Is Sun any better than Microsoft??
Mark Mealman
I'm a big Java fan, really. I've even been studying my ass off the past
few weeks to take the Java cert programming exam.
But I'm really starting to get pissed with Sun's apparent attitude that
"100% Pure Java" means Windows NT/95 and Solaris(solaris being the
"preferred" development OS, I'm sure).
Sun only seems to care about porting the JDK, JRE, Swing, etc.
libraries to the above three OSes. Everyone else, it seems, has to wait
patiently for dedicated programmers to donate their free time and port
the libs over(thus "expanding" Sun's vision of write once, run
anywhere).
Java Workshop, which is supposedly written in Java, is also available
only for Windows NT/95 and Solaris. Again, a "workaround" was in place
to get it to work on other systems which had (ported) JDKs on them,
until the workaround itself got blown to hell when the final product was
released.
Then there's HotJava, Sun's shining example of a "100% Pure Java"
application which, by the way, you can only download and install if you
have a Windows NT/95 or Solaris machine.
WTF?!?!
It's bad enough that private developers have to port the friggen JDK
over to other OSes without Sun apparently deciding these OSes don't even
exist.
What next? A System.WhatIsOS() call embedded into the core of the JDK
that forces a System.exit() if it's not a Windows or Solaris machine???
Mark
Peter van der Linden
>Sun only seems to care about porting the JDK, JRE, Swing, etc.
>libraries to the above three OSes. Everyone else, it seems, has to wait
>patiently for dedicated programmers to donate their free time and port
>the libs over(thus "expanding" Sun's vision of write once, run
>anywhere).
So it's not enough that you can get all this software for free for
the high volume platforms? You want to complain and kick up a storm
because Javasoft won't spend hundreds of thousands of dollars to port
it to your grand-dad's Frobutronic-2000?
It's not just the money -- it's the opportunity cost. When the programmers
are porting the JDK to a system that only a few use, they are not able to
improve the implementation on the high volume platforms. What basis would
you recommend for prioritising, if you reject "benefits the most people"?
--
Peter van der Linden, Java Programmer's FAQ
Certified Java Programmer. ---> http://www.best.com/~pvdl
Roy Murphy
Mark Mealman (la...@mail.fwi.com) screamed into the ether:
: But I'm really starting to get pissed with Sun's apparent attitude that
: "100% Pure Java" means Windows NT/95 and Solaris(solaris being the
: "preferred" development OS, I'm sure).
100% Pure Java" has nothing to do with platform. It is a platform
independant series of tests.
: Sun only seems to care about porting the JDK, JRE, Swing, etc.
: libraries to the above three OSes. Everyone else, it seems, has to wait
: patiently for dedicated programmers to donate their free time and port
: the libs over(thus "expanding" Sun's vision of write once, run
: anywhere).
How many platforms do you expect Sun to support? That's why they have
licensees for many other operating systems. If MS weren't a hostile
licensee, and if they used a Sun code base, Sun wouldn't even have to
produce a Windows JDK. As it is, MS is hostile and Windows represents
such a huge installed base, they are forced to directly support it.
: Java Workshop, which is supposedly written in Java, is also available
: only for Windows NT/95 and Solaris. Again, a "workaround" was in place
: to get it to work on other systems which had (ported) JDKs on them,
: until the workaround itself got blown to hell when the final product was
: released.
: Then there's HotJava, Sun's shining example of a "100% Pure Java"
: application which, by the way, you can only download and install if you
: have a Windows NT/95 or Solaris machine.
I have located and installed a version of HotJava that runs on the
Linux port of the JDK. HotJava runs on a lot of other JVM's. It is
provided under license to to several NC OEMs. Though I agree that the
distribution method sucks. A .tar.gz should be made available.
: It's bad enough that private developers have to port the friggen JDK
: over to other OSes without Sun apparently deciding these OSes don't even
: exist.
: What next? A System.WhatIsOS() call embedded into the core of the JDK
: that forces a System.exit() if it's not a Windows or Solaris machine???
Maybe on the MS version. But porters of the JVM would just strip it out.
But really, how can you campare Sun and Microsoft? Sun has published
the complete specification for Java and the JVM. Anyone can produce a
Java-compatable product because its all specified. In contrast, how
much of the Windows 32 API has been fully documented? There are still
many important, undocumented functions that MS never intends to
document. As Sun discovered with WABI, MS is not very happy with
anyone who wants to make a Windows-compatable OS. Sun is very happy to
have others make Java-compatable JVMs.
--
Roy Murphy \ "For a successful technology, reality must take precedence
mur...@panix.com \ over public relations, for Nature cannot be fooled"
\ R.P. Feynman
Philip Brown
On 18 Dec 1997 17:34:53 -0500, mur...@panix.com wrote:
>
>I have located and installed a version of HotJava that runs on the
>Linux port of the JDK. HotJava runs on a lot of other JVM's. It is
>provided under license to to several NC OEMs. Though I agree that the
>distribution method sucks. A .tar.gz should be made available.
On the other hand, if you have "jar", which you should have, if you have
a decent JVM... you have something to unpack .zip files.
--
[trim the no-bots from my address to reply to me by email!]
--------------------------------------------------
"initiating.. 'getting the hell out of here' maneouver" - Lennier, babylon5
Raffael Cavallaro
Peter van der Linden wrote in message <67c7uu$5oq$1...@shell15.ba.best.com>...
>So it's not enough that you can get all this software for free for
>the high volume platforms? You want to complain and kick up a storm
>because Javasoft won't spend hundreds of thousands of dollars to port
>it to your grand-dad's Frobutronic-2000?
We're talking about Mac/OS and Linux, which between them have about 75
million active users, worldwide, on some 30 million machines. Is this what
you mean by "grand-dad's Frobutronic-2000?"
You can't have it both ways, Peter. Either Java is WORA, and Sun supports at
least the major platforms (and that had better include Mac/OS and Linux) or
Sun doesn't really support anything other than Win32 and Solaris, and Write
Once, Run Anywhere is a joke.
Which is it?
Raf
Tony Juricic
Roy Murphy <mur...@panix.com> wrote
> 100% Pure Java" has nothing to do with platform. It is a platform
> independant series of tests.
And COM works fine on all platforms. Yeah. And those tests
are supposed to be done by little green guys who have enough
green paper to make Sun happy.
> How many platforms do you expect Sun to support?
All. Cheaper licenses and test suites would make it possible.
> But really, how can you campare Sun and Microsoft? Sun has published
> the complete specification for Java and the JVM. Anyone can produce a
> Java-compatable product because its all specified. In contrast, how
> much of the Windows 32 API has been fully documented?
People who never did any programming for living, would-be-
consultants, article and book writers are worst Java defenders.
Windows 32 API is lousily documented but it is *much more*
documented than Sun's Java. You have to work to discover that,
though.
> Sun is very happy to have others make Java-compatable JVMs.
They must be as happy as is your employer.
Tony
Bjørn Vaggen Konestabo
[Raffael Cavallaro]
| You can't have it both ways, Peter. Either Java is WORA, and Sun supports at
| least the major platforms (and that had better include Mac/OS and Linux) or
| Sun doesn't really support anything other than Win32 and Solaris, and Write
| Once, Run Anywhere is a joke.
But Java _has_ been ported to Linux and I'm sure it has been ported to
Mac/OS aswell. Why do you claim that Sun should be the one doing this?
--
Every action has an equal and opposite satisfaction.
Mark Mealman
Philip Brown wrote:
> On 18 Dec 1997 17:34:53 -0500, mur...@panix.com wrote:
> >
> >I have located and installed a version of HotJava that runs on the
> >Linux port of the JDK. HotJava runs on a lot of other JVM's. It is
> >provided under license to to several NC OEMs. Though I agree that the
> >distribution method sucks. A .tar.gz should be made available.
>
> On the other hand, if you have "jar", which you should have, if you have
> a decent JVM... you have something to unpack .zip files.
>
I'd agree with you except HotJava 1.1 doesn't come in zip format. Only Win
.exe and Solaris self-extracting archives.
It's not that creating easily installed binaries is wrong, it's just that it'd
be nice to be access Sun's grand example of a Java application without having
to own one of three OSes.
Mark
Mark Mealman
Peter van der Linden wrote:
> >Sun only seems to care about porting the JDK, JRE, Swing, etc.
> >libraries to the above three OSes. Everyone else, it seems, has to wait
> >patiently for dedicated programmers to donate their free time and port
> >the libs over(thus "expanding" Sun's vision of write once, run
> >anywhere).
>
> So it's not enough that you can get all this software for free for
> the high volume platforms? You want to complain and kick up a storm
> because Javasoft won't spend hundreds of thousands of dollars to port
> it to your grand-dad's Frobutronic-2000?
>
Your point is moot. Linux/Mac/OS2 are not my grand-dad's Frobutronic 2000.
Each arguably reaches a larger market than Solaris. So should Sun drop the
Solaris distribution in favor of, say, Mac?
Of course not. Sun created Java in part to promote Solaris, but I'd wish
they'd be a little more generous in their support of "rogue ports".
After all, more widespread use of Java means more applications that run on
Solaris.
> It's not just the money -- it's the opportunity cost. When the programmers
> are porting the JDK to a system that only a few use, they are not able to
> improve the implementation on the high volume platforms. What basis would
> you recommend for prioritising, if you reject "benefits the most people"?
>
Valid point. My main bitch isn't with the base JDK. Unfortunately I've grown
quite accustomed to relying on the donated time of bright people who port the
drivers, applications, etc I use over to my fringe OS.
But is there any reason Java Workshop won't run on my 1.1.5 JDK? Is there any
reason I can't download HotJava 1.1 and install it on my machine?
HotJava is, after all, Sun's prime example of what a Java application should
be. So where's the platform independent .jar of it??
How about swing? It's pure 100% Java, yet the install instructions(and
download compressed formats) are still:
A. If you have Solaris, do this.
B. If you have Windows NT, do this.
C. If you have Windows 95, do this.
Mac users, OS/2 users, Linux users... either we're so bright we don't need
install instructions(generalized ones would be nice), or I guess several
million users just aren't a big enough "niche" market for Sun to bother with.
Does "Write Once, Run Anywhere" really mean "any computer, anywhere" to Sun?
Or is it just "our OS" and, "Oh ya, that Wintel organization we can't ignore"?
Mark
Tim N. van der Leeuw
Bjørn Vaggen Konestabo wrote:
>
> [Raffael Cavallaro]
> | You can't have it both ways, Peter. Either Java is WORA, and Sun supports at
> | least the major platforms (and that had better include Mac/OS and Linux) or
> | Sun doesn't really support anything other than Win32 and Solaris, and Write
> | Once, Run Anywhere is a joke.
>
> But Java _has_ been ported to Linux and I'm sure it has been ported to
> Mac/OS aswell. Why do you claim that Sun should be the one doing this?
>
Since it has been ported to those OSes, why not make HotJava,
JavaWorkshop, and Swing versions for those platforms? Shouldn't be too
much pain to add that to the distributions.
Adding Linux/MacOS distributions of the JVM otoh would be a pain. I
can't blame them for leaving that to Apple or the Linux community. Well,
Apple was going to include their own Java runtime anyways, so there was
little point -- But the rest of the JDK, ie the compiler and debugger,
they could have made those readily available. And having a Solaris
version of JDK/JRE, how much work would it be to port it to Linux -
another Unix?
--Tim
> --
> Every action has an equal and opposite satisfaction.
I like that one ;)
Joseph T. Adams
Is Political Party A, which isn't in power, any better than Political
Party B, which currently does hold power?
Maybe, maybe not; but Party B, by virtue of being in power, can do a
LOT more damage.
If Party B, in power, abuses its power, it can reasonably be expected
that some of those who oppose its abuse of power will support Party A.
Even if it is not in fact better.
Now suppose both Party A and Party B have a long and not so glorious
history of abusing power. They each take turns doing it, and the
public is getting screwed no matter who temporarily holds power.
Enter Party C: in the political realm, the Libertarian Party; in the
computer realm, Open Standards (including but not limited to non-Sun
supporters of Java).
The basic premise of Party C is that it's wrong for ANY party, even
itself, to impose its will on others. It proposes to change the
fundamental structure of the whole system, so that there hopefully
will be both real competition AND autonomy for the common folks who
make up the great majority of the political and computer landscape.
This way, even though not only Party A and Party B but even Party C
may suffer from the universal human tendency to abuse power, most of
the power continues to reside with the People, and it is difficult for
anyone to control and manipulate and directly or indirectly commit
force or fraud against them.
Liberty and open standards benefit almost everyone. They don't
benefit those whose livelihood depends on the status quo, those whose
hold on political or economic power rests on the systematic and
"legal" exercise of force and/or fraud. It is to be expected that
those in the latter group, who are numerically small but control most
of society's resources, will oppose those in the former. It's also to
be expected that they will try to "divide and conquer" the opposition
in order to continue to dominate. This is why even though liberty and
open standards are such good ideas, they very seldom actually happen
in practice.
Joe
Mats Olsson
In article <01bd0c3e$9f1fa9c0$55346020@pp040219>,
Tony Juricic <to...@magg.net> wrote:
>People who never did any programming for living, would-be-
>consultants, article and book writers are worst Java defenders.
>Windows 32 API is lousily documented but it is *much more*
>documented than Sun's Java.
Eh... more in terms of amount of text, certainly. In term of
information contents in relation to how much you need to know, I'd
say not - at least in most areas.
You have a specific example in mind?
/Mats
Roy Murphy
Tony Juricic (to...@magg.net) screamed into the ether:
: Roy Murphy <mur...@panix.com> wrote
: > 100% Pure Java" has nothing to do with platform. It is a platform
: > independant series of tests.
: And COM works fine on all platforms. Yeah. And those tests
: are supposed to be done by little green guys who have enough
: green paper to make Sun happy.
Tony, I've been following you're comments here for several months.
From what I can tell, you'd like Sun to give away "100% Java" awards
and JVMs for hundreds of platforms. Get real. "100% Java" is a
marketing sticker like "Windows 95 Compatable". You should expect to
pay for it if you want to use it. But, that still dosn't affect the
fact that while "100% Java" dosn't guarantee that it'll run on every
platform, it does mean that it should run. Nobody can guarantee
bug-free JVMs, just like you can't guarantee bug-free software.
:
: > How many platforms do you expect Sun to support?
: All. Cheaper licenses and test suites would make it possible.
Again, get real. How can Sun ever adequately support even the top 10
platforms. This just takes time away from improving the JDK.
: > But really, how can you campare Sun and Microsoft? Sun has published
: > the complete specification for Java and the JVM. Anyone can produce a
: > Java-compatable product because its all specified. In contrast, how
: > much of the Windows 32 API has been fully documented?
: People who never did any programming for living, would-be-
: consultants, article and book writers are worst Java defenders.
: Windows 32 API is lousily documented but it is *much more*
: documented than Sun's Java. You have to work to discover that,
: though.
Ouch! Are you accusing me of being a book writer?
Windows has great big chunks of it that are not and never will be
documented. Important chunks even. No comparison.
: > Sun is very happy to have others make Java-compatable JVMs.
: They must be as happy as is your employer.
And who do you think that my employer is? And what does that have to
do with the price of Beans in Redmond?
Peter van der Linden
> Either Java is WORA, and Sun supports at
> least the major platforms (and that had better include Mac/OS and Linux) or
> Sun doesn't really support anything other than Win32 and Solaris, and Write
> Once, Run Anywhere is a joke.
You are confused. WORA does not apply to the Java Virtual Machine. That
needs to be individually ported and hand crafted for each individual
platform.
Raffael Cavallaro
Peter van der Linden wrote in message <67e9k0$nc5$1...@shell15.ba.best.com>...
>> Either Java is WORA, and Sun supports at
>> least the major platforms (and that had better include Mac/OS and Linux)
or
>> Sun doesn't really support anything other than Win32 and Solaris, and
Write
>> Once, Run Anywhere is a joke.
>
>You are confused. WORA does not apply to the Java Virtual Machine. That
>needs to be individually ported and hand crafted for each individual
>platform.
Peter, don't try to weasel out of this one. I'm not confused at all. The
post you were replying to made the following point:
"Sun only seems to care about porting the JDK, JRE, Swing, etc.
libraries to the above three OSes [i.e., Win95/NT & Solaris]. Everyone else,
it seems, has to wait
patiently for dedicated programmers to donate their free time and port
the libs over(thus "expanding" Sun's vision of write once, run
anywhere)."
You replied that Sun shouldn't have to support "grand-dad's
Frobutronic-2000."
The fact is, that WORA isn't a real possibility if newer features/libraries
have to wait for some non-Sun party to port them to the platform in
question. Sun's atitude seems to be "We make WORA noises, but we really mean
'Write Once, Run on Win32 & Solaris' and the rest of you will have to port
low level stuff (JDK, JRE, Swing, etc.) if you want it to run on your
platform."
You have personally added fuel to this perception by referring to Linux as
"grand-dad's Frobutronic-2000" - in comp.os.linux.advocacy no less! (you
should look at those headers you know, just 'cause you see it in
comp.lang.java.advocacy don't mean that's the only place your reply is
seen).
Raf
Raffael Cavallaro
Peter van der Linden wrote in message <67e9k0$nc5$1...@shell15.ba.best.com>...
>> Either Java is WORA, and Sun supports at
>> least the major platforms (and that had better include Mac/OS and Linux)
or
>> Sun doesn't really support anything other than Win32 and Solaris, and
Write
>> Once, Run Anywhere is a joke.
>
>You are confused. WORA does not apply to the Java Virtual Machine. That
>needs to be individually ported and hand crafted for each individual
>platform.
Or is it that Sun means Write Once on Win95/NT or Solaris, Run where the
ports are up to date?
Raf
Raffael Cavallaro
Peter van der Linden wrote in message <67e9k0$nc5$1...@shell15.ba.best.com>...
>You are confused. WORA does not apply to the Java Virtual Machine. That
>needs to be individually ported and hand crafted for each individual
>platform.
Or is it that Sun means Write Once on Win95/NT or Solaris, Run where the
JVM ports are up to date?
Raf
Matt Boersma
Tim N. van der Leeuw <tn...@xs4all.nl*.spamprotector*> wrote:
> Since it has been ported to those OSes, why not make HotJava,
> JavaWorkshop, and Swing versions for those platforms? Shouldn't be too
> much pain to add that to the distributions.
Swing 0.61 seems to work ok with the Linux 1.1.4 and MRJ 2.0 current beta
ports. It hasn't been part of the Win32 or Solaris 1.1.x distributions
either, BTW, nor has HotJava or Java Workshop.
I think Sun's strategy of covering two major bases and letting others port
to less strategic platforms makes sense.
Tony Juricic
Mats Olsson <ma...@dtek.chalmers.se> wrote
> You have a specific example in mind?
How is parent/child or owner/owned relation solved in AWT?
Why add() component doesn't set the parent? Or does it?
Or which function does it?
I think I can easily find much more. This is just one that I
happened ( with some other people ) to bump into.
Tony
Tony Juricic
Roy Murphy <mur...@panix.com> wrote
> Tony, I've been following you're comments here for several months.
> From what I can tell, you'd like Sun to give away "100% Java" awards
> and JVMs for hundreds of platforms. Get real. "100% Java" is a
> marketing sticker like "Windows 95 Compatable". You should expect to
> pay for it if you want to use it. But, that still dosn't affect the
> fact that while "100% Java" dosn't guarantee that it'll run on every
> platform, it does mean that it should run. Nobody can guarantee
> bug-free JVMs, just like you can't guarantee bug-free software.
Well, really, let's get real. What was the requirement for "Windows 95
Compatible"? To implement some technologies in your app.
Not an easy thing to do for some legacy apps, true, but there was no
charge. You did not have to buy some compatibility test tools from
Microsoft.
> : > How many platforms do you expect Sun to support?
>
> : All. Cheaper licenses and test suites would make it possible.
>
> Again, get real. How can Sun ever adequately support even the top 10
> platforms. This just takes time away from improving the JDK.
Exactly. That is why they should make it easire and less expensive
for others to do it.
> : > But really, how can you campare Sun and Microsoft? Sun has published
> Ouch! Are you accusing me of being a book writer?
No, how would I know if you are or not? I'm just saying that
there is a definite difference between babbling about
Java advantages and trying to do something today that you can
sell tomorrow.
.... Sun and employer's happines ...
.
> And who do you think that my employer is? And what does that have to
> do with the price of Beans in Redmond?
Be real. Why should Sun be happy if somebody produces
fast and commercially distributable JVM without paying
them license? They can sing songs of support on Web site but
tomorrow there is a new cute package coming out from JavaSoft
that somehow you must support if you want to be compatible.
Just like MS APIs.
Tony
Matt Kennel (Remove 'NOSPAM' to reply)
:Peter, don't try to weasel out of this one. I'm not confused at all. The
:post you were replying to made the following point:
:
:"Sun only seems to care about porting the JDK, JRE, Swing, etc.
:libraries to the above three OSes [i.e., Win95/NT & Solaris]. Everyone else,
:it seems, has to wait
:patiently for dedicated programmers to donate their free time and port
:the libs over(thus "expanding" Sun's vision of write once, run
:anywhere)."
:
:You replied that Sun shouldn't have to support "grand-dad's
:Frobutronic-2000."
:
:The fact is, that WORA isn't a real possibility if newer features/libraries
:have to wait for some non-Sun party to port them to the platform in
:question. Sun's atitude seems to be "We make WORA noises, but we really mean
:'Write Once, Run on Win32 & Solaris' and the rest of you will have to port
:low level stuff (JDK, JRE, Swing, etc.) if you want it to run on your
:platform."
The solution is simple. People ought to pay Sun to have them port the
stuff to the Frobutronic-2000's, like my Linux box.
Unfortunately, I doubt Sun would be amused, considering that Linux (and a few
NT) workstations have replaced what used to be a true-purple Sun shop
here at INLS since 1987 or so.
--
* Matthew B. Kennel/Institute for Nonlinear Science, UCSD
*
* According to California Assembly Bill 3320, it is now a criminal offense
* to solicit any goods or services by email to a CA resident without
* providing the business's legal name and complete street address.
*
Philip Brown
On Fri, 19 Dec 1997 00:35:13 -0500, la...@mail.fwi.com wrote:
>> On the other hand, if you have "jar", which you should have, if you have
>> a decent JVM... you have something to unpack .zip files.
>>
>
>I'd agree with you except HotJava 1.1 doesn't come in zip format. Only Win
>.exe and Solaris self-extracting archives.
Ah. yes, it looks that way.
Except they lie.
Its the old "standard" messy-dos "self-extracting archive" trick,
where you can run the archive if you're lazy.. OR... if you have a
decent version of an unzip program, it will skip the initial 6K junk
executable, and find the plain-ordinary-zipfile that comprises the rest
of the data.
Yes, they should document this.
No, this should not stand in your way ;-)
[Unfortunately though, "jar" isn't one of the aformentioned "decent versions"
:-( ]
>It's not that creating easily installed binaries is wrong, it's just that it'd
>be nice to be access Sun's grand example of a Java application without having
>to own one of three OSes.
Well, fyi, if you were really bored, you could write a simple C prog
(or java prog? :-)
"junk-header extractor", that ran on all platforms,
and junked everything before the bytes
"PK\003\004\n"
and then you would have the pure file suitable to shove through "jar"
:-)
Mindspring
Joseph T. Adams wrote in message <67e1m4$34v$1...@nerd.apk.net>...
>
>Is Political Party A, which isn't in power, any better than Political
>Party B, which currently does hold power?
>
>Maybe, maybe not; but Party B, by virtue of being in power, can do a
>LOT more damage.
>
So say the Nazi party is Political Party A and just about any other
political party in human history is Party B...
This line of argument is really inane. Sun and Microsoft are after the
same thing their motives are the same, and in as much as Sun's position
allows it, their tactics are the same.
Mindspring
Mats Olsson wrote in message <67e2ic$sjr$1...@nyheter.chalmers.se>...
>In article <01bd0c3e$9f1fa9c0$55346020@pp040219>,
>Tony Juricic <to...@magg.net> wrote:
>>People who never did any programming for living, would-be-
>>consultants, article and book writers are worst Java defenders.
>>Windows 32 API is lousily documented but it is *much more*
>>documented than Sun's Java.
>
> Eh... more in terms of amount of text, certainly. In term of
>information contents in relation to how much you need to know, I'd
>say not - at least in most areas.
Strangely, Sun has found that there is enough Win32 documentation necessary
to write a Windows version of the JVM...
Nasser Abbasi
Also, when Sun releases its JDK first to windows (and Solaris )platforms
only, it is actually helping Windows and Microsoft become more popular,
becuase the programmers who want to use the JDK and don't want to wait few
more months for it to be ported by some other third party to their platform,
will swich to windows from say Linux or the Mac just to use the JDK, with
time, they'll spend more and more time on Windows becuase of this.
So, Sun by not releasing the JDK to Linux as well, is making some people
leave Linux to go to Windows (for those who want to program in java and use
the latest JDK's ). So , to some extent, Sun is helping make Unix less
popular.
Nasser Abbasi
Nasser Abbasi
Mark Mealman wrote in message
>Sun only seems to care about porting the JDK, JRE, Swing, etc.
>libraries to the above three OSes. Everyone else, it seems, has to wait
>patiently
good point Mark. I was just thinking the same last night when I downloaded
JDK 1.2 Beta2, it is only avaliable on M$ and Solaris, I would have liked to
see it on Linux also. For now I have to boot into M$ to play with this
latest JDK.
I have Solaris x86, but it will not install on my PC, it is not happy about
something, Sun support have no clue what it is, but that is another story.
Nasser Abbasi
Nathan Hand
Mark Mealman <la...@mail.fwi.com> writes:
> Then there's HotJava, Sun's shining example of a "100% Pure Java"
> application which, by the way, you can only download and install if you
> have a Windows NT/95 or Solaris machine.
>
> WTF?!?!
That is because the install scripts only exist for these two
platforms. However HotJava runs fine on all of the JVMs I've
thrown at it, including the Linux JDK.
--
The sticker on the side of the box said "Supported Platforms: Windows 95,
Windows NT 4.0, or better", so clearly Linux was a supported platform.
Sriram N
Well if a jdk exists for a variety for platforms besides
win32 and solaris, it is the height of idiocy to distribute
a pure java app as a win32 exe or whatever (atleast using
a zip format would be more forgivable).
Recently I had the pleasure of installing a shareware
chat server (called fastchat) and guess what ? The
same gui install should work on any jdk system. Perhaps
they could do a textmode install as well.
It used a java install procedure that looked
suspiciously like the installshield in win32. Turns out
the folks at installshield.com have a solution for
packaging java apps that is truly x-platform. (it
worked fine on my os/2 jdk 1.1.4 system)
I really dont see why javaworkshop or any other
app that claims to be pure java should fail to
use x-platform installation procedures as well.
Sriram
Joseph T. Adams
Mindspring (gsi...@mindspring.com) wrote:
: >
: >Is Political Party A, which isn't in power, any better than Political
: >Party B, which currently does hold power?
: >
: >Maybe, maybe not; but Party B, by virtue of being in power, can do a
: >LOT more damage.
:
: So say the Nazi party is Political Party A and just about any other
: political party in human history is Party B...
: This line of argument is really inane. Sun and Microsoft are after the
: same thing their motives are the same, and in as much as Sun's position
: allows it, their tactics are the same.
I think you missed the point I was trying to make. Some perceive the
necessity of opposing the party in power when it abuses power. The
dangers of the party *not* in power are seldom obvious, unless that
party has been in power recently. (The evils of even the Nazis were
not apparent to the average German until well after they took power,
though anyone who'd read _Mein Kampf_ should have had ample warning.)
And, finally, some - for instance, those who favor liberty in the
political realm and/or open standards in the computer realm - don't
want Sun, Microsoft, IBM, the government, or ANYONE else abusing their
power, by (among other things) establishing proprietary,
non-competitive standards.
If you don't look beyond Sun vs. Microsoft, you aren't seeing the
bigger picture. I'm not necessarily defending *or* bashing either of
these entities, though I've been critical of both in the past; I am
stressing the need for genuine open standards; I applaud progress
toward that goal. I'll oppose the actions of either of them when they
impede it.
Joe
Glenn Sills
I think that perhaps people should give Sun a little bit of a break. There
is an awful lot of "stuff" that needs to be done in order to make "Write
Once Run Anywhere" significantly more than a marketing slogan. Sun has a
finite number of programmers. They can spend their development time writing
the WORA "stuff" or they can spend ther time writing good cross platform
installation programs.
Sriram N wrote in message <67gaft$8...@nuscc.nus.sg>...
Reality is a point of view
+---- pv...@shell15.ba.best.com wrote (18 Dec 1997 14:26:38 -0800):
| >Sun only seems to care about porting the JDK, JRE, Swing, etc.
| >libraries to the above three OSes. Everyone else, it seems, has to wait
| >the libs over(thus "expanding" Sun's vision of write once, run
| >anywhere).
|
| the high volume platforms? You want to complain and kick up a storm
| because Javasoft won't spend hundreds of thousands of dollars to port
| it to your grand-dad's Frobutronic-2000?
|
| are porting the JDK to a system that only a few use, they are not able to
| improve the implementation on the high volume platforms. What basis would
| you recommend for prioritising, if you reject "benefits the most people"?
Solaris is a high volume platform? Microsoft Windows is a high
volume platform. Singular, not plural.
Java(TM) isn't simply Sun's way of keeping MSNT away from Sun's
profits? It is disingenuous to rally mercenary revolutionaries
with empty war cries. And dangerous if you leave them on the
battlefield staring at their empty hands.
100% pure projects, _from Sun_, aren't a myth? Do as Sun says,
not as they do?
When you say improvements do you mean yet more extensions
intended to claim niche market share, or actual fundamental
improvements? From here it looks like marketeers are now in
charge of R&D.
The long term mistake Sun/Javasoft has made is 'open specs,
implementors compete'. Because they will eventually lose on
implementation, even with the road blocks and delays. Once the
advantage of the moving spec is gone Microsoft won't be the
only source of faster, better, cheaper. Microsoft is already
winning in the faster category.
I like Java. Sun/Javasoft is starting to smell a lot like a
Microsoft wannabe. Just like Netscape has in the past. Given
the current directions, actions, and press releases; I look
forward to the separation of the technology and the company.
--
Gary Johnson gjoh...@season.com
Privacy on the net is still illegal.
Cocomax
>
>I like Java. Sun/Javasoft is starting to smell a lot like a
>Microsoft wannabe. Just like Netscape has in the past. Given
>the current directions, actions, and press releases; I look
>forward to the separation of the technology and the company.
The news stories that I have been reading about Sun indicate that Sun is
moving out of the big iron Unix world focusing almost all its energies into
defeating Windows NT with Solarius.
To do so they need to make deals with Intel to aid in porting Solarius to the
Merced Chip and produce cheap work stations in the sub $5,000 range that use
standard PC market parts. Both things they have done.
In order for Solarius to make it as the NT killer they will need programmers
to write apps for it, that is were JAVA comes in, you can program for both
Solarius and Windows at the same time, thus making Solarius useful when people
begin writting real apps in JAVA that are as fast as C++ apps using hotspot
some time in the future. . . yada yada. yada . . .
MAC, Linix ? Sun will not even bother with porting JVM's to those platforms,
why should they? They are not the targets, MS Windows is the target.
Microsoft evil? Of course, the enemy Sun is trying to defeat will always be
considered the evil makers of bad things by Sun and all the fans of Sun.
This is a war of lies and nit picks, by very smart men over very large
dollars, nether Sun or Microsoft are victims, rather they are both fighters,
each with their own armies.
Microsoft can be defeated, but it will take someone smarter than Sun. Hyp,
JAVA rings, slow JVM, that dream of WORA and calling Bill Gates a butt head is
not enough to defeat Microsoft.
If you read carefully why JAVA is such a great language, it is not its lack
of bugs, speed or WORA. . . it is because M$ sucks, C sucks and the only reason
JAVA does not take over the world of programming and solve all problems is
because M$ is standing in the way by not supporting Suns version of native
calls to the OS (that defeats the WORA anyway) or it is M$ is not supporting a
tech that Sun will phase out in six months and breaks backward compatiblity.
JAVA is a very neat idea.
Tim
Rich
> Does "Write Once, Run Anywhere" really mean "any computer, anywhere" to
Sun?
> Or is it just "our OS" and, "Oh ya, that Wintel organization we can't
ignore"?
>
> Mark
>
kill off the Win 95/98 line soon and won't even acknowledge the continued
existence of Win 3.x and DOS. So in the near future, Win NT and Unix will
be the only major platforms and portability won't be that big of a concern.
****************************************
Rich ev...@nfnarg.pbz
Spam Stopper - Apply ROT13 to
unscramble my e-mail address.
****************************************
Charles R. Lyttle
copies of Redhat 5.0 to meet the demand even with the limited shelf
space allocated to non MS products. Caldera seems to be doing well, and
Applix (sp?) ware wasn't just sitting around. It seems that the MS - Sun
thing isn't of much concern outside this news group. Too many people are
running up against the MS brick wall by themselves. And they seem to
have enough savvy to recognize it. I propose that we have more talent
available here than either of the two corporations and what they do is
of little matter. If we want Java, we should make it.
--
Russ Lyttle, PE
<http://www.flash.net/~lyttlec>
Bowie Poag
: Then there's HotJava, Sun's shining example of a "100% Pure Java"
: application which, by the way, you can only download and install if you
: have a Windows NT/95 or Solaris machine.
Psst.. Java Sucks.
A) It isnt platform independant. Java apps require tweaking to get them
to work on different platforms.
B) Its not an open standard. Sun controls it.
C) Its not write-once, run-anywhere -- Its write-15-times, crash-everywhere.
--
o--+--------------------------------------------------------------------+--+--o
|()|Bowie J. Poag | b...@primenet.com | http://www.primenet.com/~bjp |[]|{}|
+==+====================================================================+==+==+
| |
| If Microsoft began to slaughter infants and barbecue them, the press |
| would cover the story as, "Microsoft Invests In Alternative Food Products". |
| |
L+=+=======================================================================+==J
Joseph T. Adams
Bowie Poag (b...@primenet.com) wrote:
:
: Psst.. Java Sucks.
:
: A) It isnt platform independant. Java apps require tweaking to get them
: to work on different platforms.
The *language* supports cross platform code. It does not require it,
and Java implementations to date have not lived up to that promise
with respect to the more difficult and/or rapidly changing portions of
the Java spec (for example, threading, which is difficult to
implement in an OS-independent way). It is likely that once the
graphics libraries have stabilized, Java-based front end applications
truly will be "write once, run everywhere." And that is the niche I
see Java as being most likely to dominate, at least for the near future.
: B) Its not an open standard. Sun controls it.
It is "open" in the sense that it is published and anyone is free to
make their own implementation of it. That is not true of any
competing technology that I'm aware of. It most especially is not
true of Win32, which, otherwise, might have become a universal,
cross-platform, hardware-independent API. (In a technical sense it *is*
true of C++; however, C++ implementations differ enough that writing
cross-platform C++ is tremendously difficult, even compared to Java in
its present, relatively immature state.)
:
: C) Its not write-once, run-anywhere -- Its write-15-times, crash-everywhere.
Yes, it's possible for Java apps and/or VMs to crash, especially if
they're poorly written. The same is true for all languages. Bad
implementations, and bad code, are not necessarily the fault of the
language.
Java is a young language that's taking the world by storm in spite of
many competing languages (not just C++ but also Eiffel, Smalltalk,
Ada, Objective-C, and countless RAD tools like VB) that are all at
least arguably more mature, more stable, and faster. You need to ask
yourself why. Is this all a matter of marketing hype, or does Java
represent at least a realistic promise of changing the very way that
developers think and work and market their products? If the former,
then do you really think the developer community (which has seen many
promising languages come and go) is as naive as it would then appear?
If the latter, then is it possible that we know something you don't?
Bottom line is this: no one knows what Java will become tomorrow,
though all of us who use it do have high hopes for it. We do know
that it offers us, TODAY, something that no other product does:
platform, hardware, OS, and vendor independence. The "WORA" paradigm
has not been totally realized, but Java has made great strides toward
that goal. And it is, today, a valuable, useful and productive tool.
Not for all purposes, but for a large and growing set of problems that
simply require cross-platform solutions.
Joe
Nathan Hand
Bowie Poag <b...@primenet.com> writes:
> Psst.. Java Sucks.
>
> A) It isnt platform independant. Java apps require tweaking to get them
> to work on different platforms.
It is platform independent, even if it does require tweaking. The
design of the language is such that it works across all platforms
and it does this by being well defined (defining bitness and even
endianess for all types) and using standards for representing the
common types (ie unicode for characters).
If Java applications require "tweaking" then that's an indication
of poorly implemented JVM's. It is not the same thing as the Java
language being platform dependent, because Java most certainly is
not dependent on any given platform.
> B) Its not an open standard. Sun controls it.
It *is* an open standard, despite Sun controlling it. If you have
no idea what it means for a standard to be open, as this point of
yours would seem to suggest, then I recommend you expend a little
effort and *learn* what an open standard means.
> C) Its not write-once, run-anywhere -- Its write-15-times, crash-everywhere.
HotJava, a 3.2 compliant web browser, runs flawlessly on all five
JVM's I've tried. I daresay you've never actually tried running a
Java app on different JVM's, but that you're merely spouting what
all good anti-Java proponents spout: FUD, FUD and more FUD.
Simon Kinahan
> >information contents in relation to how much you need to know, I'd
> >say not - at least in most areas.
>
> Strangely, Sun has found that there is enough Win32 documentation necessary
> to write a Windows version of the JVM...
Its different. The JVM docs give enough information to rewrite the thing
from scratch, and a reference implementation to clear up any
ambiguities. The Windows docs only give enough information to
write applications, and in fact there are significant areas where
they don't even fo that (WinExec on 3.1 has some wonderful
undocumented behaviours).
Simon
Simon Kinahan
I think people are missing the point here. Sun provides Java as am
open standard, and anyone who wants to implement it, can implement
it. They only provide there *reference* implementation (which is
the JDK's role in life) on 2 platforms. So what ? You can
even get the source if you want to port it yourself ! This
has no impact on Java's openness or legitimacy as a standard.
It is like saying ANSI C is not a legitate standard, because ANSI
didn't provide an implementation for <insert favourite plaform>.
Expecting SUN to port their implementation to every plaform in
existence seeme extreme. Almost all plaforms have a port of JDK 1.1.4.
What is the problem ?
Simon
Tony Juricic
Simon Kinahan <sim...@cadence.com> wrote
> Its different. The JVM docs give enough information to rewrite the thing
> from scratch, and a reference implementation to clear up any
> ambiguities.
Theoretically yes. In reality rewriting JVM from scratch makes
no sense if you can not provide core classes as well.
And in combination of these two you find a lot of
undocumented behavior.
And any reliance on reference implementation could be
interpreted as reverse engineering and not "clean room".
You certainly can not apply test suites without
licensing them from JavaSoft, along the likes of IBM.
All in all, Sun/JavaSoft did all they could to advertise their
openness while, in practice, licensing from them is the
only realistic choice.
>The Windows docs only give enough information to
> write applications, and in fact there are significant areas where
> they don't even fo that (WinExec on 3.1 has some wonderful
> undocumented behaviours).
There has been quite a few books published on undocumented
Windows behavior. Anything similar regarding JVM could
bring lawsuit from Sun/JavaSoft.
Sun is indeed somewhat better than MS. MS doesn't
license Win32 but there are some known attempts to
port Win32 to UNIX platforms, there were some HP
or Siemens ( I think) specific implementations of
Windows so there must be some licensing going on.
Sun did license Java to many more. That does account
for some difference because, for example, there
are more Java IDEs to choose from on PC.
I say "somewhat" because I see it all as quite weak in
comparison to big PR "freedom and WORA" blab they emit.
Tony
Glenn Sills
Simon Kinahan wrote in message <349E6258...@cadence.com>...
>Mindspring wrote:
>>
>> Mats Olsson wrote in message <67e2ic$sjr$1...@nyheter.chalmers.se>...
>> > Eh... more in terms of amount of text, certainly. In term of
>> >information contents in relation to how much you need to know, I'd
>> >say not - at least in most areas.
>>
>> Strangely, Sun has found that there is enough Win32 documentation
necessary
>> to write a Windows version of the JVM...
>
>from scratch, and a reference implementation to clear up any
>write applications, and in fact there are significant areas where
>they don't even fo that (WinExec on 3.1 has some wonderful
>undocumented behaviours).
Would documentation necessary to rewrite Window NT 4.0 make it any more
useful for the 99.999% of the population uninterested in writing an
operating system?
Glenn Sills
Simon Kinahan wrote in message <349E63B4...@cadence.com>...
I think what people are complaining about is the Sun's marketing of Java.
They are obviously over selling the"Write Once Run Everywhere" and the "100%
Pure" slogans. The implication of both campaigns is that an application
developer can write a Java program that he can actually sell on multiple
platforms. This is clearly not all that true at this time. That's not to
say that Sun isn't working on it, but the bottom line is that Sun is
dependent of a licensees to make those marketing slogans more than just cool
ideas.
Greg Chien
Tony Juricic wrote in message <01bd0f35$aa74b320$b6346020@pp040219>...
>
>Sun is indeed somewhat better than MS. MS doesn't
>license Win32 but there are some known attempts to
>port Win32 to UNIX platforms, there were some HP
>or Siemens ( I think) specific implementations of
>Windows so there must be some licensing going on.
There are Win32/MFC porting tools from Bristol Technology
(http://www.bristol.com) and Mainsoft (http://www.mainsoft.com). There is
also a less expensive one call Willow (don't have the web addr handy). I
suppose IE4 was ported to Un*x with Bristol's tool.
Greg
David Brooks
Nathan Hand <h930...@student.anu.edu.au> writes:
>> B) Its not an open standard. Sun controls it.
>
>It *is* an open standard, despite Sun controlling it. If you have
>no idea what it means for a standard to be open, as this point of
>yours would seem to suggest, then I recommend you expend a little
>effort and *learn* what an open standard means.
I'm curious: what do *you* think an open standard means?
"Open" isn't the same as "published". To be open, a standard must be
change-controlled by a body of unrestricted membership, which has no stake
in any specific implementation. Sun can change the spec of Java whenever
it likes unilaterally, as is its privilege under the current rules (I
heard, unconfirmed, that changes to floating-point handling were in the
small-print of the Merced deal).
Sun understands the distinction. After all, their price for joining OSF and
the CDE project was that OSF should cede ownership of the Motif
specification to X/Open. OSF, despite its name, was also the licensing
source for Motif implementation and that put them the wrong side of the
wall.
(to be clear, I'm interested in this distinction as an ex-employee of OSF
and The Open Group. I'm not expressing an opinion in the context of my
current employment).
--
David Brooks
The views expressed above are mine and not representative of
Microsoft Corporation.
Nathan Hand
"Glenn Sills" <gsi...@mindspring.com> writes:
> Would documentation necessary to rewrite Window NT 4.0 make it any more
> useful for the 99.999% of the population uninterested in writing an
> operating system?
No but it's very necessary for the 0.001 percent who are
interested in writing an operating system. In particular
the WINE developers and Sun's WABI project would both be
ecstatic if Microsoft had documented Win32 well.
However Microsoft has not documented Win32 very well, so
typically you need to get the NT source code in order to
get enough information to make a clone in any reasonable
amount of time (aka SoftWindows 95).
Unfortunately this excludes free clones (WINE), and also
excludes Sun for some unknown reason *cough* *cough*.
Fortunately Sun is very kind and they have released Java
specifications in sufficient detail to make clones, both
free and non-free. The Kaffe project is actually kicking
some serious butt, now you just need free class libs.
r.e.b...@usa.net
In article <67n7ap$t...@news.microsoft.com>,
dbr...@microsoft.com (David Brooks) wrote:
> Nathan Hand <h930...@student.anu.edu.au> writes:
> >> B) Its not an open standard. Sun controls it.
> >It *is* an open standard, despite Sun controlling it. If you have
> >no idea what it means for a standard to be open, as this point of
> >yours would seem to suggest, then I recommend you expend a little
> >effort and *learn* what an open standard means.
vi to support virtually any terminal from a KSR-33 teletype to an Xterm
with a SUN Keyboard.
Ed Gosling experienced first hand the effect of 1000 programmers and a
General Public License (Stallman wrote the GPL to prevent Gosling from
creating proprietary dirivative products without returning them to the
public archive, Gosling Emacs was quickly obsoleted by GNU Emacs.
Sun is acutely aware that there is already a guavac compiler and a kaffe
virtual machine, waiting in the wings if Sun decides it doesn't want to
"play nice". By supporting the Linux implementation of Java with a
modified GPL (gives Sun the right to royalties for commercial sales and
first right to port to Solaris).
> I'm curious: what do *you* think an open standard means?
Vendors have played with this term since the 80's. The working
definition, the one that seems to be effective in actual practice, must
meet the following criteria:
1. The language, protocols, and structures must be documented and
published in a manner which allows retrieval without fee. All changes
should also be available in this manner.
2. A reference implementation - providing the core functionality, is
available in source format that can be compiled on an industry standard
GPL compiler such as GCC or on a compiler that can be compiled using GPL
source code and a GCC compiler.
3. The reference implementation will either not offer GUI features, or
will implement GUI features as an isolated library which can be
implemented in X11/R6 (or earlier) libraries, and can be ported to
Microsoft Foundation classes with little effort.
4. Sample data, programs, clients, servers, or test data generators are
provided in a manner than can be generated as a reference
implementation.
This seems to be the "working definition". Variations that have been
tried include published protocol but no reference models (Motif), but
they are vulnerable to a standard that meets the criteria above (OLIT,
Athena, HP).
> "Open" isn't the same as "published". To be open, a standard must be
> change-controlled by a body of unrestricted membership, which has no stake
> in any specific implementation.
This isn't always true. There are many GNU programs which are managed by
the FSF. The FSF membership is rather loose, but RMS has final say.
There are times when people (the Debian folks), wish it weren't so.
> Sun can change the spec of Java whenever
> it likes unilaterally, as is its privilege under the current rules (I
> heard, unconfirmed, that changes to floating-point handling were in the
> small-print of the Merced deal).
Not exactly. As I mentioned before, there is a GNU implementation called
guavac and kaffe. If Sun oversteps it's authority, the user support
could quickly shift to the GPL implementations, leaving SUN with a nifty
trademark and very little innovation resource. Many of the innovations
on Web servers were implemented in NCSA and Apache long before Netscape
or IIS got them. When Netscape tried to claim exclusive use of SSL,
three GPL implementations were proposed and Netscape "contributed" it's
own SSL implementation.
On the other side of the table. Micrsoft could find itself in a bad
place as well. If the GPL support chooses to, they can release patches
and upgrades to Apache and Arena (or an HTML 4 GPL browser) that would
break both Netscape and IE. Since the code would be distributed freely,
it would take little to break the "Microsoft Monopoly".
Unfortunately, guavac and kaffe do little to help Microsoft, since they
are protected by an even Stricter General Public License that makes
the Sun deal look like a "gift". For one thing, Microsoft is very
unlikely to start contibuting free "reference copies" of Microsoft
Word in order to integrate activeX controls into guavac and kaffe. For
another Micrsoft is very restricted in it's ability to run "UNIX" code.
The most dramatic illustration of this was the ISO OSI protocol. The ISO
decided to charge for distribution, charging $30 per book for over 100
books and specifications. The price to train one programmer to
understand OSI could go as high as $50,000. The supporters of OSI
(primarily DEC, IBM, and AT&T) felt that would protect them from
competitors who had been supporting TCP/IP.
The strategy backfired. Not only was the industry reluctant to support
OSI (even with the government mandating GOSIP), the developers of TCP/IP
began distributing free copies of TCP/IP for various implementations of
UNIX, DOS, and eventually Windowms. The Trumpet TCP/IP stack, coupled
with the Mosaic Browser, turned TCP/IP into the undisputable winner of
the protocol wars. The IETF will need to get similar support to gain the
acceptance of IPv6.
> Sun understands the distinction. After all, their price for joining OSF and
> the CDE project was that OSF should cede ownership of the Motif
> specification to X/Open. OSF, despite its name, was also the licensing
> source for Motif implementation and that put them the wrong side of the
> wall.
OSF (especially IBM, DEC, and HP) was notorious for putting steep
ristrictions and charges on source code, not providing reference
implementations, and trying to "lock out" competitors, including Sun, MIT
(Athena), Berkely, and Linux. Even SCO found itself shuffled between the
"ins" and the "outs" during the early days.
The joke is that X/Open has been trying to put similar restrictions on
the Unix/98 standard, which has resulted in an industry rally around the
FSF gclib and Linux. If Caldera or Red Hat starts trying to get cute and
"close the doors", Slackware, Debian, and S.U.S.E are waiting in the
wings. When he finishes school, Pat Volkerding could be a significant
player. So could any of another 200 people.
Of course, the biggest player of them all is Richard M Stallman, whose
passion, integrity, courage, and ethics have made him the virtual leader
of the worlds largest information systems organization (the Internet,
TCP/IP, most of the Linux Distribution, and most of the contributed
applications on the "supplemental" or "contrib" CD-ROM are under the
costodial care of RMS.
Though he is not a billionaire, RMS has effectively lead, through his
writings, vision, and inspiration, an organization that encompasses 200
million computers world-wide, 2 million developers and administrators
world-wide, and 100,000 Linux developers world-wide. In effect, RMS has
generated an industry worth over $1 trillion. Not bad for an NPO with a
budget of less (much less) than $1 million. Bill Gates just happened to
come to the river with an oil tanker.
Many have tried to stop the FSF with billions of dollars worth of
advertising "hype", providing the press with "press releases" to portray
TCP/IP as unreliable, insecure, slow, and "incapable of handling more
than 2 million users". The press never mentioned that TCP/IP was
designed by the military as a system designed to guarantee the delivery
of the "last strike" AFTER a nuclear haulocaust.
Microsoft has also provided press releases claiming that UNIX is slow,
unreliable, unsupported, and "unusable by the general population". The
press forgot to mention that UNIX was designed by AT&T to manage nearly 1
billion telephone circuits, switches, and datacomm components, performed
as much as 8 times the work as NT with 1/2 the resources, ran free
software that had been contributed by 100,000 people and had been tested
by over 1 million people for over 25 years (approximately 10 million
staff-years of "non-support"), and been used by over 200 million people
(over 2/3 of the servers on the Internet run some variant of UNIX,
including Linux). They even forgot to mention that Linux distributions
were funded, managed, and coordinated by 5 different tightly held
companies. The press forgot to mention that Micrsoft isn't doing a UNIX
because it CAN'T (Microsoft sold all future rights to the UNIX and UNIX
compatibility market to SCO, along with Xenix). Bill Gates still owns
25% of SCO, but can never gain controlling interest. Even if he did, he
couldn't get control of FSF or GPL.
> (to be clear, I'm interested in this distinction as an ex-employee of OSF
> and The Open Group. I'm not expressing an opinion in the context of my
> current employment).
The OSF came up with some very good standards, including Motif, OSF1,
and DCE. The problem was that Sun and Linux were offering http, rpc,
onc, and nis. They weren't technically superior, but they were available
under terms that made further support and innovation (CORBA, html3, Java,
OpenLook, and 3D Athena) more practical.
If you really want to serve your masters, you should take the time to get
intimately familiar with Linux, especially Red Hat 5.0 and Caldera 1.1 to
the point where you know every strength and weakness. Many technophiles
will be getting Linux this Christmas and the the usual rhetoric referring
to the "clunky character mode interface" and the "incredibly difficult
installation and configuration" isn't going to help your credibility.
Windows will be "sharing the drive" with a new neighbor - Linux.
> David Brooks
> The views expressed above are mine and not representative of
> Microsoft Corporation.
read "the Cathedral and the Bazaar"
http://locke.ccil.org/~esr/writings/cathedral.html
Rex Ballard
http://www.access.digex.net/~rballard
-------------------==== Posted via Deja News ====-----------------------
http://www.dejanews.com/ Search, Read, Post to Usenet
Todd Bandrowsky
Peter van der Linden <pv...@shell15.ba.best.com> wrote in article
<67c7uu$5oq$1...@shell15.ba.best.com>...
> >Sun only seems to care about porting the JDK, JRE, Swing, etc.
> >libraries to the above three OSes. Everyone else, it seems, has to wait
> >patiently for dedicated programmers to donate their free time and port
> >the libs over(thus "expanding" Sun's vision of write once, run
> >anywhere).
>
Ok, that argument works quite well for saying, why not just buy Windows.
If that's the line, then Java is pointless.
Nathan Hand
dbr...@microsoft.com (David Brooks) writes:
> Nathan Hand <h930...@student.anu.edu.au> writes:
> >> B) Its not an open standard. Sun controls it.
> >
> >It *is* an open standard, despite Sun controlling it. If you have
> >no idea what it means for a standard to be open, as this point of
> >yours would seem to suggest, then I recommend you expend a little
> >effort and *learn* what an open standard means.
>
> I'm curious: what do *you* think an open standard means?
Any ISO standard is open, as far as I'm concerned, and the
new PAS guidelines allow Sun to have control over the Java
standard yet still make Java an ISO standard. OK Sun isn't
a PAS submitter *yet* but they're getting there.
The PAS guidelines basically state that the specifications
must be publically available, any interested bodies have a
right to participate, and the public must not be prevented
from creating and distributing their own implementations.
Sun has met all three of these requirements and it is only
a matter of time before Java becomes an open ISO standard.
> "Open" isn't the same as "published". To be open, a standard must be
> change-controlled by a body of unrestricted membership, which has no stake
> in any specific implementation.
I don't know what gave you the impression that this is the
definition of an open standard, especially the last part.
Steve Mading
r.e.b...@usa.net wrote:
: In article <67n7ap$t...@news.microsoft.com>,
: dbr...@microsoft.com (David Brooks) wrote:
: > Nathan Hand <h930...@student.anu.edu.au> writes:
: > >> B) Its not an open standard. Sun controls it.
: > >It *is* an open standard, despite Sun controlling it. If you have
: > >no idea what it means for a standard to be open, as this point of
: > >yours would seem to suggest, then I recommend you expend a little
: > >effort and *learn* what an open standard means.
: Sun's leadership has been the leader in "open" systems. Bill Joy wrote
: vi to support virtually any terminal from a KSR-33 teletype to an Xterm
: with a SUN Keyboard.
He wrote it to work with just about any terminal, true, but I doubt
that he was thinking of an xterm at the time, given that they didn't
exist yet at the first inception of 'vi'. Of course, the fact that
vi could easily be ran later on an xterm is evidence of its terminal
independance.
: > I'm curious: what do *you* think an open standard means?
: Vendors have played with this term since the 80's. The working
: definition, the one that seems to be effective in actual practice, must
: meet the following criteria:
: 1. The language, protocols, and structures must be documented and
: published in a manner which allows retrieval without fee. All changes
: should also be available in this manner.
: 2. A reference implementation - providing the core functionality, is
: available in source format that can be compiled on an industry standard
: GPL compiler such as GCC or on a compiler that can be compiled using GPL
: source code and a GCC compiler.
I would say that this is not so much a requirement of openness as
it is just a conveinience. An open standard could exist without
a reference implementation, but it would be a little more frustrating
to work with. I don't think this 'requirement' is nearly as strong
as requirement #1. I say this because there is a certain company
who shall remain nameless that feels that having a reference
implementation is all it takes to have an open standard. Or,
worse yet, that when the reference implementation is different
than the specification docs that it's the docs that are wrong.
This is precisely the wrong attitude to foster if you want
open standards.
: 3. The reference implementation will either not offer GUI features, or
: will implement GUI features as an isolated library which can be
: implemented in X11/R6 (or earlier) libraries, and can be ported to
: Microsoft Foundation classes with little effort.
Why restrict this idea to just GUI's? It's not so much the fact that
the GUI is specified that bothers you, but rather it's the idea
that the spec is tied to one archetecture (I assume). There are
other things than GUI's that can have this effect. For example,
you could make the spec assume that the programmer is using VC++
with the MFC's, and even if it wasn't GUI, it would still have
the kind of problem you are talking about here.
: 4. Sample data, programs, clients, servers, or test data generators are
: provided in a manner than can be generated as a reference
: implementation.
Again, I feel like reference implementations, while a useful tool,
are not really a part of what it takes for a standard to be "open".
Remember that not all standards are programs. For example, how
do I write a "reference implementation" for the "ascii" standard?
Some standards refer mearly to data, and not to processes.
: This seems to be the "working definition". Variations that have been
: tried include published protocol but no reference models (Motif), but
: they are vulnerable to a standard that meets the criteria above (OLIT,
: Athena, HP).
Eh? Athena does not pretend to implement the "Motif" standard. They
are both subsets of the 'standard' called 'an X toolkit'. I don't
quite understand what you are trying to say here.
--
Steve Mading: mad...@execpc.com http://www.execpc.com/~madings
Bob O
On Tue, 23 Dec 1997 02:23:21, dbr...@microsoft.com (David Brooks) wrote:
> >It *is* an open standard, despite Sun controlling it. If you have
> >no idea what it means for a standard to be open, as this point of
> >yours would seem to suggest, then I recommend you expend a little
> >effort and *learn* what an open standard means.
>
> I'm curious: what do *you* think an open standard means?
>
> "Open" isn't the same as "published". To be open, a standard must be
> change-controlled by a body of unrestricted membership, which has no stake
> in any specific implementation. Sun can change the spec of Java whenever
> it likes unilaterally, as is its privilege under the current rules (I
> heard, unconfirmed, that changes to floating-point handling were in the
> small-print of the Merced deal).
Not true. New open standard bodies are being formed all the time. If you
want the definition of "not open" you probably should start with Microsoft
and then move towards open from there. This is not a black or white
concept but a concept in degrees. Additionally, Sun is only the submitter
of the technology. When Sun submits the technology to the ISO they will be
giving up the controls you are concerned about. Sun has some latitude
that you abhor during the development stage, but this is necessary, IMO, to
spur development quickly. If Sun abuses this position then the results
will not be accepted by the ISO.
Bob O - Computing for fun
Christopher Estep
r.e.b...@usa.net wrote in message <882940398...@dejanews.com>...
>In article <67n7ap$t...@news.microsoft.com>,
For one thing, Microsoft is very
>unlikely to start contibuting free "reference copies" of Microsoft
>Word in order to integrate activeX controls into guavac and kaffe. For
>another Micrsoft is very restricted in it's ability to run "UNIX" code.
Any ActiveX controls Microsoft (or any other member of the ActiveX Working
Group) creates MUST be turned over to the group as a whole for inspection
and approval (Microsoft does NOT control the AXWG, as it is part of The Open
Group, and was designated as such SPECIFICALLY to prevent Microsoft, or
anyone else, to control how ActiveX controls are implemented in browsers ort
anything else that can use ActiveX controls).
ActiveX controls themselves aren't even limited to Windows (or even to C++);
it is possible to write ActiveX controls in (surprisingly) 100% Pure Java
(both Sun and Symantec have demonstrated this); and if Java is truly "write
once/run anywhere" then ActiveX on UNIX (or any other non-Windows OS) is
quite workable (and doable).
>
>The most dramatic illustration of this was the ISO OSI protocol. The ISO
>decided to charge for distribution, charging $30 per book for over 100
>books and specifications. The price to train one programmer to
>understand OSI could go as high as $50,000. The supporters of OSI
>(primarily DEC, IBM, and AT&T) felt that would protect them from
>competitors who had been supporting TCP/IP.
>
>The strategy backfired. Not only was the industry reluctant to support
>OSI (even with the government mandating GOSIP), the developers of TCP/IP
>began distributing free copies of TCP/IP for various implementations of
>UNIX, DOS, and eventually Windowms. The Trumpet TCP/IP stack, coupled
>with the Mosaic Browser, turned TCP/IP into the undisputable winner of
>the protocol wars. The IETF will need to get similar support to gain the
>acceptance of IPv6.
That is indeed what the IETF is doing with IPv6. Led by Cisco Systems,
NorTel, and 3Com, they are making IPv6 a little-cost (or no cost) add-in (in
most cases, by including IPv6 support in existing products, such as
routers).
However, Microsoft CAN (and has) supported TCP/IP (in Windows NT from the
beginning, and as far back as Windows for Workgroups 3.11 on LANs; even the
much-criticized LAN Manager offered TCP/IP as an option). Windows 95 became
Microsoft's first desktop OS for everyone to include dialup TCP/IP "out of
the box" (along with, for the most part, complete application compatibility
with existing TCP/IP applications for Windows NT and Trumpet Winsock). The
new Winsock 2 (included as part of Windows 98 and Windows NT 5.0) further
extends this to the third generation of Winsock applications (NOTE: Both the
original Winsock 1.1 and Winsock 2 were NOT written either at Microsoft or
even with Microsoft development tools, even though both are part and parcel
of Microsoft operating systems.)
>If you really want to serve your masters, you should take the time to get
>intimately familiar with Linux, especially Red Hat 5.0 and Caldera 1.1 to
>the point where you know every strength and weakness. Many technophiles
>will be getting Linux this Christmas and the the usual rhetoric referring
>to the "clunky character mode interface" and the "incredibly difficult
>installation and configuration" isn't going to help your credibility.
>Windows will be "sharing the drive" with a new neighbor - Linux.
Most versions of Linux HAVE been notoriously hard for non-technophiles (and
even some tech-heads) to get up and running (especially in the installation
phase). The installation phase of ANY Linux is almost wholly
character-based (if this has changed for any of the "major" flavors of
Linux, please feel free to point me in their direction!). Compared to even
Red Hat Linux, Windows NT Workstation is a breeze to install, configure, and
get up and running, even on an otherwise all-UNIX network.
ddav...@netcom.ca.nosppamplease
>
> Most versions of Linux HAVE been notoriously hard for non-technophiles (and
> even some tech-heads) to get up and running (especially in the installation
> phase). The installation phase of ANY Linux is almost wholly
> character-based (if this has changed for any of the "major" flavors of
> Linux, please feel free to point me in their direction!). Compared to even
> Red Hat Linux, Windows NT Workstation is a breeze to install, configure, and
> get up and running, even on an otherwise all-UNIX network.
>
Unless you yourself have installed Linux, please do not generalize the
installaton of redhat linux as difficult simply because the installation is
character based. I simply can not see the difference between installing linux
full screen sequentially prompting the user for information as being any
different than having an install wizard doing the same sequences prompting the
user in a graphical environment. Just because an installation is graphical or
character based does not make it any more or less easy.
--
David A Davidson <ddav...@netcom.ca.nospamplease>
Micro$oft Corporation, 1 The Only Way (ours), Redmond Washington
The Information Superhighway (AKA "The Road Ahead"), use to be called
the freeway, now the freeway is a path of road blocks, dead ends and
tollways. Theres only a one way street, thanks to the folks in Redmond.