IE4 to run in a
>separate process. You do KNOW about this feature?
>
>>
>> >will have crash protection -- better than OS/2. (NT does resource tracking
>> >and will recover all resources used by a task which dies a horrible death.
>> >OS/2 doesn't, and resources can fall on the floor and get lost. That's why
>> >when Nav 2.02 dies, it sometimes leaves permanent trash on the screen. OS/2
>> >doesn't know how to find it and scavenge it.)
>>
>> Wrong. The problem of Netscape on OS/2 is that it sometimes fails to
>> release shared resources, as normally killed applications in OS/2 does.
>> That has nothing to do with tracking, but more like
>> internally, Netscape for OS/2 can be a lot more integrated to the
>> operating system and the UI than it looks on the surface. For all the
>> problems Communicator beta still has, this is one area where
>> Communicator for OS/2 has it done over Netscape 2.02.
>
>If integration is important, IE4 has all beat. I've had it crash and not
>does it NOT leave a hole, but it doesn't take the desktop with it.
>Why?
>Because I have it set to run in a separate process.
>
>> Also, as a matter of fact, I have both NT and Windows 98 unable to cope
>> with dead areas as well involving killed or hanged applications.
>>
>>
>> >
>> >It will have better multi-tasking.
>>
>> Better multitasking? NT takes up more resource than OS/2 just to
>> multitask and multithread.
>
>Based on what? Evidence, please.
>
>> >
>> >I'm not impressed by RAM efficiency when RAM is as cheap as it is. My
>> >computer has 128M now, and the one I'll be buying in about a month and a
>> >half is going to have 256M.
>> >
>> >And I do expect to be near the front of the line to buy it when it comes
>> >out. Because it will give me all this without requiring me to discard and
>> >replace all the software I own.
>> >
>> >In the meantime, Win98 works quite well and I'm very pleased with it.
>> >
>>
>> It works well, but it still hangs completely dead on some games.
>>
>> Rgds,
>>
>> Chris
>
>As does OS/2 on some applications and it is supposed to be the stronger,
>more robust system.
Based on what? Evidence, please.
Bob O - Computing for fun
Bob O
wrote:
>In article <35f2c3a5.27269538@news-server>, sden...@san.rr.com says...
>>
>> Bob, I don't particularly care to go make a census of eighteen thousand
>> files just to come up with a figure which will only be slightly more
>> precise than what I just described. The fact is that the change from a 4K
>> cluster to a 256 byte cluster will not save very much space.
>>
>> If you don't like my numbers, fine.
>>
>
>Bob has time to perform such a count because there are fewer applications,
>and thus, fewer files to track.
Heeheehee! I happened to have over 30,000 files. So I guess I both
need more applications than
Steven does and they are available.
And as far as making a count, I have an application that makes the count
take less than a minute. Ztree Bold. I simply log a disk, sort by
size, tag the files of the appropriate size and look at the log window.
I realize that with Microsoft software it probably takes a week to do
the job, but it isn't a big job here.
Steven C. Den Beste
the Deja-News archive:
>On Sat, 5 Sep 1998 23:30:35, to...@tkpowers.com (Todd) wrote:
>
>>NT 4.0 is selling like mad *now*, and comes preinstalled on most, if
>>not all, medium-high end servers. This just wasn't the case a year
>>ago even.
>
>CompUSA did not have one machine with NT on it when I was in there the
>day before yesterday. Seems the NT fad has come and gone.
>
>>Customers are demanding NT where just two years ago, Novell and OS/2
>>still had a shot.
>
>Both OS/2 and Novell had more shelfspace at CompUSA than NT. You can
>make these claims but it seems the leading retailers don't hear you.
Bob, from *one* visit to *one* CompUSA store you are extrapolating an
industry trend?
It's time for a brush-up course on statistics, with especial emphasis on
valid sample sizes.
Steven C. Den Beste
the Deja-News archive:
>On Sun, 6 Sep 1998 02:15:42, dmccoy@REMOVE_MEmnsinc.com (David H. McCoy)
>wrote:
>
>>In article <35f2c3a5.27269538@news-server>, sden...@san.rr.com says...
>>>
>>> Bob, I don't particularly care to go make a census of eighteen thousand
>>> files just to come up with a figure which will only be slightly more
>>> precise than what I just described. The fact is that the change from a 4K
>>> cluster to a 256 byte cluster will not save very much space.
>>>
>>> If you don't like my numbers, fine.
>>>
>>
>>Bob has time to perform such a count because there are fewer applications,
>>and thus, fewer files to track.
>
>Heeheehee! I happened to have over 30,000 files. So I guess I both
>need more applications than
> Steven does and they are available.
I guess I didn't mention that I have five other 2G partitions besides that
C: partition I mentioned.
C: 18602 files.
D: 4112 files
E: 24821 files
F: 18968 files
G: 4948 files
H: 1281 files
Not to mention all the files on the 10 1G JAZ cartridges I have, nor the
files I've rolled out onto CDs using my CDR drive.
...you were saying?
Bob O
wrote:
>On 6 Sep 1998 16:17:44 GMT, osb...@deletemeibm.net (Bob O) added this to
>the Deja-News archive:
>
>>On Sat, 5 Sep 1998 23:30:35, to...@tkpowers.com (Todd) wrote:
>>
>>>NT 4.0 is selling like mad *now*, and comes preinstalled on most, if
>>>not all, medium-high end servers. This just wasn't the case a year
>>>ago even.
>>
>>CompUSA did not have one machine with NT on it when I was in there the
>>day before yesterday. Seems the NT fad has come and gone.
>>
>>>Customers are demanding NT where just two years ago, Novell and OS/2
>>>still had a shot.
>>
>>Both OS/2 and Novell had more shelfspace at CompUSA than NT. You can
>>make these claims but it seems the leading retailers don't hear you.
>
>Bob, from *one* visit to *one* CompUSA store you are extrapolating an
>industry trend?
>
>It's time for a brush-up course on statistics, with especial emphasis on
>valid sample sizes.
Actually the 40 different or so brands and models carried by CompUSA
make up a statistically valid sample in their own right. However, any
statistical study operates on only a confidence factor. So perhaps you
could fill us in on the major retailers in your area that do carry NT
preloaded options and we can then extrapolate further.
Or is this one of those twisted posts of yours like the one earlier this
week where you indicated a tendency about preload licenses in general in
order to provide a defense of Microsoft licensiing practices when in
fact it appears the only examples of preload licenses you apparently had
were Microsoft licenses.
Still waiting on your reply to those other licenses and you can add to
it any major retailers in your area still pushing NT.
Bob O
wrote:
>On 6 Sep 1998 16:17:52 GMT, osb...@deletemeibm.net (Bob O) added this to
>the Deja-News archive:
>
>....you were saying?
>
Sorry! I only counted my C: partition also. I didn't count my data
partitions, or my other computers, or my tape/disk libraries, those
would amount to multiple of 100,000 files, but we were talking about
applications primarily were we not?
Of course I don't have many game files though. But then again I don't
want game files either. I find computer games to be boring. I play an
occasional short game like mahhjong or solitaire but I only do that when
I am really bored.
Bob O
wrote:
>On 6 Sep 1998 03:55:19 GMT, bo...@wasatch.com (Bob Hauck) added this to the
>Deja-News archive:
>
>>In article <35f2c3a5.27269538@news-server>,
>> sden...@san.rr.com (Steven C. Den Beste) writes:
>>
>>> The upshot is that over a period of weeks, after two or three
>>> defrags,
>>
>>You people are still defragging disks? What is this, the stone
>>age or something?
>
>
>Oh, it isn't really necessary in the sense of the disk becoming really
>fragmented; the last couple of times I defragged it was precisely to take
>advantage of this feature.
>
>Each time I defrag, the programs I use most commonly are optimized to load
>as fast as possible.
>
>Does OS/2 do this?
>
Yes, but you don't have to defrag for the OS/2 optimizations to be in
effect.
Jack Troughton
wrote:
> On 6 Sep 1998 03:55:19 GMT, bo...@wasatch.com (Bob Hauck) added this to the
> Deja-News archive:
>
> >In article <35f2c3a5.27269538@news-server>,
> > sden...@san.rr.com (Steven C. Den Beste) writes:
> >
> >> The upshot is that over a period of weeks, after two or three
> >> defrags,
> >
> >You people are still defragging disks? What is this, the stone
> >age or something?
>
>
> Oh, it isn't really necessary in the sense of the disk becoming really
> fragmented; the last couple of times I defragged it was precisely to take
> advantage of this feature.
>
> Each time I defrag, the programs I use most commonly are optimized to load
> as fast as possible.
>
> Does OS/2 do this?
>
Pretty much; it's a feature that has been in HPFS since the beginning,
since it uses a far superior bin packing algorithm than FAT has ever
used. It's not based on use so much as it's based on finding the
optimal sized free disk area for the file in question. Eventually,
files will be moved around to create these areas for new files, if you
use the disk heavily enough. This is the reason you _might_ have to
defrag the disk after a year or two of heavy use. This can be easily
accomplished by doing an xcopy to a backup, and then xcopying back.
You can use xcopy to create bootable systems on an hpfs drive, by the
way; you don't need to use a program like sys.com in the FAT world to
create a bootable partition. Info-zip's freeware zip program will do
this as well; so restoring a boot partition is very simple. All you
need to do is to unzip the file to a suitably formatted partition,
mark it as bootable in the boot manager, and you're done.
One place where warp is still weak is that it doesn't take very kindly
to changing drive letters, so you do have to be sure that you do keep
you bootable partition as the same drive letter that it was
previously. Whenever the LVM gets around to trickling down from the
corporate world to us lowly kitchentop users (it'll be in public beta
soon for Aurora; I imagine we'll get it in another six months to a
year) that problem will go away.
As I understand it, MS did the initial design for HPFS; probably the
best piece of coding they have ever done.
Jack Troughton ICQ:7494149
http://modemcable162.163.mmtl.videotron.net:8000/
jaft at adan.kingston.net
jack.troughton at videotron.ca
Montréal PQ Canada
There are two kinds of people in this world: the kind that
divides everybody into two kinds of people, and everybody else.
Kris Kwilas
>On Sun, 6 Sep 1998 15:29:15, sden...@san.rr.com (Steven C. Den Beste)
>>as fast as possible. Does OS/2 do this?
>effect.
No.
HPFS's optimizations are oriented toward limiting the number of
extents for any particular file. Windows 98's defrag/optimization routine
(based in large part on Intel's work in this area I believe) go a bit beyond
that in actually optimizing the layout of programs and files on the disk
to speed loading based on how the machine is used. Note that means
that files (particularly the resources/exes/dlls used to load programs)
may actually wind up fragmented on the disk because the application
itself will load more efficiently that way.
Kris
Steven C. Den Beste
the Deja-News archive:
>On Sun, 6 Sep 1998 15:29:15, sden...@san.rr.com (Steven C. Den Beste)
>wrote:
>
>>On 6 Sep 1998 03:55:19 GMT, bo...@wasatch.com (Bob Hauck) added this to the
>>Deja-News archive:
>>
>>>In article <35f2c3a5.27269538@news-server>,
>>> sden...@san.rr.com (Steven C. Den Beste) writes:
>>>
>>>> The upshot is that over a period of weeks, after two or three
>>>> defrags,
>>>
>>>You people are still defragging disks? What is this, the stone
>>>age or something?
>>
>>
>>Oh, it isn't really necessary in the sense of the disk becoming really
>>fragmented; the last couple of times I defragged it was precisely to take
>>advantage of this feature.
>>
>>Each time I defrag, the programs I use most commonly are optimized to load
>>as fast as possible.
>>
>>Does OS/2 do this?
>>
>
>Yes, but you don't have to defrag for the OS/2 optimizations to be in
>effect.
>
>Bob O - Computing for fun
Let's try this again:
How does OS/2 reorganize its disk to make program load-times faster, then?
Remember that when a program loads, it also loads DLLs and other things as
well.
The Win 98 defragger brings those other things to the front so that they
are in sequence with the parts of the program, so that when it is loaded
*EVERYTHING* it needs is right there, in the order it needs them. This cuts
way down on disk seek time during program loading.
How does OS/2 accomplish this?
David H. McCoy
119.ca.us.ibm.net>, osb...@deletemeibm.net says...
> On Sun, 6 Sep 1998 15:29:15, sden...@san.rr.com (Steven C. Den Beste)
> wrote:
>
> >On 6 Sep 1998 03:55:19 GMT, bo...@wasatch.com (Bob Hauck) added this to the
> >Deja-News archive:
> >
> >>In article <35f2c3a5.27269538@news-server>,
> >> sden...@san.rr.com (Steven C. Den Beste) writes:
> >>
> >>> The upshot is that over a period of weeks, after two or three
> >>> defrags,
> >>
> >>You people are still defragging disks? What is this, the stone
> >>age or something?
> >
> >
> >Oh, it isn't really necessary in the sense of the disk becoming really
> >fragmented; the last couple of times I defragged it was precisely to take
> >advantage of this feature.
> >
> >Each time I defrag, the programs I use most commonly are optimized to load
> >as fast as possible.
> >
> >Does OS/2 do this?
> >
>
> Yes, but you don't have to defrag for the OS/2 optimizations to be in
> effect.
>
> Bob O - Computing for fun
>
OS/2 DOESN'T do this, Bob. Damn, do you know anything about the OS you are
using?
Clearly not.
David H. McCoy
> On 6 Sep 1998 16:17:52 GMT, osb...@deletemeibm.net (Bob O) added this to
> the Deja-News archive:
>
> >wrote:
> >
> >>In article <35f2c3a5.27269538@news-server>, sden...@san.rr.com says...
> >>>
> >>> Bob, I don't particularly care to go make a census of eighteen thousand
> >>> files just to come up with a figure which will only be slightly more
> >>> precise than what I just described. The fact is that the change from a 4K
> >>> cluster to a 256 byte cluster will not save very much space.
> >>>
> >>> If you don't like my numbers, fine.
> >>>
> >>
> >>Bob has time to perform such a count because there are fewer applications,
> >>and thus, fewer files to track.
> >
> >Heeheehee! I happened to have over 30,000 files. So I guess I both
> >need more applications than
> > Steven does and they are available.
>
> I guess I didn't mention that I have five other 2G partitions besides that
> C: partition I mentioned.
>
> C: 18602 files.
> D: 4112 files
> E: 24821 files
> F: 18968 files
> G: 4948 files
> H: 1281 files
>
> Not to mention all the files on the 10 1G JAZ cartridges I have, nor the
> files I've rolled out onto CDs using my CDR drive.
>
>
>
He wasn't saying anything.
Bob O
>In article <Yqzh521VLZKl-p...@slip129-37-55-119.ca.us.ibm.net>, osb...@deletemeibm.net (Bob O) wrote:
>>On Sun, 6 Sep 1998 15:29:15, sden...@san.rr.com (Steven C. Den Beste)
>>>Each time I defrag, the programs I use most commonly are optimized to load
>>>as fast as possible. Does OS/2 do this?
>>Yes, but you don't have to defrag for the OS/2 optimizations to be in
>>effect.
>
>No.
>
>HPFS's optimizations are oriented toward limiting the number of
>extents for any particular file. Windows 98's defrag/optimization routine
>(based in large part on Intel's work in this area I believe) go a bit beyond
>that in actually optimizing the layout of programs and files on the disk
>to speed loading based on how the machine is used. Note that means
>that files (particularly the resources/exes/dlls used to load programs)
>may actually wind up fragmented on the disk because the application
>itself will load more efficiently that way.
>
>Kris
Sounds like this is not a solution for preventing fragmentation. One of
the benefits of limiting extents is files do not fragment in the first
place and thus while they may not stay put, you know the disk isn't
going to thrash to load the file when the time comes to load it.
If you still allow fragmentation of files how do you prevent the rapid
deterioration you experience with Windows? I have never run a disk
defragger on OS/2, see no need to as file access times simply do not
crop up as an issue as the disk never seems to fragment. I have run
utilities from time to time to see if a defrag is necessary but HPFS
seems to stay at about 99% unfragmented all the time.
HPFS also has a mechanism that cuts down on seek time in general. Since
HPFS has twice as much disk within the same reach as a FAT disk, to get
a benefit from putting file fragments in order as opposed to putting
files in order would mean you would have to have less than one half of
a file accessed on average to have a benefit from doing it this way as
opposed to the HPFS way assuming you use a file order methodology with
HPFS.
Seems unlikely that would be the case.
Personally I prefer not having to even think about fragmentation or file
ordering for that matter. All the reading and writing and maintenance
associated with defraggers is just mileage on the machinery and a
personal hassle as far as I am concerned. I prefer having a vehicle
that just goes and goes and doesn't need a tune up like the new 100,000
mile autos out there today. While I am sure you can tune these cars for
more performance you no doubt lose the 100,000 worry-free miles.
joseph
>
> the Deja-News archive:
>
> >
> >They might come crawling back to IBM and try to resell MS OS/2. It
> >works and all you guys say it needs is MS's marketing. MS OS/2, the
> >follwo-on to NT and WIN98 -- it's better faster, and cheaper. Sure
> >beats NT 2001.
>
> Fat chance of this. It doesn't run WIN32 software.
Neither does WindowsCE. MS has spent 1 billion is cash for WebTV and
TCI cables boxes to run WindowsCE. Then they are spending $ for game
console and PDA support. These WindowsCE devices will replace many
Win98 PCs form laptps to home systems for e-mail and browsing.
Most critical Win32 software at MS is going to be radically redesigned
to mimmic the fine granulaity of Java software which is designed for a
distributed environment. MS will also build new software using
windowsCE's sub-sets of Win32 API to fit the WindowsCE platform. MS is
going to down size Office this way so it will work like eSuite over a
network. They of course use MS's OO technlogy like COM or COM+, not
Javabeans. Like eSuite, additional suite functionality will be
optionally downloaded to the user.
That's the future. It's based on today's win32 software.
WRT to OS/2, it would be a snap to support win32 APIs on OS/2. The
infrastructure and hooks are all there and IBM's even show it's possible
with their Open32 APIs. Th eOS is stable -- all MS needs is to hook the
APIs into the stable OS. Or maybe MS would hire a few OS/2 hobbiests
and finish the Win32 application translator. That way the 400+
developers at MS could still tinker on NT 5.0 well into the next
century.
Either way, MS would finish with OS/2 Win32 support long before NT 5.0
ships. Intel would like this because OS/2 is optimized so it is tied to
the x86 CPU. It would weaken Java and LINUX becasue MS would have IBM
supporting their Win32 APIs.
Greed and egos will stop MS from rejoining IBM and going back to OS/2.
MS would have to admit that they cannot competitively finish building an
enterprise class version of NT 5.0.
Kris Kwilas
>On Sun, 6 Sep 1998 18:16:59, kwi...@stardock.net (Kris Kwilas) wrote:
>>extents for any particular file. Windows 98's defrag/optimization routine
>>(based in large part on Intel's work in this area I believe) go a bit beyond
>>that in actually optimizing the layout of programs and files on the disk
>>to speed loading based on how the machine is used. Note that means
>the benefits of limiting extents is files do not fragment in the first
To the best of my knowledge, there is nothing inherent in the Windows
95/98 implementations of FAT that attempt to prevent fragrmentation.
But, I don't think anyone has made that claim. ;-)
>place and thus while they may not stay put, you know the disk isn't
>going to thrash to load the file when the time comes to load it.
And after running the defragmenter in Windows 98, the same thing
applies (just like running any defragmenter for the the last how
many years? when used with the proper options).
>If you still allow fragmentation of files how do you prevent the rapid
>deterioration you experience with Windows?
Windows implementation of FAT/FAT32(whatever) does not attempt
to prevent extents when writing files (as opposed to HPFS). To
correct this, you optimize the disk with a defragmenter. "Rapid
deterioration" highly subjective.
>I have never run a disk
>defragger on OS/2, see no need to as file access times simply do not
>crop up as an issue as the disk never seems to fragment. I have run
>utilities from time to time to see if a defrag is necessary but HPFS
>seems to stay at about 99% unfragmented all the time.
Experiences may vary. I certainly agree that HPFS doesn't really require
much in the way of optimization (unless you start filling up the partition,
which is another matter entirely). However, 99% of the files on your
HPFS partition are almost certainly _not_ contiguous (note: this statement
is slightly different than yours above) either as files or in the total layout
of all files on the disk. Unless you XCOPY everything off and back on
a regular basis. ;-)
>HPFS also has a mechanism that cuts down on seek time in general. Since
That would be the B-Tree's (in part).
>HPFS has twice as much disk within the same reach as a FAT disk, to get
>a benefit from putting file fragments in order as opposed to putting
>files in order would mean you would have to have less than one half of
>a file accessed on average to have a benefit from doing it this way as
>opposed to the HPFS way assuming you use a file order methodology with
>HPFS.
Ask Microsoft (and Intel, since they developed it ;-) for the benchmarks.
>Seems unlikely that would be the case.
But the point is not the global state of an arbitrary file on the partition.
>Personally I prefer not having to even think about fragmentation or file
>ordering for that matter.
Excellent point. Perhaps that's why the Windows 98 defragmenter
schedules itself to run (once initially configured) w/o user intervention.
It handles this on its own, including background monitoring to determine
"how" it should optimize the machine for improved performance (rather
than just unfragmenting files).
>All the reading and writing and maintenance
>associated with defraggers is just mileage on the machinery and a
>personal hassle as far as I am concerned.
The hassle isn't there (unless that initial 30 seconds to activate the
feature counts). Mileage on the machinery? Better only run those
servers at certain hours of the day. We wouldn't want any unnecessary
disk access happening, particularly if it was working to speed up
the machine...
>I prefer having a vehicle
>that just goes and goes and doesn't need a tune up like the new 100,000
>mile autos out there today. While I am sure you can tune these cars for
>more performance you no doubt lose the 100,000 worry-free miles.
OK, but you still were incorrect in responding to Den Beste. ;-)
Kris
Bob O
>In article <Yqzh521VLZKl-p...@slip129-37-55-60.ca.us.ibm.net>, osb...@deletemeibm.net (Bob O) wrote:
>>On Sun, 6 Sep 1998 18:16:59, kwi...@stardock.net (Kris Kwilas) wrote:
>>>extents for any particular file. Windows 98's defrag/optimization routine
>>>(based in large part on Intel's work in this area I believe) go a bit beyond
>>>that in actually optimizing the layout of programs and files on the disk
>>>to speed loading based on how the machine is used. Note that means
>>Sounds like this is not a solution for preventing fragmentation. One of
>>the benefits of limiting extents is files do not fragment in the first
>
>To the best of my knowledge, there is nothing inherent in the Windows
>95/98 implementations of FAT that attempt to prevent fragrmentation.
>But, I don't think anyone has made that claim. ;-)
Well Steven seems to be claiming that disk defragging isn't necessary
anymore. Witness this exchange:
">You people are still defragging disks? What is this, the stone
>age or something?
Steven Den Beste:
Oh, it isn't really necessary in the sense of the disk becoming really
fragmented; the last couple of times I defragged it was precisely to
take
advantage of this feature."
End quote.
I guess the question becomes why run it more than once?
>>place and thus while they may not stay put, you know the disk isn't
>>going to thrash to load the file when the time comes to load it.
>
>And after running the defragmenter in Windows 98, the same thing
>applies (just like running any defragmenter for the the last how
>many years? when used with the proper options).
>>If you still allow fragmentation of files how do you prevent the rapid
>>deterioration you experience with Windows?
>
>Windows implementation of FAT/FAT32(whatever) does not attempt
>to prevent extents when writing files (as opposed to HPFS). To
>correct this, you optimize the disk with a defragmenter. "Rapid
>deterioration" highly subjective.
It didn't seem subjective back when I ran Windows and felt compelled to
defrag about once a month.
>>I have never run a disk
>>defragger on OS/2, see no need to as file access times simply do not
>>crop up as an issue as the disk never seems to fragment. I have run
>>utilities from time to time to see if a defrag is necessary but HPFS
>>seems to stay at about 99% unfragmented all the time.
>
>Experiences may vary. I certainly agree that HPFS doesn't really require
>much in the way of optimization (unless you start filling up the partition,
>which is another matter entirely). However, 99% of the files on your
>HPFS partition are almost certainly _not_ contiguous (note: this statement
>is slightly different than yours above) either as files or in the total layout
>of all files on the disk. Unless you XCOPY everything off and back on
>a regular basis. ;-)
Well I suppose that is what I do now when I buy a larger disk. Not sure
what method Disk Copy uses though.
Oh you mean my comment that "Yes OS/2 does that". You are right, I
should have said "Yes OS/2 optimizes disks but in a different manner".
Kris Kwilas
>On Sun, 6 Sep 1998 19:37:52, kwi...@stardock.net (Kris Kwilas) wrote:
>Well Steven seems to be claiming that disk defragging isn't necessary
>anymore. Witness this exchange:
>">You people are still defragging disks? What is this, the stone
>>age or something?
>Steven Den Beste:
>Oh, it isn't really necessary in the sense of the disk becoming really
>fragmented; the last couple of times I defragged it was precisely to
There are three issues on the table here:
1. not defragmenting at all (and HPFS fits in here w/Windows coming
in on FAT w/the problems of fragmentation that it has had as part
of the DOS/FAT legacy for years).
2. defragmenting a disk/files to increase continuity and lessen
"disk thrashing" (standard defrag stuff, also relatively unnecessary
with HPFS as you've pointed out).
3. optimizing applications and their position on the disk based on
system usage.
Steven's statement above is dealing with 3 (as opposed to defragging
in general which is covered by 2). With Windows 98 (and only Windows
98, though I suppose Norton may be doing something similar ;-), this
is a factor.
>I guess the question becomes why run it more than once?
Two reasons: files are going to fragment with FAT and, more importantly
with Windows 98, application usage doesn't stay constant. At the most
basic level, you add a new application to the system. As you use it,
Windows 98 will take note of that and the next time the "defrag/optimize"
operation is invoked to do its thing, it will take that into account and
include that application (assuming it meets the internal threshold in
the monitoring process/optimizer for inclusion) in the process.
>>correct this, you optimize the disk with a defragmenter. "Rapid
>>deterioration" highly subjective.
>It didn't seem subjective back when I ran Windows and felt compelled to
>defrag about once a month.
Substitute a virus scanner for disk optimizer and look at this approach
again. Windows 98 takes care of this in the background (if requested).
>>OK, but you still were incorrect in responding to Den Beste. ;-)
>Oh you mean my comment that "Yes OS/2 does that". You are right, I
>should have said "Yes OS/2 optimizes disks but in a different manner".
Close. OS/2 (HPFS) optimizes files.
It's really an apple/oranges comparison (i.e. OS/2 to Windows 98). Windows
95 and NT w/NTFS on the other hand, that's a whole 'nother kettle of sih. ;-)
Kris
Bob O
wrote:
>On 6 Sep 1998 18:01:46 GMT, osb...@deletemeibm.net (Bob O) added this to
>the Deja-News archive:
>
>>On Sun, 6 Sep 1998 15:29:15, sden...@san.rr.com (Steven C. Den Beste)
>>wrote:
>>
>>>On 6 Sep 1998 03:55:19 GMT, bo...@wasatch.com (Bob Hauck) added this to the
>>>Deja-News archive:
>>>
>>>>In article <35f2c3a5.27269538@news-server>,
>>>> sden...@san.rr.com (Steven C. Den Beste) writes:
>>>>
>>>>> The upshot is that over a period of weeks, after two or three
>>>>> defrags,
>>>>
>>>>You people are still defragging disks? What is this, the stone
>>>>age or something?
>>>
>>>
>>>Oh, it isn't really necessary in the sense of the disk becoming really
>>>fragmented; the last couple of times I defragged it was precisely to take
>>>advantage of this feature.
>>>
>>>Each time I defrag, the programs I use most commonly are optimized to load
>>>as fast as possible.
>>>
>>>Does OS/2 do this?
>>>
>>
>>Yes, but you don't have to defrag for the OS/2 optimizations to be in
>>effect.
>>
>>Bob O - Computing for fun
>
>Let's try this again:
>
>How does OS/2 reorganize its disk to make program load-times faster, then?
>Remember that when a program loads, it also loads DLLs and other things as
>well.
>
>The Win 98 defragger brings those other things to the front so that they
>are in sequence with the parts of the program, so that when it is loaded
>*EVERYTHING* it needs is right there, in the order it needs them. This cuts
>way down on disk seek time during program loading.
>
>How does OS/2 accomplish this?
>
OS/2 accomplishes by bringing the partition tables closer to the data
rather than bringing data closer to a favored corner of the disk.
What you are explaining that Microsoft did was to an effort to reduce
average seek time by making infrequently accessed files the slow files
to access and the more frequently accessed files close to the partition
tables.
I tried this years ago with Windows to try to get better performance
(ordering files) however, it really didn't make for a hill of beans
difference in speed. I think I shaved about 5% off the time that
Windows took to load.
OS/2 on the other hand uses a partition table scheme that minimizes the
average movement of the read/write heads by distributing the tables
within the data (one direction for the heads to go [a result of putting
the partition table at the front of the disk] results in half the area
you can cover with one movement of the heads. By distributing the
tables within the disk seek times are dramatically reduced by reducing
the average distance the heads have to travel.
Your system essentially depends upon having a large amount of little
used data on your disk. The more crap you have on the disk the more
good the system will help you avoid FAT partition table weaknesses.
If everything on your disk has an equal opportunity to be used it is of
no help whatsoever. Not true with the OS/2 system. The OS/2 system
dramatically reduces the average seek time to the corners of your disk
by shortening the distance to them all the while maintaining an equal to
Windows best access time to the closest data. The Windows method
doesn't speed up your disk it simply tries to make the distant places to
travel on your disk a place to store the stuff you don't use much. With
OS/2 there are not really any distant places on your disk.
My approach to disk clutter is to put different kinds of data in
different partitions and to quickly erase programs that I don't use.
This approach saves money on buying hard drives as well.
Bob O
wrote:
>In article <Yqzh521VLZKl-pn2-8fdqcR3007km@slip129-37-55-
>119.ca.us.ibm.net>, osb...@deletemeibm.net says...
>> On Sun, 6 Sep 1998 15:29:15, sden...@san.rr.com (Steven C. Den Beste)
>> wrote:
>>
>> >On 6 Sep 1998 03:55:19 GMT, bo...@wasatch.com (Bob Hauck) added this to the
>> >Deja-News archive:
>> >
>> >>In article <35f2c3a5.27269538@news-server>,
>> >> sden...@san.rr.com (Steven C. Den Beste) writes:
>> >>
>> >>> The upshot is that over a period of weeks, after two or three
>> >>> defrags,
>> >>
>> >>You people are still defragging disks? What is this, the stone
>> >>age or something?
>> >
>> >
>> >Oh, it isn't really necessary in the sense of the disk becoming really
>> >fragmented; the last couple of times I defragged it was precisely to take
>> >advantage of this feature.
>> >
>> >Each time I defrag, the programs I use most commonly are optimized to load
>> >as fast as possible.
>> >
>> >Does OS/2 do this?
>> >
>>
>> Yes, but you don't have to defrag for the OS/2 optimizations to be in
>> effect.
>>
>> Bob O - Computing for fun
>>
>
>OS/2 DOESN'T do this, Bob. Damn, do you know anything about the OS you are
>using?
>
>Clearly not.
Sure it does David. OS/2 HPFS has a modified partition table that makes
the average piece of data closer to the tables, but it is a done deal
you don't have to defrag to take advantage of it. Don't you know
anything about OS/2?
Bob O
wrote:
>In article <35f3ba4b.5388799@news-server>, sden...@san.rr.com says...
>> On 6 Sep 1998 16:17:52 GMT, osb...@deletemeibm.net (Bob O) added this to
>> the Deja-News archive:
>>
>> >On Sun, 6 Sep 1998 02:15:42, dmccoy@REMOVE_MEmnsinc.com (David H. McCoy)
>> >wrote:
>> >
>> >>In article <35f2c3a5.27269538@news-server>, sden...@san.rr.com says...
>> >>>
>> >>> Bob, I don't particularly care to go make a census of eighteen thousand
>> >>> files just to come up with a figure which will only be slightly more
>> >>> precise than what I just described. The fact is that the change from a 4K
>> >>> cluster to a 256 byte cluster will not save very much space.
>> >>>
>> >>> If you don't like my numbers, fine.
>> >>>
>> >>
>> >>Bob has time to perform such a count because there are fewer applications,
>> >>and thus, fewer files to track.
>> >
>> >Heeheehee! I happened to have over 30,000 files. So I guess I both
>> >need more applications than
>> > Steven does and they are available.
>>
>> I guess I didn't mention that I have five other 2G partitions besides that
>> C: partition I mentioned.
>>
>> C: 18602 files.
>> D: 4112 files
>> E: 24821 files
>> F: 18968 files
>> G: 4948 files
>> H: 1281 files
>>
>> Not to mention all the files on the 10 1G JAZ cartridges I have, nor the
>> files I've rolled out onto CDs using my CDR drive.
>>
>> ...you were saying?
>>
>>
>
>He wasn't saying anything.
As usual David you were wrong. But that has never stopped you from
making a comment; now has it?
Bob O
wrote:
>On 6 Sep 1998 18:01:46 GMT, osb...@deletemeibm.net (Bob O) added this to
>the Deja-News archive:
>
>>On Sun, 6 Sep 1998 15:29:15, sden...@san.rr.com (Steven C. Den Beste)
>>wrote:
>>
>>>On 6 Sep 1998 03:55:19 GMT, bo...@wasatch.com (Bob Hauck) added this to the
>>>Deja-News archive:
>>>
>>>>In article <35f2c3a5.27269538@news-server>,
>>>> sden...@san.rr.com (Steven C. Den Beste) writes:
>>>>
>>>>> The upshot is that over a period of weeks, after two or three
>>>>> defrags,
>>>>
>>>>You people are still defragging disks? What is this, the stone
>>>>age or something?
>>>
>>>
>>>Oh, it isn't really necessary in the sense of the disk becoming really
>>>fragmented; the last couple of times I defragged it was precisely to take
>>>advantage of this feature.
>>>
>>>Each time I defrag, the programs I use most commonly are optimized to load
>>>as fast as possible.
>>>
>>>Does OS/2 do this?
>>>
>>
>>Yes, but you don't have to defrag for the OS/2 optimizations to be in
>>effect.
>>
>>Bob O - Computing for fun
>
>Let's try this again:
>
>How does OS/2 reorganize its disk to make program load-times faster, then?
>Remember that when a program loads, it also loads DLLs and other things as
>well.
>
>The Win 98 defragger brings those other things to the front so that they
>are in sequence with the parts of the program, so that when it is loaded
>*EVERYTHING* it needs is right there, in the order it needs them. This cuts
>way down on disk seek time during program loading.
>
>How does OS/2 accomplish this?
Actually I don't know if that Windows feature is very desirable or not.
The one program I want to load fast is really boot up as it has to be
done before I start doing anything. I do that less than once a day on
average. I load my organizer, Lotus 123, the dialer, Netscape,
ProNews/2 all more often than I boot up, but I would not want them to be
the ones optimized.
Typically when I boot up, I boot, start my organizer, start the dialer,
start Netscape, start ProNews/2, and start 123. What I want is for the
organizer to be running a quickly as possible and the rest I could give
a shuck about because by the time I am done with one, the others are all
loaded.
I am not sure that frequency of use is really all that closely
correlated with how much waiting for use you actually do as a user in a
"true" multitasking system as I open and close a lot of objects multiple
times during one session, (sometimes I only reboot once a week) but it
is usually the main session of whatever it is that I really want loaded
fastest.
David H. McCoy
149.ca.us.ibm.net>, osb...@deletemeibm.net says...
> >
> >OK, but you still were incorrect in responding to Den Beste. ;-)
>
> Oh you mean my comment that "Yes OS/2 does that". You are right, I
> should have said "Yes OS/2 optimizes disks but in a different manner".
>
>
>
OS/2 does not optimize the disk with respect to file loading, which was
clearly Steven's question. Everything you have mentioned is also done by
NTFS, so OS/2 has nothing special there.
You were just wrong.
Bob O
>In article <Yqzh521VLZKl-p...@slip129-37-55-149.ca.us.ibm.net>, osb...@deletemeibm.net (Bob O) wrote:
>>On Sun, 6 Sep 1998 19:37:52, kwi...@stardock.net (Kris Kwilas) wrote:
>>Well Steven seems to be claiming that disk defragging isn't necessary
>>anymore. Witness this exchange:
>>">You people are still defragging disks? What is this, the stone
>>>age or something?
>>Steven Den Beste:
>>Oh, it isn't really necessary in the sense of the disk becoming really
>>fragmented; the last couple of times I defragged it was precisely to
>
>There are three issues on the table here:
>
>1. not defragmenting at all (and HPFS fits in here w/Windows coming
> in on FAT w/the problems of fragmentation that it has had as part
> of the DOS/FAT legacy for years).
>2. defragmenting a disk/files to increase continuity and lessen
> "disk thrashing" (standard defrag stuff, also relatively unnecessary
> with HPFS as you've pointed out).
>3. optimizing applications and their position on the disk based on
> system usage.
>
>Steven's statement above is dealing with 3 (as opposed to defragging
>in general which is covered by 2). With Windows 98 (and only Windows
>98, though I suppose Norton may be doing something similar ;-), this
>is a factor.
>
>>I guess the question becomes why run it more than once?
>
>Two reasons: files are going to fragment with FAT and, more importantly
>with Windows 98, application usage doesn't stay constant. At the most
>basic level, you add a new application to the system. As you use it,
>Windows 98 will take note of that and the next time the "defrag/optimize"
>operation is invoked to do its thing, it will take that into account and
>include that application (assuming it meets the internal threshold in
>the monitoring process/optimizer for inclusion) in the process.
Again if it is true that you want this app to push your boot up sequence
back down the priority track. It sounds like it might be an interesting
feature for other than application partitions, such as data partitions,
but I don't use many big data files so I don't see that as an issue for
me anyway.
Personally, if I were into file ordering (or partial file ordering) I
would probably choose to use the software that has been around a while
that allows you to select directories that you want to move close to the
front to make items such as your environment or shells to boot faster.
However, that has always been such a dubious benefit to me I never did
that beyond some initial experimentation with Norton Utiliites in trying
to get a faster system when I was running Windows. I don't even know if
such file ordering is available for OS/2's approach to partition tables.
It seems to be really Microsoftish that they would come up with a
non-selective system to use that doesn't let me tell it what programs I
want to load fast but figure that it knows best. Does this system have
detection for when I am sitting at the workstation with a blank stare or
when I went to the coffee machine, or when I am already working in
another application while files load in the background?
>>>correct this, you optimize the disk with a defragmenter. "Rapid
>>>deterioration" highly subjective.
>>It didn't seem subjective back when I ran Windows and felt compelled to
>>defrag about once a month.
>
>Substitute a virus scanner for disk optimizer and look at this approach
>again. Windows 98 takes care of this in the background (if requested).
Something else with OS/2 I don't do. I didn't run virus scanners for
DOS/Windows until after being struck the first time and have been
running the same philosophy here with OS/2 for 6 1/2 years now but have
been living lucky I guess.
>>>OK, but you still were incorrect in responding to Den Beste. ;-)
>>Oh you mean my comment that "Yes OS/2 does that". You are right, I
>>should have said "Yes OS/2 optimizes disks but in a different manner".
>
>Close. OS/2 (HPFS) optimizes files.
OK they both optimize files (or in the case of Windows file parts).
>It's really an apple/oranges comparison (i.e. OS/2 to Windows 98). Windows
>95 and NT w/NTFS on the other hand, that's a whole 'nother kettle of sih. ;-)
Well I have always been impressed with how quiet my drives work with
OS/2 compared to booting to my second C: partition and running DOS and
Windows. I formatted my NT partition with NTFS but that has been a big
mistake as NT does not read my HPFS partitions still doesn't work with
my CDROM, and in general has made a pain of itself. I had planned upon
a casual evaluation of NT but it isn't so simple to get going that it is
not without more pain than I had budgeted for it so I really haven't had
the time to check it out.
Bob O
wrote:
>In article <Yqzh521VLZKl-pn2-GbOWpzXQrLEE@slip129-37-55-
>149.ca.us.ibm.net>, osb...@deletemeibm.net says...
>> >
>> >OK, but you still were incorrect in responding to Den Beste. ;-)
>>
>> Oh you mean my comment that "Yes OS/2 does that". You are right, I
>> should have said "Yes OS/2 optimizes disks but in a different manner".
>>
>> Bob O - Computing for fun
>>
>>
>
>OS/2 does not optimize the disk with respect to file loading, which was
>clearly Steven's question. Everything you have mentioned is also done by
>NTFS, so OS/2 has nothing special there.
>
>You were just wrong.
No David. We were talking about Windows 98 not NT.
I am sure NTFS is a fine file system but it was designed with servers in
mind. Properly NTFS should be compared to the 32 bit HPFS.
A better question might be why isn't NTFS available for Win98? All I
can do is guess, but my guess is that Win98 still uses DOS structures
for file systems and NTFS is just too big of a file system to load into
DOS and still have room to run whatever else MS runs there in Win98.
Bob O
wrote:
>In article <Yqzh521VLZKl-pn2-GbOWpzXQrLEE@slip129-37-55-
>149.ca.us.ibm.net>, osb...@deletemeibm.net says...
>> >
>> >OK, but you still were incorrect in responding to Den Beste. ;-)
>>
>> Oh you mean my comment that "Yes OS/2 does that". You are right, I
>> should have said "Yes OS/2 optimizes disks but in a different manner".
>>
>> Bob O - Computing for fun
>>
>>
>
>OS/2 does not optimize the disk with respect to file loading, which was
>clearly Steven's question. Everything you have mentioned is also done by
>NTFS, so OS/2 has nothing special there.
>
>You were just wrong.
You are only really partly right there David. It is true that OS/2 does
not "optimize" the disk for lile loading. That is because the disk is
already optimized for file loading when you install the system and
format the drive with HPFS.
Now you can get into an argument about which kind of optimization is
faster, but that would be an admission that you are wrong.
Steven C. Den Beste
the Deja-News archive:
>On Sun, 6 Sep 1998 18:16:59, kwi...@stardock.net (Kris Kwilas) wrote:
>
>>In article <Yqzh521VLZKl-p...@slip129-37-55-119.ca.us.ibm.net>, osb...@deletemeibm.net (Bob O) wrote:
>>>On Sun, 6 Sep 1998 15:29:15, sden...@san.rr.com (Steven C. Den Beste)
>>>>Each time I defrag, the programs I use most commonly are optimized to load
>>>>as fast as possible. Does OS/2 do this?
>>>Yes, but you don't have to defrag for the OS/2 optimizations to be in
>>>effect.
>>
>>No.
>>
>>HPFS's optimizations are oriented toward limiting the number of
>>extents for any particular file. Windows 98's defrag/optimization routine
>>(based in large part on Intel's work in this area I believe) go a bit beyond
>>that in actually optimizing the layout of programs and files on the disk
>>to speed loading based on how the machine is used. Note that means
>>that files (particularly the resources/exes/dlls used to load programs)
>>may actually wind up fragmented on the disk because the application
>>itself will load more efficiently that way.
>>
>>Kris
>
>Sounds like this is not a solution for preventing fragmentation. One of
>the benefits of limiting extents is files do not fragment in the first
>place and thus while they may not stay put, you know the disk isn't
>going to thrash to load the file when the time comes to load it.
>
>If you still allow fragmentation of files how do you prevent the rapid
>deterioration you experience with Windows? I have never run a disk
>defragger on OS/2, see no need to as file access times simply do not
>crop up as an issue as the disk never seems to fragment. I have run
>utilities from time to time to see if a defrag is necessary but HPFS
>seems to stay at about 99% unfragmented all the time.
For many kinds of files, you don't necessarily load the entire thing.
(That's quite common for DLLs, for instance.) What Win 98 is doing is
monitoring the actual load patterns of the applications used on a
particular machine by the user(s) of that particular machine, and optimizes
the layout of files and file fragments so as to optimize the loading time
of the applications used by the user(s).
There's no heuristic where someone ahead of time tried to make a general
guess of what would work best collectively for the entire user base. The
Win 98 algorithm adapts *each* machine to the specific behavior of its
user(s) determined by direct observation.
If the usage pattern changes radically, the solution is to run the
defragger again. The monitor demon is always running, and it has a
relatively short memory (4 loads per application, as it turns out). So it
rarely takes more than a few days of use before it has enough info for the
defragger to respond properly to the new usage pattern.
What this is doing is to live up to the spirit of defragging, not the
letter. The traditional reason for defragging the files on a partition was
to bring the pieces of those files together; by doing so it was expected
that this would enhance performance. Microsoft's new approach takes a step
further and deliberately fragments a file *IF* it has direct observation
from the monitoring demon that doing so will speed performance on that
particular machine.
>HPFS also has a mechanism that cuts down on seek time in general. Since
>HPFS has twice as much disk within the same reach as a FAT disk, to get
>a benefit from putting file fragments in order as opposed to putting
>files in order would mean you would have to have less than one half of
>a file accessed on average to have a benefit from doing it this way as
>opposed to the HPFS way assuming you use a file order methodology with
>HPFS.
>
>Seems unlikely that would be the case.
You lost me somewhere in there. Could you explain that again in more
detail?
How can HPFS have twice as much disk within the same reach? Space is space;
to scan to a file stored at the 100M point, you have to seek over 100M of
the drive.
If you're referring to the efficiency of usage, don't forget that FAT uses
different cluster sizes with different sizes of drives. If a FAT drive is
relatively small, the cluster size can be such that efficiency is really
quite good. It's not a given that HPFS is twice as efficient as FAT (in
fact, I don't believe it's ever going to be that good in practice).
For instance, a 256M FAT partition uses a 4K cluster. An HPFS partition
that size isn't going to have much of an advantage in wastage/efficiency.
>Personally I prefer not having to even think about fragmentation or file
>ordering for that matter. All the reading and writing and maintenance
>associated with defraggers is just mileage on the machinery and a
>personal hassle as far as I am concerned. I prefer having a vehicle
>that just goes and goes and doesn't need a tune up like the new 100,000
>mile autos out there today. While I am sure you can tune these cars for
>more performance you no doubt lose the 100,000 worry-free miles.
>
>Bob O - Computing for fun
Actually, HPFS partitions *do* get fragmented; all disk structure formats
are subject to this problem. HPFS isn't immune. It's more a matter of rate;
it's quantitative rather than qualitative. Some formats become fragmented
more rapidly than others do.
What's different about HPFS is that IBM didn't provide you with a utility
to do anything about it. It doesn't mean the problem isn't there.
IBM has managed to make it so that the problem doesn't become severe
anything like as rapidly as it does on a FAT drive.
But I believe that third parties sell HPFS defraggers. Evidently they
thought there was a market for them.
What I don't understand is why no-one is doing the one really simple thing
which would speed disk operations. The reason I don't understand it is that
it is easy and essentially free: put the disk control blocks in the center
of the partition instead of at the beginning. Then the maximum seek
distance from the FAT/whosis/whatsis demanded by whichever filesystem is in
use is half the length of the partition. It is literally changed by
altering a constant in an assign in the formatter and a few other places to
a calculation based on the size of the partition.
Steven C. Den Beste
the Deja-News archive:
>On Sun, 6 Sep 1998 19:37:52, kwi...@stardock.net (Kris Kwilas) wrote:
>
>>In article <Yqzh521VLZKl-p...@slip129-37-55-60.ca.us.ibm.net>, osb...@deletemeibm.net (Bob O) wrote:
>>>On Sun, 6 Sep 1998 18:16:59, kwi...@stardock.net (Kris Kwilas) wrote:
>>>>extents for any particular file. Windows 98's defrag/optimization routine
>>>>(based in large part on Intel's work in this area I believe) go a bit beyond
>>>>that in actually optimizing the layout of programs and files on the disk
>>>>to speed loading based on how the machine is used. Note that means
>>>Sounds like this is not a solution for preventing fragmentation. One of
>>>the benefits of limiting extents is files do not fragment in the first
>>
>>To the best of my knowledge, there is nothing inherent in the Windows
>>95/98 implementations of FAT that attempt to prevent fragrmentation.
>>But, I don't think anyone has made that claim. ;-)
>
>Well Steven seems to be claiming that disk defragging isn't necessary
>anymore. Witness this exchange:
>
>">You people are still defragging disks? What is this, the stone
>>age or something?
>
>Steven Den Beste:
>Oh, it isn't really necessary in the sense of the disk becoming really
>fragmented; the last couple of times I defragged it was precisely to
>take
>advantage of this feature."
>End quote.
>
>I guess the question becomes why run it more than once?
OK, I didn't make myself clear. I've run the defragger on my C drive five
times since I installed Win 98. It wasn't necessary to run it that often;
the last three times I ran it because I wanted it to optimize my load
patterns, not because the drive had become heavily fragmented.
The rate at which a drive becomes fragmented is very much a function of the
rate at which files are deleted and files are written. My C drive is very
stable in that regard; I keep programs which I install and uninstall
(mostly games) on my D drive instead. The C driver is for utilities which I
expect to install and leave on the system.
As such, the rate at which my C drive fragments is very low. If
fragmentation were all I was worried about, I'd probably only have to
defrag my C drive about once a year.
But when I do install new utilities on the C drive and have been running
them for a while, I have decided it's a good idea to run the defragger to
permit it to re-optimize loading performance, in as much as the drive usage
pattern has changed.
And like many other things, hardware takes a hand. Fragmentation has a lot
less effect on disk operatings than the fundamental speed of the hardware
doing the work. My C partition is on an Ultra-Wide SCSI drive spinning 7200
RPM with a very fast seek time.
My disk performance is quite acceptable with FAT32, thank you.
>>>I prefer having a vehicle
>>>that just goes and goes and doesn't need a tune up like the new 100,000
>>>mile autos out there today. While I am sure you can tune these cars for
>>>more performance you no doubt lose the 100,000 worry-free miles.
>>
>>OK, but you still were incorrect in responding to Den Beste. ;-)
>
>Oh you mean my comment that "Yes OS/2 does that". You are right, I
>should have said "Yes OS/2 optimizes disks but in a different manner".
>
>Bob O - Computing for fun
The biggest difference is that OS/2 optimizes disks using a global
heuristic which was designed ahead of time to try to maximize performance
for the median user. All OS/2 systems are optimized the same way.
Win 98 optimizes the disk based on the specific usage patterns of the
particular user(s) of a particular machine. Different machines with
different usage patterns will be optimized differently.
It's a difference between predefined optimization and adaptive
optimization.
Steven C. Den Beste
the Deja-News archive:
>It seems to be really Microsoftish that they would come up with a
>non-selective system to use that doesn't let me tell it what programs I
>want to load fast but figure that it knows best. Does this system have
>detection for when I am sitting at the workstation with a blank stare or
>when I went to the coffee machine, or when I am already working in
>another application while files load in the background?
It monitors each and every request to load an application. As the
application loads, it monitors each and every file or portion of a file
which is loaded, and it monitors and logs the order in which they take
place. So yes, it does applications loaded in the background, in as much as
it monitors every load request no matter what.
When the system is defragged, part of what is taken into account is how
*often* a given application is used. It gives preference to the programs
which are used more often (which, if you think about it, makes a great deal
of sense).
As to not providing control to the user, you have to remember who the
target audience is for Win 98. Many of its target customers are home users
who are a little lost; it isn't just for hard core computer geeks.
Steven C. Den Beste
the Deja-News archive:
>On Sun, 6 Sep 1998 21:24:35, dmccoy@REMOVE_MEmnsinc.com (David H. McCoy)
>wrote:
>
>>In article <Yqzh521VLZKl-pn2-GbOWpzXQrLEE@slip129-37-55-
>>149.ca.us.ibm.net>, osb...@deletemeibm.net says...
>>> >
>>> >OK, but you still were incorrect in responding to Den Beste. ;-)
>>>
>>> Oh you mean my comment that "Yes OS/2 does that". You are right, I
>>> should have said "Yes OS/2 optimizes disks but in a different manner".
>>>
>>> Bob O - Computing for fun
>>>
>>>
>>
>>OS/2 does not optimize the disk with respect to file loading, which was
>>clearly Steven's question. Everything you have mentioned is also done by
>>NTFS, so OS/2 has nothing special there.
>>
>>You were just wrong.
>
>No David. We were talking about Windows 98 not NT.
>
>I am sure NTFS is a fine file system but it was designed with servers in
>mind. Properly NTFS should be compared to the 32 bit HPFS.
>
>A better question might be why isn't NTFS available for Win98? All I
>can do is guess, but my guess is that Win98 still uses DOS structures
>for file systems and NTFS is just too big of a file system to load into
>DOS and still have room to run whatever else MS runs there in Win98.
>
>Bob O - Computing for fun
Actually, it has nothing whatever to do with that. FAT32 is too big for
DOS, too; but they load it into extended memory, not into base memory. DOS
7.whatsis when FAT32 is in use is really quite different from DOS 6.X.
The real reason is that there's an all-or-nothing difference between Win NT
and Win 95: user validation.
You cannot use NT without logging in. And after you've logged in, you have
an official user identity which NT uses to permit/deny many of your
operations. For instance, on NTFS every file has a file owner. Every user
has a group. File operations can be permitted or denied on an owner or
group basis, or by a list of specific user IDs.
When Win 95 was created, they decided that having user identities was
overkill for the target audience of home/hobby users who mainly used their
computers (at that time) stand-alone. So while it's possible to set up user
idents and passwords on Win 95, it isn't required and most users don't use
it that way. And files don't have owners under Win 95.
(Unless, that is, they're accessing an NT server. For a Win 95 user to
access a Win NT server, the Win 95 user has to go through a validated login
sequence. At that point NT has the user-ID it needs for all its security.)
For Win 95 to support NTFS, user login would have become mandatory. Too
much of NTFS simply would make no sense without it.
I have to say that I was very, very disappointed to discover that NTFS
wasn't going to be supported by Win 95.
Steven C. Den Beste
the Deja-News archive:
>On Sun, 6 Sep 1998 18:43:23, sden...@san.rr.com (Steven C. Den Beste)
>wrote:
>
>>How does OS/2 reorganize its disk to make program load-times faster, then?
>>Remember that when a program loads, it also loads DLLs and other things as
>>well.
>>
>>The Win 98 defragger brings those other things to the front so that they
>>are in sequence with the parts of the program, so that when it is loaded
>>*EVERYTHING* it needs is right there, in the order it needs them. This cuts
>>way down on disk seek time during program loading.
>>
>>How does OS/2 accomplish this?
>>
>
Thank you. That was much more clear.
In actual practice, disk usage like so much else in computer science tends
to follow the "90/10" rule, which is to say that 90% of the usage is of 10%
of the data. (Another application of the 90/10 rule is "90% of the time is
spent in 10% of the code." By identifying that critical 10% you can
massively increase the speed of your program through optimization, without
wasting your time on code which only runs rarely.)
>My approach to disk clutter is to put different kinds of data in
>different partitions and to quickly erase programs that I don't use.
>This approach saves money on buying hard drives as well.
>
>Bob O - Computing for fun
I do that, too; otherwise I'd need at least four times as much disk space
as the 12G I have now. Few games that I load stay loaded; they stay around
until I know I'm tired of them, then they get uninstalled.
Do you know that "Starship Titanic" required over 700M of diskspace for a
full install? Eeek!
Robato Yao
I tried one of those. If that is defragging, those defraggers have a
much different definition in mind. They act not like as if they're
defragging a disk like you would with Scandisk, but rather, all they
seem to do is backup a file, then delete the original then copy the file
back to a disk. Imagine doing that to a hard disk full of files.
After that, I never tried those defraggers again. The speed benefits
are too little or non existant, and I never had problems with the
original files anyway. I have never "defragged" an HPFS driver for all
these few years.
Rgds,
Chris
(counting down from top 50 oxymorons...)
10. Tight slacks
9. Definite maybe
8. Pretty ugly
7. Twelve-ounce pound cake
6. Diet ice cream
5. Rap music
4. Working vacation
3. Exact estimate
2. Religious tolerance
And the NUMBER ONE top oxy-MORON
1. Microsoft Works
---From the Top 50 Oxymorons (thanks to Richard Kennedy)
David H. McCoy
149.ca.us.ibm.net>, osb...@deletemeibm.net says...
> On Sun, 6 Sep 1998 21:24:35, dmccoy@REMOVE_MEmnsinc.com (David H. McCoy)
> wrote:
>
> >In article <Yqzh521VLZKl-pn2-GbOWpzXQrLEE@slip129-37-55-
> >149.ca.us.ibm.net>, osb...@deletemeibm.net says...
> >> >
> >> >OK, but you still were incorrect in responding to Den Beste. ;-)
> >>
> >> Oh you mean my comment that "Yes OS/2 does that". You are right, I
> >> should have said "Yes OS/2 optimizes disks but in a different manner".
> >>
> >>
> >>
> >
> >clearly Steven's question. Everything you have mentioned is also done by
> >NTFS, so OS/2 has nothing special there.
> >
> >You were just wrong.
>
> No David. We were talking about Windows 98 not NT.
I know, which is why I said "also."
> I am sure NTFS is a fine file system but it was designed with servers in
> mind. Properly NTFS should be compared to the 32 bit HPFS.
Why on earth for? If we are discussing clients or workstations, both are
available on their respective clients, so all client features are up for
grabs.
> A better question might be why isn't NTFS available for Win98? All I
> can do is guess, but my guess is that Win98 still uses DOS structures
> for file systems and NTFS is just too big of a file system to load into
> DOS and still have room to run whatever else MS runs there in Win98.
How much memory does NTFS consume? No guessing, let's stick to fact and go
from there.
I would say that it is more likely that Win95/98 doesn't support NTFS for
security reasons.
> Bob O - Computing for fun
>
--
Peter Koehlmann
> > the Deja-News archive:
> >
> > >On Sun, 6 Sep 1998 02:15:42, dmccoy@REMOVE_MEmnsinc.com (David H. McCoy)
> > >wrote:
> > >
> > >>In article <35f2c3a5.27269538@news-server>, sden...@san.rr.com says...
> > >>>
> > >>> Bob, I don't particularly care to go make a census of eighteen thousand
> > >>> files just to come up with a figure which will only be slightly more
> > >>> precise than what I just described. The fact is that the change from a 4K
> > >>> cluster to a 256 byte cluster will not save very much space.
> > >>>
> > >>> If you don't like my numbers, fine.
> > >>>
> > >>
> > >>Bob has time to perform such a count because there are fewer applications,
> > >>and thus, fewer files to track.
> > >
> > >Heeheehee! I happened to have over 30,000 files. So I guess I both
> > >need more applications than
> > > Steven does and they are available.
> >
> > I guess I didn't mention that I have five other 2G partitions besides that
> > C: partition I mentioned.
> >
> > C: 18602 files.
> > D: 4112 files
> > E: 24821 files
> > F: 18968 files
> > G: 4948 files
> > H: 1281 files
> >
> > Not to mention all the files on the 10 1G JAZ cartridges I have, nor the
> > files I've rolled out onto CDs using my CDR drive.
> >
> > ...you were saying?
> >
> >
>
> He wasn't saying anything.
> -----------------------------------
> David H. McCoy
> dmccoy@REMOVE_MEmnsinc.com
> -----------------------------------
256 BYTE (!!!!) cluster on ANY PC. I'd like to see that one, 'cause the
cluster size can not be any lower than the sector size of a disk, which
is
512 Bytes. To this day there is NO OS, which would use partial SECTORS
for
a cluster, so just keep yout mouth shut and let this whole issue where
it
belongs, in the closet.
Just to keep things in order,
anyone who still uses FAT (FAT32 / FAT xx) just shows he still uses shit
as an OS, while there are far better solutions out there (HPFS, NTFS)
David H. McCoy
says...
If Steven type 256 either he made a typo or he knows something you do not.
I suspect it is a typo.
David H. McCoy
> In <35f1cad7...@nntp1.ba.best.com>, to...@tkpowers.com (Todd) writes:
> >On 5 Sep 1998 23:56:42 GMT, cro...@kuentos.guam.net (Robato Yao)
> >wrote:
> >
> >Heck, many DX 5.0 games run under NT 4.0. This should be impossible
> >since NT has only DX 3.0 available for it.
>
> Bullshit. By far most DX 5.0 games actually don't run on NT 4 with SP3.
> They are not supported in NT 4, many would not install under NT 4, and
> those that do run, often don't so reliably enough. Like crashing in
> the middle of a game, or rendering not properly. If you have problems,
> don't call the technical support. Most NT-DX game compatibility lists I
> have seen look badly outdated.
>
> You even contradicted yourself saying "very few" about DirectX5 games
> and now you say "many" DirectX5 games. Hahaha.
>
> >
> >You can get DX 5.0 for NT 4.0 if you need it though...
> >
> >But, the most killer games run on NT 4.0 right now, like Quake II,
> >Hexen II, Unreal (the most ultimate game ever) , StarCraft, etc.
>
> Unreal and Quake II is ugly compared to Forsaken, Incoming and Turok,
> which runs only in Windows 9X.
Again, especially when you are talking about Turok and Forsaken, I believe
that Unreal looks better, by far. The only one close is Incoming.
I would recommend either if you wish to show off your machines power,
though.
> Starcraft is a strategy game that does not stress graphics. Wargames
> does, and it does not run on NT.
>
> The good games that run on NT right now are those courtesy of 3DFX
> Glide. If your video card doesn't have OpenGL ICDs for it that can run
> under NT , sorry.
>
> Descent Freespace ships with a GLIDE version---this is the only reason
> why this runs on NT. You can't run the Direct3D version. It's good
> enough, but it's not Wing Commander Prophecy---which does not run on NT
> at all. DF is one of few recent titles that runs on NT---for every one,
> there must be at least five or more that won't run on NT.
I've got to call you out here. :-) I've got both WC:Prophecy and
Descent:Freespace and Descent, IMO, looks and plays better.
> Glide itself has a doubtful future because it only works directly with a
> 3DFX card, and a lot of 3D cards out there are not made by 3DFX. Doing
> Glide only games means that your D3D accelerator (i740, MGA G200, NVidia
> Riva, Permidia Rendition, etc,.) won't have direct hardware support
> (unless you have to use a D3D->Glide wrapper, which adds overhead).
> Games that support Glide have to support D3D or they won't be so
> marketable, but many more just don't bother with Glide.
However, 3dfx has a performance that none of the others you mention can
match. For hardcore gaming, 3dfx is king and that, as you say, works under
NT.
The rest just don't stack up, so I must disagree with your assertion that
Glide has a doubtful future.
>
>
> >
> >Note that these are some of the latest games you can buy... most
> >developers aren't touching DX 5.0/6.0 until more people get the
> >version pre-installed.
>
> Hahaha. You are staying in Japan too long. You should check the
> shelves---DirectX5 required in about every new game coming out since
> LAST YEAR.
> People are now talking about DirectX6 optimized games, as D3D on
> DirectX6 is much more improved, and if you have an AMD K6-2, 3DNow SIMD
> on DirectX6 can greatly boost the performance of the 3D game. This is
> another reason why developers are so enticed for developing for D3D
> 4.06.
>
> >
> >All of the above run with 3D acceleration under NT using the mini
> >OpenGL 3D drivers.
> >
> >It's pretty fun... think I'll play a little Quake II CTF right now!
> >
>
> And there are many games that won't by the way.
>
> Rgds,
>
> Chris
>
Indeed, but NONE of those work under OS/2.
Robato Yao
>In article <Yqzh521VLZKl-p...@slip129-37-55-119.ca.us.ibm.net>, osb...@deletemeibm.net (Bob O) wrote:
>>On Sun, 6 Sep 1998 15:29:15, sden...@san.rr.com (Steven C. Den Beste)
>>>Each time I defrag, the programs I use most commonly are optimized to load
>>>as fast as possible. Does OS/2 do this?
>>Yes, but you don't have to defrag for the OS/2 optimizations to be in
>>effect.
>
>No.
>
>HPFS's optimizations are oriented toward limiting the number of
>(based in large part on Intel's work in this area I believe) go a bit beyond
>that in actually optimizing the layout of programs and files on the disk
>to speed loading based on how the machine is used. Note that means
>may actually wind up fragmented on the disk because the application
>itself will load more efficiently that way.
If you ever install OS/2 on a clean FAT disk, and then boot a floopy
to run SCANDISK to let you see the cluster/sector map, you can see that
OS/2 installs itself deliberately (strategically) fragmented all over
the disk. OS/2 also locks the sectors so a FAT defragger like Scandisk
cannot change or relocate the contents of a choosen sector, so the
deliberate fragmentation is permanent. Years later, even after heavy
use, the same sectors stay locked and the original scattered pattern
stays the same.
For that matter, it didn't make sense to me to defragment OS/2 FAT
drives as well, since the deliberate scattering means that it affects
the contigiouty of nonessential files, and that even if you defrag an
OS/2 FAT disk, you really don't get any benefit, since the files are not
really defragged, just rearranged but still fragmented. Nonetheless, I
still get an awfully good performance out of my OS/2 FAT drives, which I
specifically format so, so Windows partitions can read and exchange data
with.
Robato Yao
>On 5 Sep 1998 23:56:42 GMT, cro...@kuentos.guam.net (Robato Yao)
>wrote:
>
>>>On Sat, 5 Sep 1998 14:50:00 -0700, josco <jo...@sea.monterey.edu> added
>>>this to the Deja-News archive:
>>>
>>>
>>>I don't quite understand how you can describe it as NT technology when it
>>>works on Win 98 but doesn't yet work on NT. For that matter, I don't
>>>believe it's on CE yet, either.
>>>
>>>Each of those target operating systems requires its own version of DirectX
>>>6. The one for Win 98, developed specifically *for* Win 98, was just
>>>released and I have downloaded and installed it. That sounds a great deal
>>>to me like Win 98 development.
>>
>>I am not completely sure if the current DirectX6 is specifically for
>>Windows 98 only.
For a Windows blowhard, you got to be the most ignorant even on Windows
issues.
>
>I don't think it is only for 98. It is in NT 5.0 as well.
>
Wrong. NT 5.0 is not out yet and DirectX5 on NT is still in beta.
(Looks like a frozen and unfinished project already. That beta still
has bugs that affect the playing of many games.) DirectX5 also requires
many drivers optimized or specific to it as well---and NT drivers are
not optimized or designed for DirectX use. DX5 for NT is simply not the
same as DX5 for Windows 9X---really different animals.
Nobody is expecting a timely release of NT 5, which is a monster project
out of control. The truth is DirectX is so specific to Windows 9X
architecture, that the NT version of DirectX is a really different
animal and is simply not the same DirectX. DirectX does a lot of
direct hardware manipulation, and DX under NT means totally different
game because you either have to access the HAL or DX has to break past
NT's architectural model. DX itself does not fit NT's architectural
model, just as direct hardware access by DOS or Windows 3.1 is not
allowed by NT. For a common DX model to run on NT, NT's architecture
would have to be compromised for it, and Microsoft had to put priority
on another issue---making NT's architecture scalable.
>But, most games still only use mostly directx 2.0/3.0 functionality
>anyway. Very few games even use DX 5.0.
Bullshit. Just about every new Windows game requires DirectX5
functionality since last year.
DX3 is not even considered a reliable API for games playing.
>
>Heck, many DX 5.0 games run under NT 4.0. This should be impossible
>since NT has only DX 3.0 available for it.
Bullshit. By far most DX 5.0 games actually don't run on NT 4 with SP3.
They are not supported in NT 4, many would not install under NT 4, and
those that do run, often don't so reliably enough. Like crashing in
the middle of a game, or rendering not properly. If you have problems,
don't call the technical support. Most NT-DX game compatibility lists I
have seen look badly outdated.
You even contradicted yourself saying "very few" about DirectX5 games
and now you say "many" DirectX5 games. Hahaha.
>
>You can get DX 5.0 for NT 4.0 if you need it though...
>
>But, the most killer games run on NT 4.0 right now, like Quake II,
>Hexen II, Unreal (the most ultimate game ever) , StarCraft, etc.
Unreal and Quake II is ugly compared to Forsaken, Incoming and Turok,
which runs only in Windows 9X.
Starcraft is a strategy game that does not stress graphics. Wargames
does, and it does not run on NT.
The good games that run on NT right now are those courtesy of 3DFX
Glide. If your video card doesn't have OpenGL ICDs for it that can run
under NT , sorry.
Descent Freespace ships with a GLIDE version---this is the only reason
why this runs on NT. You can't run the Direct3D version. It's good
enough, but it's not Wing Commander Prophecy---which does not run on NT
at all. DF is one of few recent titles that runs on NT---for every one,
there must be at least five or more that won't run on NT.
Glide itself has a doubtful future because it only works directly with a
3DFX card, and a lot of 3D cards out there are not made by 3DFX. Doing
Glide only games means that your D3D accelerator (i740, MGA G200, NVidia
Riva, Permidia Rendition, etc,.) won't have direct hardware support
(unless you have to use a D3D->Glide wrapper, which adds overhead).
Games that support Glide have to support D3D or they won't be so
marketable, but many more just don't bother with Glide.
>
>Note that these are some of the latest games you can buy... most
>developers aren't touching DX 5.0/6.0 until more people get the
>version pre-installed.
Hahaha. You are staying in Japan too long. You should check the
shelves---DirectX5 required in about every new game coming out since
LAST YEAR.
People are now talking about DirectX6 optimized games, as D3D on
DirectX6 is much more improved, and if you have an AMD K6-2, 3DNow SIMD
on DirectX6 can greatly boost the performance of the 3D game. This is
another reason why developers are so enticed for developing for D3D
4.06.
>
>All of the above run with 3D acceleration under NT using the mini
>OpenGL 3D drivers.
>
>It's pretty fun... think I'll play a little Quake II CTF right now!
>
And there are many games that won't by the way.
Rgds,
Robato Yao
>On Sat, 05 Sep 1998 23:40:01 GMT, to...@tkpowers.com (Todd) added this to
>the Deja-News archive:
>
>>On 5 Sep 1998 23:56:42 GMT, cro...@kuentos.guam.net (Robato Yao)
>>wrote:
>>
>>>In <35f3b2ba.22937490@news-server>, sden...@san.rr.com (Steven C. Den Beste) writes:
>>>>On Sat, 5 Sep 1998 14:50:00 -0700, josco <jo...@sea.monterey.edu> added
>>>>this to the Deja-News archive:
>>>>
>>>>>On Sat, 5 Sep 1998, Steven C. Den Beste wrote:
>>>>
>>>>I don't quite understand how you can describe it as NT technology when it
>>>>works on Win 98 but doesn't yet work on NT. For that matter, I don't
>>>>believe it's on CE yet, either.
>>>>
>>>>Each of those target operating systems requires its own version of DirectX
>>>>6. The one for Win 98, developed specifically *for* Win 98, was just
>>>>released and I have downloaded and installed it. That sounds a great deal
>>>>to me like Win 98 development.
>>>
>>>I am not completely sure if the current DirectX6 is specifically for
>>>Windows 98 only.
>>
>
>customized to each of them. The customized version for Win 95 was released
>in July. The customized version for Win 98 was released a couple of weeks
>ago. The customized version for NT won't be available for quite some time.
>
>>But, most games still only use mostly directx 2.0/3.0 functionality
>>anyway. Very few games even use DX 5.0.
>
>them come with DX5 on them and use it. In most cases game developers have
>been designing for the tip version of DirectX. Even if they don't use the
>features of the new releases, each new release of DirectX has included more
>drivers and has included speed increases for the older capabilities.
>
>>Heck, many DX 5.0 games run under NT 4.0. This should be impossible
>>since NT has only DX 3.0 available for it.
>
>the speedups in DirectX 5 for functions available in DirectX 3. But if
>they're not there, the program will still run -- but not as fast.
..and not as reliable either.
>
>>You can get DX 5.0 for NT 4.0 if you need it though...
>>
>>But, the most killer games run on NT 4.0 right now, like Quake II,
>>Hexen II, Unreal (the most ultimate game ever) , StarCraft, etc.
>>
>>Note that these are some of the latest games you can buy... most
>>developers aren't touching DX 5.0/6.0 until more people get the
>>version pre-installed.
>
>DirectX is routinely included on the install disk with various games.
>
A lot of games would check for DX version numbers and ask you to install
the right one or don't run at all.
Robato Yao
>On 6 Sep 1998 03:55:19 GMT, bo...@wasatch.com (Bob Hauck) added this to the
>Deja-News archive:
>
>>In article <35f2c3a5.27269538@news-server>,
>>> defrags,
>>
>>age or something?
>
>
>fragmented; the last couple of times I defragged it was precisely to take
>advantage of this feature.
>
>as fast as possible.
>
>Does OS/2 do this?
>
Not sure. The patterns created by an OS/2 FAT defragger look a bit
different from a DOS FAT defragger (SCANDISK can be used on an OS/2 FAT
disk). OS/2 also seems to deliberately scatter the system files all
over the disk, already knowing full well what pattern is best for
optimization (and locks them).
Bob O
wrote:
>In article <Yqzh521VLZKl-pn2-ehYUzvS0TC4O@slip129-37-55-
>149.ca.us.ibm.net>, osb...@deletemeibm.net says...
>> On Sun, 6 Sep 1998 21:24:35, dmccoy@REMOVE_MEmnsinc.com (David H. McCoy)
>> wrote:
>>
>> >In article <Yqzh521VLZKl-pn2-GbOWpzXQrLEE@slip129-37-55-
>> >149.ca.us.ibm.net>, osb...@deletemeibm.net says...
>> >> >
>> >> >OK, but you still were incorrect in responding to Den Beste. ;-)
>> >>
>> >> Oh you mean my comment that "Yes OS/2 does that". You are right, I
>> >> should have said "Yes OS/2 optimizes disks but in a different manner".
>> >>
>> >> Bob O - Computing for fun
>> >>
>> >>
>> >
>> >OS/2 does not optimize the disk with respect to file loading, which was
>> >clearly Steven's question. Everything you have mentioned is also done by
>> >NTFS, so OS/2 has nothing special there.
>> >
>> >You were just wrong.
>>
>> No David. We were talking about Windows 98 not NT.
>
>I know, which is why I said "also."
>
>> I am sure NTFS is a fine file system but it was designed with servers in
>> mind. Properly NTFS should be compared to the 32 bit HPFS.
>
>Why on earth for? If we are discussing clients or workstations, both are
>available on their respective clients, so all client features are up for
>grabs.
Sure they are but they are only features of any real interest to server
users.
>> A better question might be why isn't NTFS available for Win98? All I
>> can do is guess, but my guess is that Win98 still uses DOS structures
>> for file systems and NTFS is just too big of a file system to load into
>> DOS and still have room to run whatever else MS runs there in Win98.
>
>How much memory does NTFS consume? No guessing, let's stick to fact and go
>from there.
>I would say that it is more likely that Win95/98 doesn't support NTFS for
>security reasons.
Thats an interesting theory but why would security make it technically
infeasible unless it is a memory restriction? Or are you saying that
the security features are not desirable in the client? But then you
have to ask why put it in NT workstation?
I don't know how much memory NTFS consumes but I heard a figure of 512K
somewhere along the way from a source I don't recall and thus I cannot
offer it as a fact. If so, that is certainly runnable in DOS but not
together with Windows 3.1 and probably much less with Win 95.
Forrest Gehrke
>
> Greed and egos will stop MS from rejoining IBM and going back to OS/2.
> MS would have to admit that they cannot competitively finish building an
> enterprise class version of NT 5.0.
Egos and also bent noses, yes! But greed, no. MS is passing up
lots of bucks by not adopting your suggestion, Joe; both in the
additional software sold and the reduction in losses they
would save by tossing in the towel with NT v5.0.
Not only would MS get back help from IBM but in one fell
swoop they'd get DOJ off their backs.
//
Bob O
wrote:
Well you have to first of all correctly envision how a hard drive works.
A harddrive does not have to "search" a disk incrementally for an
address, it can just go to an address depending upon how it is
instructed to move. Seek time is affected by how far the heads have to
travel to reach an address. OS/2 effectively puts it partition table
not at the front of the disk but in the middle of it. Thus it can move
the heads in two directions to locate a file instead of one as does a
FAT partitioned disk. That effectively makes it able to cover twice the
disk territory in the same period of time. Putting data and partitiion
tables at the front of the disk is simply obsolete technology. This
little speed up mechanism was discovered back in the mid 80's but
Windows technology still holds to the older much slower methodology.
OS/.2 uses an intelligent table structure called B-Trees or Balanced
Tree. Instead of a plodding lookup through an unsorted linear table,
HPFS quickly traverses the B-Tree structure to find data.
>If you're referring to the efficiency of usage, don't forget that FAT uses
>different cluster sizes with different sizes of drives. If a FAT drive is
>relatively small, the cluster size can be such that efficiency is really
>quite good. It's not a given that HPFS is twice as efficient as FAT (in
>fact, I don't believe it's ever going to be that good in practice).
No efficiency is usuage is not what I was referring to. Although that
does favor HPFS also.
>For instance, a 256M FAT partition uses a 4K cluster. An HPFS partition
>that size isn't going to have much of an advantage in wastage/efficiency.
We just finished discussing that. We agreed that HPFS wastes about
1/8th as much space as FAT does. You said you don't care as FAT does
not waste enough to concern you as you have plenty of money for hard
disks.
>>Personally I prefer not having to even think about fragmentation or file
>>ordering for that matter. All the reading and writing and maintenance
>>associated with defraggers is just mileage on the machinery and a
>>personal hassle as far as I am concerned. I prefer having a vehicle
>>that just goes and goes and doesn't need a tune up like the new 100,000
>>mile autos out there today. While I am sure you can tune these cars for
>>more performance you no doubt lose the 100,000 worry-free miles.
>>
>>Bob O - Computing for fun
>
>Actually, HPFS partitions *do* get fragmented; all disk structure formats
>are subject to this problem. HPFS isn't immune. It's more a matter of rate;
>it's quantitative rather than qualitative. Some formats become fragmented
>more rapidly than others do.
>
>What's different about HPFS is that IBM didn't provide you with a utility
>to do anything about it. It doesn't mean the problem isn't there.
>
>IBM has managed to make it so that the problem doesn't become severe
>anything like as rapidly as it does on a FAT drive.
>
>But I believe that third parties sell HPFS defraggers. Evidently they
>thought there was a market for them.
HPFS partitions do not seem to get fragmented at a rate faster than the
time it takes me to buy a new harddrive, which I seem to do about once
every two or three years. HPFS partitions really don't start to
fragment until the disk approachs being full as HPFS will try to locate
a spot to write a file completely in one spot before starting to write
to the disk at the first available spot like FAT does. My experience
has been that when the disk gets close to full, I am not far from the
time I start looking for a new harddisk. I suppose in the last couple
of months the disk may be fragmenting but my space availability
indicators become my primary concern, not fragmentation.
>What I don't understand is why no-one is doing the one really simple thing
>which would speed disk operations. The reason I don't understand it is that
>it is easy and essentially free: put the disk control blocks in the center
>of the partition instead of at the beginning. Then the maximum seek
>distance from the FAT/whosis/whatsis demanded by whichever filesystem is in
>use is half the length of the partition. It is literally changed by
>altering a constant in an assign in the formatter and a few other places to
>a calculation based on the size of the partition.
So you don't feel that the partitioning methodology of HPFS does not do
this? HPFS keeps its directory information near the center of the drive
to reduce head movement. FAT keeps its directory information at the
front of the disk which results in excessive disk head movement. This
has been the case for HPFS since it was invented in the 80's. So wish
no more.