Books on Working For Microsoft
Keith Watson
worked as a programmer ar Microsoft. The extracts I heard were very funny.
I remember one piece was about this guy and his colleagues going to lunch
and, because they though Bill Gates watched from his office window, one of
them struggled through the bushes because it was a more direct route than
the winding path. He thought it would demostrate his creativity and
initiative!
Trouble is I can't remember either the author or the title. Can anyone
help??
Keith Watson
Reading (no pun intended), UK
Dene Bebbington
>I heard a radio review sometime back about a book written by a guy who
>worked as a programmer ar Microsoft. The extracts I heard were very funny.
>I remember one piece was about this guy and his colleagues going to lunch
>and, because they though Bill Gates watched from his office window, one of
>them struggled through the bushes because it was a more direct route than
>the winding path. He thought it would demostrate his creativity and
>initiative!
This explains everything, the usual Microsoft programmer always take the
winding path, now we know why the words creative and innovative don't
normally exemplify Microsoft.
--
Dene Bebbington http://www.bebbo.demon.co.uk
"Beside the braes of dawn. One clear new morning. Down where the lilies
stood in bloom. I knew that I was just a stranger in this world. A wind
just passing through." - Calum & Rory Macdonald (Runrig)
Jim Collier
>
> Keith Watson <kpwa...@bcs.org.uk> wrote:
> >I heard a radio review sometime back about a book written by a guy who
> >worked as a programmer ar Microsoft. The extracts I heard were very funny.
> >I remember one piece was about this guy and his colleagues going to lunch
> >and, because they though Bill Gates watched from his office window, one of
> >them struggled through the bushes because it was a more direct route than
> >the winding path. He thought it would demostrate his creativity and
> >initiative!
>
> This explains everything, the usual Microsoft programmer always take the
> winding path, now we know why the words creative and innovative don't
> normally exemplify Microsoft.
Sounds a little legendary. Microsoft's present suite of programming
languages and Office 9x running under Windows NT make a nicely
integrated package which nothing on Unix approaches and are the
self-evident products of highly creative minds. I can think of reasons
to criticize Bill Gates, but the criticism I read levelled at
Microsoft's technical prowess comes off like the finger-wagging of the
Little Old Lady in Dubuque against New York and "Hollywood" and places
she's never been. Or like NANAE morons railing against Amazon Books.
The problem of fawning employees around bosses is ancient. I think
most bosses categorize the fawn into the martinet who acts
overtly servile around his supervisor but is a first-class shit to
everyone else, and the jerk who exhibits a bonhomie towards his boss
typified by body language like sitting on the boss's desk. Both are
energy-sapping problems.
Jim
Ron Hardin
> > This explains everything, the usual Microsoft programmer always take the
> > winding path, now we know why the words creative and innovative don't
> > normally exemplify Microsoft.
>
> Sounds a little legendary. Microsoft's present suite of programming
> languages and Office 9x running under Windows NT make a nicely
> integrated package which nothing on Unix approaches and are the
> self-evident products of highly creative minds.
The GM of a Western Electric plant we once worked in noticed that I and
a friend were always going to breakfast across the street during working
hours. Possibly crossing the lawn attracted his initial notice, but then
he saw a daily pattern.
Our bosses were duly informed, and they told him to fuck off. Our bosses
career paths are about equal to ours.
--
Ron Hardin
rhha...@mindspring.com
On the internet, nobody knows you're a jerk.
Dene Bebbington
>Dene Bebbington wrote:
>>
>> Keith Watson <kpwa...@bcs.org.uk> wrote:
>> >I heard a radio review sometime back about a book written by a guy who
>> >worked as a programmer ar Microsoft. The extracts I heard were very funny.
>> >I remember one piece was about this guy and his colleagues going to lunch
>> >and, because they though Bill Gates watched from his office window, one of
>> >them struggled through the bushes because it was a more direct route than
>> >the winding path. He thought it would demostrate his creativity and
>> >initiative!
>>
>> winding path, now we know why the words creative and innovative don't
>> normally exemplify Microsoft.
>
>Sounds a little legendary. Microsoft's present suite of programming
>languages and Office 9x running under Windows NT make a nicely
>integrated package which nothing on Unix approaches and are the
>self-evident products of highly creative minds.
Well, UNIX platforms are usually used for developing applications other
than typical office software. The main problem I have with Microsoft is
that the Operating Systems such as Win95 and NT are so resource
demanding in comparison to something like UNIX or Linux. Just to run one
application comfortably on Win95 you need quite a lot of computing
power. Oh, and for the record, there's a lot of software developed on
UNIX that are also the product of creative minds.
Considering the way things are going so far, the cynics amongst us might
think that by the time Win2010 comes out you'll need a Cray just so that
the OS can tick over.
[rest snipped]
Jorn Barger
[...]
> Trouble is I can't remember either the author or the title. Can anyone
> help??
I don't think it's the one you mean, but a journalist named Fred Moody
had extraordinary access to the MS team working on their children's
encyclopedia, and wrote a pretty appalling account of it called "I Sing
the Body Electronic". It's a classic, imho, for many reasons,
especially showing the fits and starts of poor project management.
--
<URL:http://www.mcs.net/~jorn/html/weblogs/weblog.html> Patk O'Brian:
"You might not think it, but he is a very sensitive cove, and he feels
harsh words extremely. I remember when his father called him a vile
concupiscent wastethift whoremonger he brooded over it a whole evening."
Jim Collier
>
> Jim Collier <pacc...@mindspring.com> wrote:
> >Sounds a little legendary. Microsoft's present suite of programming
> >languages and Office 9x running under Windows NT make a nicely
> >integrated package which nothing on Unix approaches and are the
> >self-evident products of highly creative minds.
>
> Well, UNIX platforms are usually used for developing applications other
> than typical office software.
As is Microsoft's extension of Office suite features into Visual C++
and Visual Basic. The built-in database engine in the two Visuals
is extensively used in scientific and AI technology for instance.
And why one would even care about X-Windows after playing with the
last three versions of Visual Basic eludes me.
> The main problem I have with Microsoft is
> that the Operating Systems such as Win95 and NT are so resource
> demanding in comparison to something like UNIX or Linux.
Unix requires a virtual memory (and uses it!) and NT requires
a virtual memory. What's the big deal? This is 1998 and memory
is hardly worth the price of discussion. I can buy a 2-gigabyte
hard drive, an obsolete size, for what I make in an hour, or
33 megabytes/minute. Really, modern operating systems aren't large
enough to take advantage of available resources.
> Just to run one
> application comfortably on Win95 you need quite a lot of computing
> power. Oh, and for the record, there's a lot of software developed on
> UNIX that are also the product of creative minds.
Yes, little freebie programs abound. I don't run a $10 million
company on give-away software, I'm afraid, and yet the price of
a *new* Microsoft installation with everything of interest is under
$1000, a pittance.
> Considering the way things are going so far, the cynics amongst us might
> think that by the time Win2010 comes out you'll need a Cray just so that
> the OS can tick over.
By the time 2010 comes around, the last Unix manual will be in
antiquarian bookstores.
(Those wishing to complain that NT doesn't have the grep function
should learn to use new features.)
Jim
hypatia
Edstrom and Marlin Eller.
hypatia
Keith Watson wrote:
> I heard a radio review sometime back about a book written by a guy who
> worked as a programmer ar Microsoft. The extracts I heard were very funny.
> I remember one piece was about this guy and his colleagues going to lunch
> and, because they though Bill Gates watched from his office window, one of
> them struggled through the bushes because it was a more direct route than
> the winding path. He thought it would demostrate his creativity and
> initiative!
>
> Trouble is I can't remember either the author or the title. Can anyone
> help??
>
SubGenius
Jim Collier (pacc...@mindspring.com) prophesied:
: By the time 2010 comes around, the last Unix manual will be in
: antiquarian bookstores.
+---------------------------------SubG------------------------------------+
Hey, yeah? And what about all the FORTRAN manuals?
This post
edited
with TECO,
Yours etc.,
SubGenius
-----BEGIN PGP SIGNATURE-----
Version: 2.6.2
iQCVAwUBNiGOBUOoGXQKy1gNAQE0iwP+LwZNezzUIksdQMy27b4oeZtGNHFotubM
avg0nFQJJiE22ms3BAAQxuI0Q+mUoYnRmAKt3qXXxcndKMjG+GLN9imiF+JNAQJT
o6bpLN5LXzqxVF/6UHNlvhQ9oqtZ3eIPC70ewXTD45hGIUo3Oh49oHaaN6wF8mtK
SqA1czAmSUo=
=JrIG
-----END PGP SIGNATURE-----
tejas
>
> -----BEGIN PGP SIGNED MESSAGE-----
>
> Jim Collier (pacc...@mindspring.com) prophesied:
>
> : By the time 2010 comes around, the last Unix manual will be in
> : antiquarian bookstores.
>
> +---------------------------------SubG------------------------------------+
> Hey, yeah? And what about all the FORTRAN manuals?
We got a few PRIMEs still chuggin' along. What about them?
--
TBSa...@richmond.infi.net (also te...@infi.net)
'Do the boogie woogie in the South American way'
Hank Snow THE RHUMBA BOOGIE
Richard Harter
>SubGenius wrote:
>>
>> -----BEGIN PGP SIGNED MESSAGE-----
>>
>> Jim Collier (pacc...@mindspring.com) prophesied:
>>
>> : By the time 2010 comes around, the last Unix manual will be in
>> : antiquarian bookstores.
>>
>> +---------------------------------SubG------------------------------------+
>> Hey, yeah? And what about all the FORTRAN manuals?
>We got a few PRIMEs still chuggin' along. What about them?
If they aren't running PRIMOS they don't count.
Richard Harter, c...@tiac.net, The Concord Research Institute
URL = http://www.tiac.net/users/cri, phone = 1-978-369-3911
If you can laugh at something it can't hurt you.
It can kill you but it can't hurt you.
Paul Ilechko
<pacc...@mindspring.com> wrote:
>As is Microsoft's extension of Office suite features into Visual C++
>and Visual Basic. The built-in database engine in the two Visuals
>is extensively used in scientific and AI technology for instance.
>And why one would even care about X-Windows after playing with the
>last three versions of Visual Basic eludes me.
I get the impression from reading Jim's posts that he has never been
involved in systems that are meant to be used by thousands of users.
Most applications built for that level of use run on mainframes or
UNIX boxes. NT don't scale, and it ain't reliable enough. It's fine
for a client workstation, but it's a miserable server platform.
Paul.
***************
Paul Ilechko
http://www.transarc.com/~pilechko/homepage.htm
Francis Muir
>
> By the time 2010 comes around, the last Unix manual will be in
> antiquarian bookstores.
By the time 2010 comes around, *all* computer manuals will be so
positioned. But you may be quite wrong about Unix since it appears
that Linux may be about to offer strong competition to NT.
FM
don_...@kvo.com
c...@tiac.net (Richard Harter) wrote:
> tejas <tbsa...@richmond.infi.net> wrote:
>
> >SubGenius wrote:
> >>
> >> -----BEGIN PGP SIGNED MESSAGE-----
> >>
> >> Jim Collier (pacc...@mindspring.com) prophesied:
> >>
> >> : antiquarian bookstores.
> >>
> >> +---------------------------------SubG------------------------------------+
> >> Hey, yeah? And what about all the FORTRAN manuals?
>
> >We got a few PRIMEs still chuggin' along. What about them?
>
> If they aren't running PRIMOS they don't count.
Anyone want to buy an NCR mini running PICK?
Don
-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
Christopher Gross
Keith Watson <kpwa...@bcs.org.uk> wrote:
>I heard a radio review sometime back about a book written by a guy who
>worked as a programmer ar Microsoft. The extracts I heard were very funny.
>I remember one piece was about this guy and his colleagues going to lunch
>and, because they though Bill Gates watched from his office window, one of
>them struggled through the bushes because it was a more direct route than
>the winding path. He thought it would demostrate his creativity and
>initiative!
An incident like that occurs in the novel _Microserfs_ by Douglas
Coupland. I have no idea if he ever actually worked for Microsoft,
though.
--Chris
______________________________________________________________________
Christopher Gross On the Internet, nobody knows I'm a dog.
chr...@gwu.edu
Dene Bebbington
>> The main problem I have with Microsoft is
>> that the Operating Systems such as Win95 and NT are so resource
>> demanding in comparison to something like UNIX or Linux.
>
>Unix requires a virtual memory (and uses it!) and NT requires
>a virtual memory. What's the big deal?
I know several people who tell me that Linux runs on a 486 with 16mb of
RAM more comfortably than Win95 runs on a Pentium with 32mb. And with
Linux you get a proper multi-tasking OS. Win95 comes across as an
inefficient bodged OS that has evolved from and still relies on MS-
DROSS.
> This is 1998 and memory
>is hardly worth the price of discussion.
Good thing for Microsoft and their users that it is.
> I can buy a 2-gigabyte
>hard drive, an obsolete size, for what I make in an hour, or
>33 megabytes/minute. Really, modern operating systems aren't large
>enough to take advantage of available resources.
Then why does Win95 running one application on a Pentium with 32mb still
page the disk to quite a degree?
>> Just to run one
>> application comfortably on Win95 you need quite a lot of computing
>> power. Oh, and for the record, there's a lot of software developed on
>> UNIX that are also the product of creative minds.
>
>Yes, little freebie programs abound.
Who said anything about freebie software.
> I don't run a $10 million
>company on give-away software, I'm afraid, and yet the price of
>a *new* Microsoft installation with everything of interest is under
>$1000, a pittance.
>
>> Considering the way things are going so far, the cynics amongst us might
>> think that by the time Win2010 comes out you'll need a Cray just so that
>> the OS can tick over.
>
>By the time 2010 comes around, the last Unix manual will be in
>antiquarian bookstores.
I wouldn't bank on it.
On a TV documentary I once saw it was commented upon how some of the
Microsoft programmers were recently graduated from University and even
slept in the office when deadlines were looming - I think that tells you
all you need to know about what goes into their products.
Jim Collier
pile...@NOSPAM.transarc.com (Paul Ilechko) wrote:
> On Sun, 11 Oct 1998 20:44:45 -0700, Jim Collier
> <pacc...@mindspring.com> wrote:
>
> >As is Microsoft's extension of Office suite features into Visual C++
> >and Visual Basic. The built-in database engine in the two Visuals
> >is extensively used in scientific and AI technology for instance.
> >And why one would even care about X-Windows after playing with the
> >last three versions of Visual Basic eludes me.
>
> I get the impression from reading Jim's posts that he has never been
> involved in systems that are meant to be used by thousands of users.
> Most applications built for that level of use run on mainframes or
> UNIX boxes. NT don't scale, and it ain't reliable enough. It's fine
> for a client workstation, but it's a miserable server platform.
You're a bit out of date there. Visual C++ and Visual Basic have nothing to
do with the number of users. NT has all the usual process-scheduling and
file-locking controls and inter-user communications that any multi-user,
multi-tasking system must have. And anyway, the typical local-area network
isn't tied to an operating system. Large companies like Boeing or
Hewlett-Packard have a common LAN architecture for all workstations. It so
happens that H-P uses primarily NT, but many engineers also have Unix
workstations at their desks.
A number of Boeing sites use only NT to serve thousands of users, and string
a few Unix workstations together in large labs on a sort of project ad-hoc
basis. For instance, the Boeing North American site in Anaheim (formerly
Rockwell Autonetics) uses Unix on one project because the developer of a
disassembler-debugger used on the project had scant knowledge of Microsoft
software and decided to do it in Unix. Since he didn't know any better, his
multi-year development was split between a "graphics" guy who wrote the
X-Windows end of it and his disassembler. He could easily have done the
whole thing in Visual C++ or even in Visual Basic himself in a few months.
His Unix audience consists of a dozen or so engineers. The other 1500 users
at Anaheim use NT, but this isn't really relevant to the LAN. Nor is it
relevant that the klutzy implementation that this guy has implemented is slow
and frequently crashes. That's his fault, not the system's. He's done a
fairly credible job for a guy who didn't have the aid of a so-called
integrated-development environment of the kind which was pioneered by Borland
and, as usual, advanced a few giant steps by everybody's favorite whipping
boy. His only real problem is that his knowhow has been obsolete for the
past four years.
Jim
Jim Collier
su...@atheist.tamu.edu (SubGenius) wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
>
> Jim Collier (pacc...@mindspring.com) prophesied:
>
> : antiquarian bookstores.
>
> +---------------------------------SubG------------------------------------+
> Hey, yeah? And what about all the FORTRAN manuals?
Well, with the phasing out of charcoal-fire starter fuel in California...
Jim Collier
Dene Bebbington <de...@bebbo.demon.co.uk> wrote:
> Jim Collier <pacc...@mindspring.com> wrote:
> >Dene Bebbington wrote:
> [snips]
> >> The main problem I have with Microsoft is
> >> that the Operating Systems such as Win95 and NT are so resource
> >> demanding in comparison to something like UNIX or Linux.
> >
> >Unix requires a virtual memory (and uses it!) and NT requires
> >a virtual memory. What's the big deal?
>
> I know several people who tell me that Linux runs on a 486 with 16mb of
> RAM more comfortably than Win95 runs on a Pentium with 32mb.
Neglecting that we're talking about Windows NT, a multi-user system, not
Windows 95 a single-user, multi-tasking system which has built-in software for
communicating with a network card (and also neglecting that 32 mb was until
late 1997 a large RAM size for Windows 95 systems), what does "runs...more
comfortably" mean?
> And with
> Linux you get a proper multi-tasking OS. Win95 comes across as an
> inefficient bodged OS that has evolved from and still relies on MS-
> DROSS.
Juvenile sloganeering.
> Then why does Win95 running one application on a Pentium with 32mb still
> page the disk to quite a degree?
Probably because the software writer decided to make disk calls "to quite a
degree." Your prose is rather like asking why a person would look
obsessively behind himself when zooming down a highway at 100 mph. Beats me,
Louie.
>
> >> Just to run one
> >> application comfortably on Win95 you need quite a lot of computing
> >> power. Oh, and for the record, there's a lot of software developed on
> >> UNIX that are also the product of creative minds.
> >
> >Yes, little freebie programs abound.
>
> Who said anything about freebie software.
Say, what exactly *do* you know? Linux is distributed for free.
While we're at it, could you quantify "need quite a lot of computing
power"?
> On a TV documentary I once saw it was commented upon how some of the
> Microsoft programmers were recently graduated from University and even
> slept in the office when deadlines were looming - I think that tells you
> all you need to know about what goes into their products.
Again, whatever Bill Gates' sins might be, this is the sort of mealy-mouthed
criticism that can only come from a juvenile non-player. Just about all
people who produce anything of any value sweat blood. Engineers working
18-hour stints near deadlines are hardly limited to Microsoft, and they are
certainly not limited to America.
m...@ptyx.com
server operating system from FreeBSD to NT. So much for MS scalability.
Whereas at the other end of the scale, the success of the Palm Pilot OS
highlights the acute embarrassment that is Windows CE.
Cordially -- Mikhail Zel...@math.ucla.edu * M...@ptyx.com ** www.ptyx.com
God: "Sum id quod sum." ** 7576 Willow Glen Road, Los Angeles, CA 90046
Descartes: "Cogito ergo sum." * 323.876.8234 (fon) * 323.876.8054 (fax)
Popeye: "Sum id quod sum et id totum est quod sum." **** www.alonzo.org
since 2.26.1958 ** itinerant assault philosopher ** will think for food
In article <70h71t$esb$1...@nnrp1.dejanews.com>,
Ted Samsel
: In article <6vs2mm$n2r$1...@atheist.tamu.edu>,
: su...@atheist.tamu.edu (SubGenius) wrote:
: > -----BEGIN PGP SIGNED MESSAGE-----
: >
: > Jim Collier (pacc...@mindspring.com) prophesied:
: >
: > : By the time 2010 comes around, the last Unix manual will be in
: > : antiquarian bookstores.
: >
: > +---------------------------------SubG------------------------------------+
: > Hey, yeah? And what about all the FORTRAN manuals?
: Well, with the phasing out of charcoal-fire starter fuel in California...
So you use naphtha? Kerosene?
--
Ted Samsel....tejas@infi.net (or tbsa...@richmond.infi.net)
"do the boogie woogie in the South American way"
Rhumba Boogie- Hank Snow (1955)
Paul Ilechko
<pacc...@mindspring.com> wrote:
>You're a bit out of date there. Visual C++ and Visual Basic have nothing to
>do with the number of users. NT has all the usual process-scheduling and
>file-locking controls and inter-user communications that any multi-user,
>multi-tasking system must have.
Yes, they just don't work as well as other systems. And you totally
missed the point I wsa making, perhaps because we work with different
industries. I'm not particularly interested in workstations - most of
my clients are looking to replace them with browsers anyway - but with
application servers. And that is where NT really falls short, as does
MS's workstation derived, tools based software strategy.
obNotEvenAGoodJoke : Microsoft Transaction Server
Richard Harter
>On Tue, 20 Oct 1998 05:28:29 GMT, Jim Collier
><pacc...@mindspring.com> wrote:
>
>
>>You're a bit out of date there. Visual C++ and Visual Basic have nothing to
>>do with the number of users. NT has all the usual process-scheduling and
>>file-locking controls and inter-user communications that any multi-user,
>>multi-tasking system must have.
>
>Yes, they just don't work as well as other systems. And you totally
>missed the point I wsa making, perhaps because we work with different
>industries. I'm not particularly interested in workstations - most of
>my clients are looking to replace them with browsers anyway - but with
>application servers. And that is where NT really falls short, as does
>MS's workstation derived, tools based software strategy.
I am told that at one of West coast ISP's the manager told the staff
that they were switching to NT for all their servers. The entire staff
quit en masse.
Michael Kagalenko
]
] NT has all the usual process-scheduling and
]file-locking controls and inter-user communications that any multi-user,
]multi-tasking system must have.
And, those are implemented poorly. Linux, OS written by hobbists, is
by all accounts more stable and runs much faster on the same hardware.
NT has video drivers in the kernel, you don't get more braindamaged design
then that.
UNIX killer, NT isn't. And, NT 5 will be even more bloated. Gates chose
tp [ut everything and kitchen sink into the new release, instead of
improving performance and stability.
Dene Bebbington
>In article <CHg72iA3...@bebbo.demon.co.uk>,
> Dene Bebbington <de...@bebbo.demon.co.uk> wrote:
>> Jim Collier <pacc...@mindspring.com> wrote:
>> >Dene Bebbington wrote:
>> [snips]
>> >> The main problem I have with Microsoft is
>> >> that the Operating Systems such as Win95 and NT are so resource
>> >> demanding in comparison to something like UNIX or Linux.
>> >
>> >Unix requires a virtual memory (and uses it!) and NT requires
>> >a virtual memory. What's the big deal?
>>
>> I know several people who tell me that Linux runs on a 486 with 16mb of
>> RAM more comfortably than Win95 runs on a Pentium with 32mb.
>
>Neglecting that we're talking about Windows NT, a multi-user system, not
>Windows 95 a single-user, multi-tasking system which has built-in software for
>communicating with a network card (and also neglecting that 32 mb was until
>late 1997 a large RAM size for Windows 95 systems), what does "runs...more
>comfortably" mean?
Not having to page the disk as much.
>> And with
>> Linux you get a proper multi-tasking OS. Win95 comes across as an
>> inefficient bodged OS that has evolved from and still relies on MS-
>> DROSS.
>
>Juvenile sloganeering.
Maybe, nevertheless Win95 still comes across (at least to me and many
people I know) as an inefficient bodged OS, and it did evolve from and
still relies on MS-DROSS.
>> Then why does Win95 running one application on a Pentium with 32mb still
>> page the disk to quite a degree?
>
>Probably because the software writer decided to make disk calls "to quite a
>degree." Your prose is rather like asking why a person would look
>obsessively behind himself when zooming down a highway at 100 mph. Beats me,
>Louie.
Or maybe some of the programming in Win95 is sloppy.
>> >> Just to run one
>> >> application comfortably on Win95 you need quite a lot of computing
>> >> power. Oh, and for the record, there's a lot of software developed on
>> >> UNIX that are also the product of creative minds.
>> >
>> >Yes, little freebie programs abound.
>>
>> Who said anything about freebie software.
>
>Say, what exactly *do* you know? Linux is distributed for free.
I specifically mentioned UNIX in that paragraph, not Linux. Seems I know
how to read.
>> On a TV documentary I once saw it was commented upon how some of the
>> Microsoft programmers were recently graduated from University and even
>> slept in the office when deadlines were looming - I think that tells you
>> all you need to know about what goes into their products.
>
>Again, whatever Bill Gates' sins might be, this is the sort of mealy-mouthed
>criticism that can only come from a juvenile non-player.
I dunno, can't remember who presented the documentary. Nevertheless, as
a software engineer in the industry for ten years and having worked on
many different projects I can make those comments without being mealy-
mouthed. I've seen enough crap code written by supposedly experienced
developers, the thought of fairly new graduates working all hours of the
day and night gives me little confidence in what they churn out.
> Just about all
>people who produce anything of any value sweat blood.
Bollocks.
> Engineers working
>18-hour stints near deadlines are hardly limited to Microsoft, and they are
>certainly not limited to America.
True, but like the rest they are apt to not be working at their best in
such conditions.
Paul Ilechko
<de...@bebbo.demon.co.uk> wrote:
>Or maybe some of the programming in Win95 is sloppy.
- which describes in detail some of the pathetic compromises made in a
last ditch effort to avoid being Windows 96. Like making all the 16
bit code (and there is plenty of it, particularly in the GDI subsystem
- can you say "Thunk" ?) dependent on a single semaphore. Which is
part of the reason why Win 95 had better support for DOS applications
than for 16 bit Windows applications.
Jim Collier
>
> When Microsoft purchased Hotmail, they briefly tried switching its web
> server operating system from FreeBSD to NT. So much for MS scalability.
> Whereas at the other end of the scale, the success of the Palm Pilot OS
> highlights the acute embarrassment that is Windows CE.
Not sure what the point is about Windows CE. As you probably know,
Windows CE is intended for tiny memory sizes, such as are found in
*hand*-held computers, ones which in general have an order of magnitude
smaller memory than say, notebook computers. The typical application
is an embedded-processor system which is not used for general-purpose
computing, for instance a test instrument.
Jim
Jim Collier
>
> Jim Collier (pacc...@mindspring.com) wrote in article <70h71t$esb$1...@nnrp1.dejanews.com>
> ]
> ] NT has all the usual process-scheduling and
> ]file-locking controls and inter-user communications that any multi-user,
> ]multi-tasking system must have.
>
> And, those are implemented poorly. Linux, OS written by hobbists,
These "hobbists" give telephone support, do they?
> by all accounts more stable and runs much faster on the same hardware.
> NT has video drivers in the kernel, you don't get more braindamaged design
> then that.
Also not true. You need to learn details about modern operating
systems, me friend. (I choose to answer your post only because
you didn't overtly confuse Windows 95 and Windows NT unlike a number of
other posts. You might very well confuse it, but you at least kept
your speculation to yourself if this is the case.)
Unix does not have preemptive scheduling, and given the same
processor, could hardly manage two or more concurrent tasks as fast
as NT where one requires real-time service... for instance, the
periodic interrupts from a signal processing application.
There is a Unix work-around with preemption glued on. It's called
QNX. Rather than lecture, I shall dangle it as a morsel before
the truly curious.
ObBook: William Stallings, _Operating Systems: Internals and Design
Principles_ (3rd ed.)
Jim Collier
Larisa Migachyov
Distribution:
Jim Collier wrote:
I don't know what's wrong with Windows CE either (except for its acronym).
In an electronic organizer, it works just fine. Just exactly how much
processing power does one need to keep track of a schedule and an address
book, anyway? And I don't like the PalmPilot OS - the device is nice, but
Graffiti is a very slow way of entering information.
--
Larisa Migachyov * Quant'e bella giovinezza
Biomechanical Engineering * Che si fugge tuttavia!
Stanford University * Chi vuol esser lieto, sia;
http://www.stanford.edu/~lvm * Di doman non c'e certezza.
Michael Kagalenko
]>
]> Jim Collier (pacc...@mindspring.com) wrote in article <70h71t$esb$1...@nnrp1.dejanews.com>
]> ]
]> ] NT has all the usual process-scheduling and
]> ]file-locking controls and inter-user communications that any multi-user,
]> ]multi-tasking system must have.
]>
]> And, those are implemented poorly. Linux, OS written by hobbists,
]
]These "hobbists" give telephone support, do they?
For a money, "they" do.
]> by all accounts more stable and runs much faster on the same hardware.
]> NT has video drivers in the kernel, you don't get more braindamaged design
]> then that.0
]
]Also not true.
It is not clear what you think is not true. That Linux and FreeBSD run much
faster on the same hardware ? That is well-established fact, Mr.Collier.
In fact, one can put together $500 FreeBSD mail server that will
outperform $4000 MS sexchange NT server, and will not crush and lose mail.
Or perhaps, you think that NT does not have video drivers in kernel ?
It didn't in early versions, but now does.
BTW, a lie about NT sometimes encountered in marketing literature
is that it is C2- certified. This lie may be the reason for US military
adopting NT for mission-critical stuff.
] You need to learn details about modern operating
]systems, me friend. (I choose to answer your post only because
]you didn't overtly confuse Windows 95 and Windows NT unlike a number of
]other posts. You might very well confuse it, but you at least kept
]your speculation to yourself if this is the case.)
Thank you so much for your kindness, Mr.Collier. I am overwhelmed.
]Unix does not have preemptive scheduling, and given the same
]processor, could hardly manage two or more concurrent tasks as fast
]as NT where one requires real-time service... for instance, the
]periodic interrupts from a signal processing application.
We are not talking real-time OS here, Mr.Collier. So far, we
discussed Internet servers. If you are interested in real-time
operating systems, then I would point out to you VxWorks, real-time
OS with unix-like kernel, which is thin enough to run on hand-held
PDAs or Mars rover, and which has microsecond latency. NT may
be OK for half-assed real-time, but then again, for half-assed
real-time one can use RT-Linux, which is free.
]There is a Unix work-around with preemption glued on. It's called
Michael Kagalenko
]Jim Collier (pacc...@mindspring.com) wrote in article <363101...@mindspring.com>
]
]] You need to learn details about modern operating
]]systems, me friend. (I choose to answer your post only because
]]you didn't overtly confuse Windows 95 and Windows NT unlike a number of
]]other posts. You might very well confuse it, but you at least kept
]]your speculation to yourself if this is the case.)
]
] Thank you so much for your kindness, Mr.Collier. I am overwhelmed.
]
]]Unix does not have preemptive scheduling, and given the same
]]processor, could hardly manage two or more concurrent tasks as fast
]]as NT where one requires real-time service... for instance, the
]]periodic interrupts from a signal processing application.
]
] We are not talking real-time OS here, Mr.Collier. So far, we
] discussed Internet servers. If you are interested in real-time
] operating systems, then I would point out to you VxWorks, real-time
] OS with unix-like kernel, which is thin enough to run on hand-held
] PDAs or Mars rover, and which has microsecond latency. NT may
] be OK for half-assed real-time, but then again, for half-assed
] real-time one can use RT-Linux, which is free.
From comp.realtime FAQ:
Is Windows NT (or windows 95, or even Windows CE now) a Real-Time Operating
System?
This question appears repeatedly in this news group. Here are the key
points:
- Despite a real-time class process, the Win32 API is not suitable to be
used for a Real-time system:
1. Too few priorities for processes and threads
2. No priority inheritance mechanism
3. Some calls are synchronous with process from the Dynamic Class
- Despite a good interface to hardware for CLASSICAL applications, this
interface is not suitable to develop a real-time system:
1. Most of the job in a device driver is done at the DPC level. And most
COTS DD take too much time in the DPC.
2. The DPC problem could have been avoided by increasing the number of DPC
levels, but this is not the case.
3. Pentium Power Management interrupt can preempt your system for an
unpredictable amount of time (depending of the BIOS)
- Real-time clock
There is a lack of programmable timer.
For a more complete view, look at article:
http://www.realtime-info.be/encyc/magazine/97q2/winntasrtos.htm
Some companies are now providing Real-Time Extensions to fill up the hole
let opened by Microsoft. (see the RTOS list)
To do so three main approaches exists: include NT as the lowest level
process in an existing RTOS, put a WIN32 API on top of an existing RTOS,
make NT coexists with a RTOS by modifying the HAL.
------------------------------
Is Windows CE 2.0 a real-time operating system?
This question appears frequently since the release of Windows CE 2.0. It is
not suitable for real-time system development because:
1.The number of priority levels is too low;
2. Interrupts can not be nested;
3. Interrupt latency is too high;
------------------------------
Jim Collier
[Good stuff about Microsoft's much told caving in on video drivers
(for the sake of speed, of course) deleted.]
I wrote:
> ]Unix does not have preemptive scheduling, and given the same
> ]processor, could hardly manage two or more concurrent tasks as fast
> ]as NT where one requires real-time service... for instance, the
> ]periodic interrupts from a signal processing application.
>
> We are not talking real-time OS here, Mr.Collier.
Now don't get bent out of shape just because I take you seriously. I
guess I am talking about real-time operating systems indeed.
Originally, we were talking about books about working at Microsoft
though.
A flexible distributed system shouldn't care what its individual users
are doing as long as each can send and receive messages in some orderly
fashion. (That of course is the whole idea behind distributing
one CPU to each user.) One user might be typing text, another might be
processing messages, another might be doing numerical processing and
another might be checking out a new electronic circuit. The latter two
activities routinely crash, which might or might not be signs that
the users are making progress in their efforts. In a properly
configured NT local loop --- and here I actually mean a star
configuration even though I call it a "loop" --- anybody can crash
without affecting other users. In a Unix loop, say one involving Sun
workstations running Solaris, with each user minding his own business,
any crash may require rebooting the entire system, and a crash on the
administrator's workstation guarantees this.
(I mention Solaris because it has features which permit a
casual experimenter to bring the entire system to a halt with only
a modest effort. Just think what he can do when he starts writing
his own shells.)
> So far, we discussed Internet servers.
Er no, the thread started with an anecdote about a guy who
slinked to lunch in Redmond. Or this may have been in the first
reply to the first post. I've lost track, and don't feel
like hunting it down. It's still implicit in the thread title, q.v.
Jim
Jim Collier
>
> Jim Collier <pacc...@mindspring.com> writes:
>
> > Unix does not have preemptive scheduling, and given the same
> > processor, could hardly manage two or more concurrent tasks as fast
> > as NT where one requires real-time service... for instance, the
> > periodic interrupts from a signal processing application.
>
> knowledge about OSes is quite clear.
>
> Unix does have preemptive scheduling. The kernel is not
> pre-emptable. Big difference.
Yes, all is not lost: Unix *can* be defeated.
>
> > ObBook: William Stallings, _Operating Systems: Internals and Design
> > Principles_ (3rd ed.)
>
> You might try to read it sometime instead of waving it around.
I only write real-time software that works on many platforms, little
man, and the twenty-three employees and six contractors who work for
me, and I, make a fair chunk of change in the process. You might want
to do some research on companies located in Corona del Mar, Calif., and
why my account has the name that it does.
Jim
Michael Kagalenko
]
]> Unix does not have preemptive scheduling, and given the same
]> processor, could hardly manage two or more concurrent tasks as fast
]> as NT where one requires real-time service... for instance, the
]> periodic interrupts from a signal processing application.
]
]I really shouldn't be bothering to answer this---Jim's extent of
]knowledge about OSes is quite clear.
]
]Unix does have preemptive scheduling. The kernel is not
]pre-emptable. Big difference.
And anyways, Collier dragged in the issue of real-time into the discussion
of the performance of UNIX vs NT. That shows he doesn't know what is
"real-time OS." It doesn't have to be faster then non-rt, and for some tasks
it may be slower. What matters is not that rt is fast, but that it's
guaranteed to be on time.
Michael Kagalenko
]
][Good stuff about Microsoft's much told caving in on video drivers
](for the sake of speed, of course) deleted.]
And yet, NT is slower and slower with each release. NT 5 will be slower yet.
]I wrote:
]
]> ]Unix does not have preemptive scheduling, and given the same
]> ]processor, could hardly manage two or more concurrent tasks as fast
]> ]as NT where one requires real-time service... for instance, the
]> ]periodic interrupts from a signal processing application.
]>
]> We are not talking real-time OS here, Mr.Collier.
]
]Now don't get bent out of shape just because I take you seriously. I
]guess I am talking about real-time operating systems indeed.
]Originally, we were talking about books about working at Microsoft
]though.
]
]A flexible distributed system shouldn't care what its individual users
]are doing as long as each can send and receive messages in some orderly
]fashion.
Oh goodie, now you are brining in "distributed systems," whatever you mean by
that.
] (That of course is the whole idea behind distributing
]one CPU to each user.)
So, by "distributed systems," you mean systems with one CPU per user. I
guess that makes both a department's LAN and CRAY with lotsa engineers
logged in "distributed systems."
] One user might be typing text, another might be
]processing messages, another might be doing numerical processing and
]another might be checking out a new electronic circuit. The latter two
]activities routinely crash, which might or might not be signs that
]the users are making progress in their efforts. In a properly
]configured NT local loop --- and here I actually mean a star
]configuration even though I call it a "loop" --- anybody can crash
]without affecting other users.
Yeah ? You mean like, for instance, the case when application server dies ?
] In a Unix loop, say one involving Sun
]workstations running Solaris, with each user minding his own business,
]any crash may require rebooting the entire system, and a crash on the
]administrator's workstation guarantees this.
Where do you get this kind of strange ideas about Solaris, I wonder.
](I mention Solaris because it has features which permit a
]casual experimenter to bring the entire system to a halt with only
]a modest effort. Just think what he can do when he starts writing
]his own shells.)
Mambo-jumbo argle bargle woosh-woosh.
]> So far, we discussed Internet servers.
Michael Kagalenko
]>
]> > Unix does not have preemptive scheduling, and given the same
]> > processor, could hardly manage two or more concurrent tasks as fast
]> > as NT where one requires real-time service... for instance, the
]> > periodic interrupts from a signal processing application.
]>
]> knowledge about OSes is quite clear.
]>
]> Unix does have preemptive scheduling. The kernel is not
]> pre-emptable. Big difference.
]Yes, all is not lost: Unix *can* be defeated.
]
]>
]> > ObBook: William Stallings, _Operating Systems: Internals and Design
]> > Principles_ (3rd ed.)
]>
]> You might try to read it sometime instead of waving it around.
]
]I only write real-time software that works on many platforms, little
]man, and the twenty-three employees and six contractors who work for
]me,
They have my pity, each and every one of them. Working for idiots sucks.
Jim Collier
>
I wrote:
> ]I only write real-time software that works on many platforms, little
> ]man, and the twenty-three employees and six contractors who work for
> ]me,
>
> They have my pity, each and every one of them. Working for idiots sucks.
Kagalenko, you do have this habit of shaking your fist at the world.
Now see here, old boy. I actually responded to a couple of your posts
this afternoon, which I hardly ever do, and then not even entirely
in a spirit of maleficence. Be good enough to take what I wrote to you,
or leave it. It makes no difference to me whether you fume at your
monitor. I'm a pretty busy guy, and my thoughts of you disappear
as soon as I read your posts. I honestly have no opinion of
whether you are the asshole in person that you are in this newsgroup,
nor do I wish to start an on-going feud with you, unless of course
you wish to enlist among the NANAE morons... and then I might
pursue you to your grave.
Enough said. I probably should let out with the you're-on-my-killfile
threat, buster. Never mind. Just wait till your father gets home.
Jim
Jim Collier
>
> And anyways, Collier dragged in the issue of real-time into the discussion
> of the performance of UNIX vs NT.
As a matter of fact, I do pretty much talk about what I want. When
the topic is something I don't want to talk about, I don't more often
than not.
This would be a sound concept for you to adopt.
> That shows he doesn't know what is
> "real-time OS." It doesn't have to be faster then non-rt, and for some tasks
> it may be slower. What matters is not that rt is fast, but that it's
> guaranteed to be on time.
Sigh. This is pretty big talk even for Mighty Mouse.
The usual measure of real-time performance, Kagalenko, is thruput (in
units of samples/sec) for some specified error. There is no
other meaningful measure, or more precisely, a criterion can always be
framed in terms of thruput and an error.
Since most signal-processing systems depend on a process occurring at
a periodic rate --- a common processing example consists of filtering
of a digitized signal --- its thruput can't be any faster than its
interrupt rate, unless of course external hardware provides for this
function with suitable buffering for samples which might be streamed
to the processor in bursts. (As you perhaps know, buffer size for this
scenario was a pretty hot theoretic topic in the early 1970s years
before such systems were realized in hardware.) Therefore, with a
properly designed buffer, with each stored sample acquired at
equally spaced intervals, "being on time" at the processor end is
irrelevant from sample to sample, and is only important over a long
interval, say thousands of samples. The more bufferred samples
between master processor interrupts, the less important the effect
of timing uncertainty. That is, increasing a buffer's length decreases
the effect of timing jitter at the CPU, at the expense of introduing
a delay into the system.
I am surprised the poster doesn't know this. Is he by any chance an
electronic tech at Northeastern U., with Administrator privileges on
a dozen Pentium-based workstations with Linux in his lab?
Am I pretty close?
Jim
Jim Collier
>
> Jim Collier <pacc...@mindspring.com> writes:
>
> > The usual measure of real-time performance, Kagalenko, is thruput (in
> > units of samples/sec) for some specified error. There is no
> > other meaningful measure, or more precisely, a criterion can always be
> > framed in terms of thruput and an error.
>
> is interrupt latency. Not thruput.
Er, no, that's completely irrelevant and shows a lack of understanding
of the issues.
The bufferred preprocessor which the poster omitted and which is
standard in signal processing schemes, typically has no latency in the
usual meaning of the term, i.e., doesn't have a variable time to
execute an interrupt service routine. There is no ISR. An analog-to-
digital converter and buffer crap out catastrophically at some sample
rate. Typically the A/D fails first.
It is only for an unbuffered system where ISR latency can have
a direct effect. There is one interrupt per sample, and if the
ISR doesn't begin at the same time relative to the time of
arrival for each sample, this is equivalent to adding angle modulation
noise to the signal. But unbuffered systems have become obsolete.
They serve only as simplified textbook examples. With programmable
array logic (PALs), it is easy to design a buffer for an A/D output
stream. One generic 22V10 (also now nearly obsolete) and a 1K x N RAM
where N is the sample-word size suffice in most cases.
However, noise on the A/D's (or more properly, the sample-and-hold's)
sampling frequency is transferred to the signal, which is akin to the
jitter in interrupt latency for unbuffered systems. Further, on the
buffered system, latency jitter introduces a secondary effect when the
buffer is nearly full, evidenced by short-interval thruput failures
where the dropout times grow in length as the buffer fills or as the
time-to-next ISR increases. A low thruput failure rate may be
acceptable for analog data but is unacceptable for other numerical data
unless error correction is present. The error correction has to be
valid over the longest expected dropout, but this interval is a
statistical phenomenon. The analysis is similar to the "click
analysis" for FM signals used by Rice in the 1940s.
(To return to the previous issue of preemption, which I previously
raised, the variability in interrupt latency time can be made
shorter than for an unpreempted system where the designer has no
control. Also, it is the *variability* in latency time, which
causes a problem. A fixed delay from interrupt to ISR execution
is usually harmless.)
This is all pretty standard fare in applied signal processing. The
poster and Kagalenko are one pair of bullshitters.
Jim
Davis
On Fri, 23 Oct 1998, Jim Collier wrote:
[SNIP]
>without affecting other users. In a Unix loop, say one involving Sun
>workstations running Solaris, with each user minding his own business,
>any crash may require rebooting the entire system, and a crash on the
>administrator's workstation guarantees this.
>
>(I mention Solaris because it has features which permit a
>casual experimenter to bring the entire system to a halt with only
>a modest effort. Just think what he can do when he starts writing
>his own shells.)
>
I nearly fell off my chair laughing, hearing someone hint that NT might be
more stable than some flavor of UNIX. Even harder than when I heard the
suggestion that NT might be faster. Any more good jokes for us?
-Davis
Jim Collier
> I nearly fell off my chair laughing, hearing someone hint that NT might be
> more stable than some flavor of UNIX. Even harder than when I heard the
> suggestion that NT might be faster. Any more good jokes for us?
>
> -Davis
A silly statement given without any justification. It does not
surprise me that the poster hears voices when he reads posts.
Given equal processors processing input data --- though it's not clear
to me that the poster even understands the need for equal processors
when he writes of "some flavor of Unix" --- the system which allows the
user to write his own ISR on a high-priority interrupt line cannot be
slower than one which doesn't, which should be pretty obvious.
I had also pointed out the relatively speedy development time
on Microsoft-based projects. The usual attack-Microsoft idiots
have about the same mentality as the attack-Amazon morons, and
are the same Neanderthals who attacked Toyota in the 1980s (often
with sledgehammers) after the Japanese company introduced features
...which American car manufacturers beginning with Ford later copied.
Jim
Richard Harter
>Rajappa Iyer writes:
>> The *only* meaningful measure of a realtime OS's performance
>> is interrupt latency. Not thruput.
>
>Er, no, that's completely irrelevant and shows a lack of understanding
>of the issues.
[snip some fairly standard signal processing stuff]
>This is all pretty standard fare in applied signal processing. The
>poster and Kagalenko are one pair of bullshitters.
>
>Jim
Well, no, they are not. The problem here seems to be a conflict in
terminology; there are a couple of different sense in which the term
"real time" is used. One sense of the word which is AFAIK still
"official" is that which Kagalenko and Rajapper are using it - response
time is guaranteed. The other, which is often used by people in signal
processing is that the data stream is processed in real time.
The difference is critical - in RT systems (in the first sense) the
response time is the critical factor. Delays due to buffering are not
(or may not be) acceptable. Examples include many military systems and,
in general, systems which are controlling a physical process in real
time.
In applied signal processing, which seems to be the area you are working
in, fixed delays (within reason) are acceptable as long as the data flow
is maintained.
Michael Zeleny
Aaron Gross <aaron+...@bfr.co.il> wrote:
>Jim Collier <pacc...@mindspring.com> writes:
>>I only write real-time software that works on many platforms, little
>>man, and the twenty-three employees and six contractors who work for
>>me, and I, make a fair chunk of change in the process.
>I've noticed before that you like to brag on this newsgroup about how
>much many you make. It's been noted. We're all very impressed.
One is reminded of the antipodean invert James Donald Collier, wont to
promulgate a philosophical justification of his sexuality by means of
broadcasting his GRE scores. There is a kinship here -- or better.
Cordially -- Mikhail Zel...@math.ucla.edu * M...@ptyx.com ** www.ptyx.com
God: "Sum id quod sum." ** 7576 Willow Glen Road, Los Angeles, CA 90046
Descartes: "Cogito ergo sum." * 213.876.8234 (fon) * 213.876.8054 (fax)
Popeye: "Sum id quod sum et id totum est quod sum." **** www.alonzo.org
established on 2.26.1958 ** itinerant philosopher * will think for food
Jim Collier
c...@tiac.net (Richard Harter) wrote:
> Jim Collier <pacc...@mindspring.com> wrote:
>
> >Rajappa Iyer writes:
> >> The *only* meaningful measure of a realtime OS's performance
> >> is interrupt latency. Not thruput.
> >
> >Er, no, that's completely irrelevant and shows a lack of understanding
> >of the issues.
> [snip some fairly standard signal processing stuff]
>
> >This is all pretty standard fare in applied signal processing. The
> >poster and Kagalenko are one pair of bullshitters.
> >
> >Jim
>
> Well, no, they are not. The problem here seems to be a conflict in
> terminology; there are a couple of different sense in which the term
> "real time" is used. One sense of the word which is AFAIK still
> "official" is that which Kagalenko and Rajapper are using it - response
> time is guaranteed.
Therefore, a system with 10-nsec instruction time capable
of servicing interrupts at the rate of 1 per second with a 3-instruction
time interrupt latency is superior to a system with the same instruction
time and capable of servicing interrupts at the rate of 10 x 10^6/sec with a
4-instruction time interrupt latency.
Of what use is the sic-'em approach in interrupt response time if you can't
process anything faster than the movement of a large ice floe in the middle of
winter?
They are official bullshitters.
Dene Bebbington
>
>
>Aaron Gross <aaron+...@bfr.co.il> wrote:
>>Jim Collier <pacc...@mindspring.com> writes:
>
>>>I only write real-time software that works on many platforms, little
>>>man, and the twenty-three employees and six contractors who work for
>>>me, and I, make a fair chunk of change in the process.
>
>>I've noticed before that you like to brag on this newsgroup about how
>>much many you make. It's been noted. We're all very impressed.
>
>One is reminded of the antipodean invert James Donald Collier, wont to
>promulgate a philosophical justification of his sexuality by means of
>broadcasting his GRE scores. There is a kinship here -- or better.
Maybe this is why r.a.b's Jim Collier also supports Micro$oft from
criticism, they make so much money that they must be good.
John J. Rushford Jr.
: when he writes of "some flavor of Unix" --- the system which allows the
: slower than one which doesn't, which should be pretty obvious.
I'm not following you here, Are you saying a UNIX user cannot write his
own device driver or are you saying that I/O device drivers run in user
mode in NT.
John
Jim Collier
> Maybe this is why r.a.b's Jim Collier also supports Micro$oft from
> criticism, they make so much money that they must be good.
What a common idiot. Has only a hazy knowledge of how things work,
writes about them anyway, pans others for working extra hard to deliver
and ultimately arrives at this.
Jim
ing. J.M.A. Oppers
operating system out of which NT has been developed, and
again, that operating system has been developed out of
another operating system, and that operating system was
the most beautiful operating system that has ever
existed ! But, there were more. The Japanese produced analog
computers in the 70s for solving differential equations. The
Americans or Brits in the 70s had a 'fair-child' computer.
That is a very special one. By listing to the sound of the relays
of that 'fair-child' computer I could determine with my ears
if the program and the operating system were running well !!!!
Why don't they make those computers anymore ?
Jac.
Jim Collier wrote:
>
> A poster who hears what he presumably reads writes:
>
> > I nearly fell off my chair laughing, hearing someone hint that NT might be
> > more stable than some flavor of UNIX. Even harder than when I heard the
> > suggestion that NT might be faster. Any more good jokes for us?
> >
> > -Davis
>
> A silly statement given without any justification. It does not
> surprise me that the poster hears voices when he reads posts.
>
> Given equal processors processing input data --- though it's not clear
> to me that the poster even understands the need for equal processors
> when he writes of "some flavor of Unix" --- the system which allows the
> user to write his own ISR on a high-priority interrupt line cannot be
> slower than one which doesn't, which should be pretty obvious.
>
Ken MacIver
Jr.) wrote:
>Jim Collier (pacc...@mindspring.com) wrote:
>: when he writes of "some flavor of Unix" --- the system which allows the
>: user to write his own ISR on a high-priority interrupt line cannot be
>: slower than one which doesn't, which should be pretty obvious.
>
>I'm not following you here, Are you saying a UNIX user cannot write his
>own device driver or are you saying that I/O device drivers run in user
>mode in NT.
Please, carry this talk over to alt.geek.compute or some such thing.
k
Dene Bebbington
>Dene Bebbington wrote:
>
>> Maybe this is why r.a.b's Jim Collier also supports Micro$oft from
>> criticism, they make so much money that they must be good.
>
>What a common idiot.
But one who knows when he's being sarcastic.
> Has only a hazy knowledge of how things work,
>writes about them anyway, pans others for working extra hard to deliver
>and
I don't pan others for working long hours to deliver projects, if that's
what they want, or have, to do then it's up to them. I don't believe
though that people working excessive hours is the best way deliver good
software products.
> ultimately arrives at this.
Is also ultimately able to not take everything as seriously as you do.
Michael Kagalenko
]mkag...@lynx01.dac.neu.edu (Michael Kagalenko) writes:
]
]> Jim Collier (pacc...@mindspring.com) wrote in article <363124...@mindspring.com>
]> ]I only write real-time software that works on many platforms, little
]> ]me,
]> They have my pity, each and every one of them. Working for idiots sucks.
]
]Let's just say that NT and Jim deserve each other.
Must you be so cruel, Mr.Iyer ?
Michael Kagalenko
]On 24 Oct 1998 16:15:41 GMT, j...@sapphire.alisa.org (John J. Rushford
Please, Mr.McIver, shut up when people who know their stuff are talking.
Michael Kagalenko
]>
]
]> And anyways, Collier dragged in the issue of real-time into the discussion
]> of the performance of UNIX vs NT.
]
]As a matter of fact, I do pretty much talk about what I want. When
]the topic is something I don't want to talk about, I don't more often
]than not.
Yes, I noticed that you delight in posting non-sequiturs to inapprorpiate
newsgroups.
]This would be a sound concept for you to adopt.
]
]> That shows he doesn't know what is
]> "real-time OS." It doesn't have to be faster then non-rt, and for some tasks
]> it may be slower. What matters is not that rt is fast, but that it's
]> guaranteed to be on time.
]
]Sigh. This is pretty big talk even for Mighty Mouse.
]
]The usual measure of real-time performance, Kagalenko, is thruput (in
]units of samples/sec) for some specified error.
No, it isn't. The usual measure is guaranteed responce time. The marginal
meaning is the one that you advocate.
] There is no
]other meaningful measure, or more precisely, a criterion can always be
]framed in terms of thruput and an error.
Again, false. Thruput does not determine the responce time.
Ken MacIver
wrote:
>mkag...@lynx02.dac.neu.edu (Michael Kagalenko) writes:
>
>> Ken MacIver (nan...@tiac.net) wrote in article <36321c97...@news.tiac.net>
>> ]Please, carry this talk over to alt.geek.compute or some such thing.
>>
>> Please, Mr.McIver, shut up when people who know their stuff are talking.
>
>Err... I hope you're not including Collier in that category.
Or Mikey himself.
K
Michael Kagalenko
]
]> Ken MacIver (nan...@tiac.net) wrote in article <36321c97...@news.tiac.net>
]> ]Please, carry this talk over to alt.geek.compute or some such thing.
]>
]> Please, Mr.McIver, shut up when people who know their stuff are talking.
]
]Err... I hope you're not including Collier in that category.
That goes without saying.
Jim Collier
> ]The usual measure of real-time performance, Kagalenko, is thruput (in
> ]units of samples/sec) for some specified error.
>
> No, it isn't. The usual measure is guaranteed responce time. The marginal
> meaning is the one that you advocate.
Kagalenko, it is pretty obvious why you are nothing but a closely
supervised electronic tech without the haziest knowledge of signal
processing.
By your criterion, two machines with the same processor could
have software which gives one a slightly faster interrupt latency
than the other, say 3 instruction cycles for one, and 4 for the other,
making the first machine about 30% "better" than the other. Right?
Now suppose the latency time for either is much shorter than the
reciprocal of the fastest possible rate at which an ISR can be
executed, a common situation in processors such as the TMS320 family
which mostly have one or two instruction-times interrupt latencies.
Then the only thing dividing the two kinds of processors is the rates
at which they execute ISRs.
And in fact, the main processor's interrupt latency doesn't even enter
the picture if the data has been previously sampled and buffered, as
I already pointed out. It is the lack of this knowledge which shows
that you obviously have no practical experience. So quit posturing,
you goddamned crybaby fraud and go back to your usual RAB sourbpuss
mode.
Jim
Erik Funkenbusch
Michael Kagalenko <mkag...@lynx01.dac.neu.edu> wrote in message
70r8ki$l...@lynx01.dac.neu.edu...
> And yet, NT is slower and slower with each release. NT 5 will be slower
yet.
Ummm.. Actually, NT has gotten faster with each release. 3.1 was glacial in
comparison to 3.5. 3.51 didn't add much in the way of speed improvements,
but it didn't get slower. NT 4 moved GDI and USER into kernel space (note
that this isn't moving GDI and USER into the kernel itself.. it just moved
them into the same execution context which saved a few context switches when
executing USER and GDI commands.) This led to an even greater speed
increase of interactive GUI applications (note that console applications ran
a bit faster as well, but not nearly the speed increase that GUI apps got).
NT5 promises to be the fastest yet, with more finely tuned task quantums and
better overall multi-tasking.
Now, each release of NT has required more RAM, that is true. But overall
speed has increased.
Note that I'm not defending the assertion that NT is as fast or faster than
Unix, since each Unix implementation is entirely different. There are some
Unix that are buggier than others (IRIX for instance has a ton of problems),
and FreeBSD 3.0 certainly has it's share of problems as well (but then it's
not a stable release either).
>](I mention Solaris because it has features which permit a
>]casual experimenter to bring the entire system to a halt with only
>]a modest effort. Just think what he can do when he starts writing
>]his own shells.)
>
> Mambo-jumbo argle bargle woosh-woosh.
I find it interesting that out of all the other gibberish Mr. Collier
spouts, you pick this which is possibly his clearest prose to act like you
don't understand it.
Francis Muir
>
> That goes without saying.
Sometimes I devoutly wish that all technogeek threads
would go without saying.
F
Jim Collier
obfuscate an earlier clarification:
> I think you need to study your terms as you called a 'star' network
> configuration a 'loop'.
And you knew I referred to a star because I clarified the
term, implying that I use "loop" to mean any network. See, Bill, that's
why people *do* clarify what they mean. So that some dolt won't
join in 24 hours later and say "you said loop when you meant star".
I wrote "loop" and in that particular instance, I meant a star network.
Nobody uses "loop" to refer to a daisy chain anymore, Bill, because
daisy chain networks are nearly non-existent in message passing.
> You neven mentioned bus style neworks.
That's true. I'm not interested in busses. I've got two cars.
(I have used the Metrolink a few times though.) We're not talking
about networks, Bill. We're talking about real-time operating systems.
However, taking a cue from my comments to a Mr. Kagalenko yesterday,
I encourage you to expand this conversation to take in the kitchen sink.
Just don't take me to task for not writing about what's on your mind.
Jim
Jim Collier
ROTFL. Okay, will cease and desist.
By the way (were we talking about books?), I read the usual pile of
books in the past couple of months while I was on the road from here to
Wazoo, the best of which I have referred to earlier: _Guns, Germs_ and
Steel_ by Jared Diamond, which prompted me to read _The Bell Curve_,
which is a sort of primer on racism through statistics. Studies
show that if you kidnap your average joe on the Gambia River to
America and keep the next 12 generations of his descendants caged
like fierce animals, his 13th great-grandson will be a chip off the
old block while the People of the Ice will go to the Ivies and
react badly to loud rap coming from the car in the next lane.
I already knew that.
Diamond says human intelligence has been diffused equally over the
planet. (Well, maybe the Aborigines... but let's not go there.)
Sedentary civilization, politics and ruling, warring classes could
not have arisen until there was agriculture. The land-mass
with the longest east-west axis, Eurasia of course, was ideally
suited to spread agriculture along roughly equal latitudes to
great distances from the founding site of agriculture, which
probably was Mesopotamia, where some 5 of 8 ancestors of modern
crops occurred naturally, as well as meat on the hoof which doubled
as beasts of burden, increasing mankind's hauling abilities by
multiple horsepower, a tremendous advantage.
(One advantageous side effect of farm animals: most of the communicable
childhood diseases originated in the barnyard, giving Eurasians a
disease immunity when they went to regions without wild game. Since
until very recent times, most children contracted all these diseases and
now are inoculted from them for life, they were never sources of
significant adult mortality, at least for animal-keeping Eurasians.
Spaniards took the poxes to Mexico for instance, and in the 19th
century, Brits took measles to Polynesia, both with disastrous effects
for the native populations.)
Diamond claims --- and here he draws on his experiences with New Guinea
tribesmen --- that a distinguishing characteristic of all primitive
farmers, people who are only slightly more advanced than their
hunter-gatherer ancestors, is detailed, exquisite knowledge of crops
around them. When they visit new places, the first thing they want to
know is how the locals feed themselves, how they farm. Hunger has been
a great motivator for a long, long time. If every adult was an
agronomist in ancient times, then surely the population which could feed
itself and have a surplus left over for its non-farming members was
bound to dominate over the groups who didn't but were just as hungry.
Hence by the sixteenth century, the marrauding Iberians bring down
the Incan empire even though outnumbered 100 to 1 but possessing germs,
steel swords and horses, all of which are overwhelming to the Quechuans.
Jim
Bill Vermillion
Michael Kagalenko <mkag...@lynx.dac.neu.edu> wrote:
>Jim Collier (pacc...@mindspring.com) wrote in article <363101...@mindspring.com>
> BTW, a lie about NT sometimes encountered in marketing literature
> is that it is C2- certified. This lie may be the reason for US
> military adopting NT for mission-critical stuff.
The person - whose name escapes me at the moment - who worked on
the NT security issues - is suing Microsoft. He left because he
said he wouldn't lie. The 3.5 is the certified version - not the
4.0. I have heard - but have not seen documentation - that 3.5 is
certified as C2 provided it is not connected to a network nor does
it have a modem.
--
Bill Vermillion bv @ wjv.com
Bill Vermillion
Jim Collier <pacc...@mindspring.com> wrote:
>Michael Kagalenko writes:
>
>[Good stuff about Microsoft's much told caving in on video drivers
>(for the sake of speed, of course) deleted.]
>I wrote:
>> ]Unix does not have preemptive scheduling, and given the same
>> ]processor, could hardly manage two or more concurrent tasks as fast
>> ]as NT where one requires real-time service... for instance, the
>> ]periodic interrupts from a signal processing application.
>>
>> We are not talking real-time OS here, Mr.Collier.
>Now don't get bent out of shape just because I take you seriously. I
>guess I am talking about real-time operating systems indeed.
Nowhere have I seen anything about 'real-time' systems in this
thread. Typical multi-user systems are not real-time, and typical
real-time systems are not the ones that typical users will operate.
Real-time means just that - RIGHT NOW - reaction to request,
interupt, etc. happens virually immediatly - timeing ranges from
nanosconds to microseconds. Used in control situations were
1/100th of a second response is far too slow.
Nothing MS does could be remotely classified as a real time system,
nor could most of the currently in use Unix systems. Often
real-time systems are embedded.
I think you need to study your terms as you called a 'star' network
configuration a 'loop'. You neven mentioned bus style neworks.
And the 'loop' can be implemented in several ways - many of which
will not let one system crash the net.
In discussions such as this the correct definition for words used
must be the same for all participants.
Matt Dillon
:Erik Funkenbusch <er...@visi.com> wrote:
:>
:>Michael Kagalenko <mkag...@lynx01.dac.neu.edu> wrote in message
:>70r8ki$l...@lynx01.dac.neu.edu...
:>> And yet, NT is slower and slower with each release. NT 5 will be slower
:>yet.
:>
:>Ummm.. Actually, NT has gotten faster with each release. 3.1 was glacial in
:>that this isn't moving GDI and USER into the kernel itself.. it just moved
:>them into the same execution context which saved a few context switches when
:>executing USER and GDI commands.) This led to an even greater speed
:>increase of interactive GUI applications (note that console applications ran
:>a bit faster as well, but not nearly the speed increase that GUI apps got).
Ummm. I don't want to throw a wrench on your parade but... well, I'm
going to throw a wrench on your parade.
What Microsoft is doing is really nothing more then what they've always
done: They throw out alpha software and use their customer base to 'test'
it, then they throw in improvements that are little better then hacks
(and are seriously destabilizing to the machine to boot!), and squeeze good
press out of the fact through relative comparisons. It never ceases to
amaze me how MS can put out something that stinks, then put a slightly
better release and have everyone singing its praises because it works
better then the previous release. By continuing to put out incremental
improvements they brainwash people into believing that they are making
real progress when, in fact, their products only approach their
competitor's in regards to reliability and features after half a dozen or
more iterations. MS does a good job of brainwashing people with
marketing hype and sweeping the less savory side of things (such as
security and reliability) under the rug.
What it comes down to is that NT is neither a multi-user machine, or
even that good of a multi-tasking machine. Hell, my old Amiga multitasked
a thousand times better the NT and it ran on a lowly 68000. NT doesn't
scale well at all, and its security mechanisms are a real mess... mostly
non-existent. MS is good at creating hacks to make benchmarks look good,
and they do fix bugs eventually (as in over a period of ,years)
but they do it at the cost of removing firewalling between tasks,
security, and creating workarounds to many of their own features. Most
of the hacks they do, such as putting the graphics drivers in ring 0,
no sane UNIX programmer would duplicate.
It's a win for Microsoft because they have always been a marketing-hype
driven company. If a hack allows them to make inflated claims in press
releases without getting sued, or allows various rags to run contrived
benchmarks that make NT look good, then Microsoft will do it.
Those of us who run UNIX systems tend to be more concerned about the
software operating properly, being bug free, and working well in typical
operating environments then making it look good in a contrived benchmark.
We are more concerned over manpower requirements to keep the systems
maintained. NT doesn't even have remote management (well, except for that
terrible, aweful graphics driver hack that they have been marketing as
their remote-management solution after NT admins started complaining).
The only way you can run server-based NT apps in a multi-user environment
is by duplicating the entire environment for each app. That is, putting
an out-of-process wrapper around all of NT's (non-multiuser) libraries.
It is extremely memory intensive and extremely inefficient. It's insane,
but it works. Add a few glitzy terms and the uninformed customer thinks
it's the next best thing since sliced bread and bet their company on it.
Microsoft has never given a damn about security or reliability. Their
idea of fixing a problem is to reboot the machine, and if that doesn't
work to reinstall the system. They protect their OS marketshare by
presenting a continuously moving target to potential clone competitors.
It works very well for them.
A day doesn't go by when I have to lockout yet another idiot whos installed
MS exchange on an NT box and left it wide-open to SMTP relaying, or has
installed a wide-open NT-based NNTP server. If MS were forced to pay
people for all the money it costs to deal with NT fuckups, they'd be stone
broke. I consider MS's profits blood money at the very best. Great
for investors... not so great for anyone else.
:>Note that I'm not defending the assertion that NT is as fast or faster than
:>Unix, since each Unix implementation is entirely different. There are some
:>Unix that are buggier than others (IRIX for instance has a ton of problems),
:>and FreeBSD 3.0 certainly has it's share of problems as well (but then it's
:>not a stable release either).
ummm... have you actually run FreeBSD-3.0 (aka -current) recently?
I do and I consider it just as stable as -stable. Even more so, now,
and the final release of the old -stable branch, 2.2.8, is also going to be
extremely stable. IRIX has certainly had its share of problems (I've
posted diatribes about IRIX), but even with all of its problems IRIX
is still an order of magnitude more efficient then NT. Amoung other
things, IRIX can scale to multi-cpu configurations much, much, MUCH better
then NT. And better then FreeBSD or Linux for that matter. But since
I don't have any particular need for 64 cpu minicomputers, I stick
with FreeBSD :-).
But I don't pretend that FreeBSD will kill MS. It may be a superior OS,
but it doesn't have a monopolistic driving position behind it.
Nor does Linux, but Linux has MS scared because there is nothing MS can
throw money at to run Linux out of business as they have done in other
segments (and are attempting to do, for the fifth time, in the embedded
OS market). And speaking of realtime... they got an earful when they
attempted to market Windows CE as a realtime OS. They played the
terminology loose and fast and were forced to backpedal when people called
them on it. MS will call their stuff whatever they think people what to
hear, whether it's true or not).
The Linux threat is simple: Linux is (a) popular and getting more so,
and (b) presents a mature, stable API to potential application makers.
The threat is that commercial software vendors will start to port their
applications to Linux, thus destroying Microsoft's OS monopoly. MS has
resorted to incecent lying in an attempt to derail Linux's popularity
(both in court, where they attributed virtually the entire Linux
codebase to a single person: Linus, when it is in fact the work of
thousands of programmers, and in news letters to their vendors such as the
one they sent to French vendors last week). They play both sides of the
coin.
Linux certainly isn't an MS killer. It may be an OS-monopoly-breaker, but
it isn't an MS or even an NT killer. I laugh whenever I see Microsoft
squirm over Linux... they know it won't kill them either, but they
really, really want to keep their monopoly and are willing to do just
about anything to make it happen.... and failing. Money can't buy
everything.
Microsoft continues to persue its 'moving target' OS... they are
concentrating on the only thing that the rest of us aren't willing
to do: Massive integration of everything under the sun into their
monolithic, proprietary system. They continue to attempt to take
over standards, but haven't been having as much luck as they used to.
For example, they are still pushing COM and denigrating CORBA... and
they are still using monopolistic practices such as purposefully
implementating incomplete interfaces for the open standards in order
to make their own proprietary solutions look better. They have already
been forced to backpedal on their attempt to destroy IPSEC in favor
of their own proprietary IP security model, and it looks like people
aren't happy with their attempt to destroy CORBA (CORBA verses COM)
by purposefully leaving out features to force developers to use COM.
All in all, I smile more then a frown these days when I think about
Microsoft. I feel almost sorry for them, but then I think of all that
blood money and say "Nah, they deserve whatever shit falls their way.
They've turned many people's lives into hell and I don't feel sorry
for them in the least".
-Matt
--
Matthew Dillon Engineering, HiWay Technologies, Inc. & BEST Internet
Communications
<dil...@best.net> (Please include original email in any response)
Davis
>
>A silly statement given without any justification. It does not
>surprise me that the poster hears voices when he reads posts.
>
Ahh, so now you feel the need to be pedantic regarding my every written
word. How amusing.
>Given equal processors processing input data --- though it's not clear
>to me that the poster even understands the need for equal processors
>when he writes of "some flavor of Unix" --- the system which allows the
>user to write his own ISR on a high-priority interrupt line cannot be
>slower than one which doesn't, which should be pretty obvious.
>
I see no relation between your "equal processors" comment and my "flavor
of Unix" comment, but I shall ignore this for the time being. With regard
to overall performance: I have seen *nix (I do not recall exactly which)
running on a Motorola 68030 processor. Allowing some leniency, this is
approximately equivalent to an Intel 386 (allowing no leniency, I recall
having heard of FreeBSD, among others, running on a 386). I am curious,
how well does Windows NT perform on a 386? This may be a rare case, but it
gives me the impression that resources are being used more efficiently if
one OS can be run on relatively slow hardware.
Furthermore, the research group I worked with this past summer
will be using Solaris (and even does currently) for critical data analysis
work in the coming year. It seems strange to me that they would choose
Solaris over NT, if NT is in fact more stable and faster. After all, large
amounts of data will be analyzed, so speed and stability are important (we
cannot have the system crash in the middle of massive analysis work). Both
operating systems are available to the group, as well. Are they simply
being foolish, in spite of the large number of collaborators, and
reasonably good resources behind them? Perhaps I misunderstand your use of
the term "faster."
>I had also pointed out the relatively speedy development time
>on Microsoft-based projects. The usual attack-Microsoft idiots
>have about the same mentality as the attack-Amazon morons, and
>are the same Neanderthals who attacked Toyota in the 1980s (often
>with sledgehammers) after the Japanese company introduced features
>...which American car manufacturers beginning with Ford later copied.
>
I am assuming that you are grouping me in with the "attack-Microsoft
idiots." I hate to disappoint you, but that is not entirely correct. While
I am not entirely happy with the corporation, I do continue to use their
software. I like having an OS which can run a wide variety of toys. Heck,
I even like MS Office. And this may be somewhat unrelated, but I
absolutely love Amazon.
Well, I guess my main point is that I have had many bad
experiences with Windows, and many good experiences with Unix (excluding
Irix 5.3). I doubt you will be able to convince me of the superiority of
NT, simply because experience tells me otherwise.
-Davis
Jim Collier
Davis <dd0...@uhura.cc.rochester.edu> wrote:
> On Sat, 24 Oct 1998, Jim Collier wrote:
>
> >
> >A silly statement given without any justification. It does not
> >surprise me that the poster hears voices when he reads posts.
> >
>
> Ahh, so now you feel the need to be pedantic regarding my every written
> word. How amusing.
Ah, so now he's going to try to give some justification for his earlier
statement and he won't enlist voices overheard while reading posts.
>
> >Given equal processors processing input data --- though it's not clear
> >to me that the poster even understands the need for equal processors
> >when he writes of "some flavor of Unix" --- the system which allows the
> >user to write his own ISR on a high-priority interrupt line cannot be
> >slower than one which doesn't, which should be pretty obvious.
> >
>
> I see no relation between your "equal processors" comment and my "flavor
> of Unix" comment,
You *omitted* the "equal processor" caveat. I put it in.
< but I shall ignore this for the time being. With regard
> to overall performance: I have seen *nix (I do not recall exactly which)
> running on a Motorola 68030 processor. Allowing some leniency, this is
> approximately equivalent to an Intel 386 (allowing no leniency, I recall
> having heard of FreeBSD, among others, running on a 386).
> I am curious, how well does Windows NT perform on a 386? This may be a rare >
case, but it
> gives me the impression that resources are being used more efficiently if
> one OS can be run on relatively slow hardware.
Windows NT is intended for the 486 and up --- it is pretty much used with
Pentiums --- although its so-called Hardware Abstraction Layer makes it
device independent in principle.
> Furthermore, the research group I worked with this past summer
> will be using Solaris (and even does currently) for critical data analysis
> work in the coming year. It seems strange to me that they would choose
> Solaris over NT, if NT is in fact more stable and faster. After all, large
> amounts of data will be analyzed, so speed and stability are important (we
> cannot have the system crash in the middle of massive analysis work). Both
> operating systems are available to the group, as well. Are they simply
> being foolish, in spite of the large number of collaborators, and
> reasonably good resources behind them? Perhaps I misunderstand your use of
> the term "faster."
Probably they are misinformed and two or three years behind knowledge of
what's on the market. I will further venture that some of them have used
Unix before and feel more comfortable going with the known than the unknown.
So, you argue that people whose opinions you respect use Unix and that's
why you feel it's superior to Windows NT. That's valid, I suppose. One
is prettty much stuck with the exigencies of his workplace. I have Unix
at my place even though it is not my preferred platform.
> >I had also pointed out the relatively speedy development time
> >on Microsoft-based projects. The usual attack-Microsoft idiots
> >have about the same mentality as the attack-Amazon morons, and
> >are the same Neanderthals who attacked Toyota in the 1980s (often
> >with sledgehammers) after the Japanese company introduced features
> >...which American car manufacturers beginning with Ford later copied.
> >
>
> I am assuming that you are grouping me in with the "attack-Microsoft
> idiots." I hate to disappoint you, but that is not entirely correct. While
> I am not entirely happy with the corporation, I do continue to use their
> software. I like having an OS which can run a wide variety of toys. Heck,
> I even like MS Office. And this may be somewhat unrelated, but I
> absolutely love Amazon.
Microsoft Office is a tiny part of the "ease of programming" to which I
referred. What is now done routinely in GUI program development under
Windows NT with Visual C++ or Visual Basic is still years away on any
Unix system I know of.
> Well, I guess my main point is that I have had many bad
> experiences with Windows, and many good experiences with Unix (excluding
> Irix 5.3). I doubt you will be able to convince me of the superiority of
> NT, simply because experience tells me otherwise.
>
And my company, CTI, will still be available to come and clean up the mess
when you're done. Generally, I find that one person can program in
Visual C++ and Basic in six months to replace what took two programmers to
do in X-Windows in two years. That figure of 6 person-months in exchange
for 48 person-months has been born out on a number of projects.
Of course our estimable clients think we have the benefit of their experience.
But the truth is, I ask a client what he wants on the screen and then ignore
what he's already done. He wouldn't have called us if he had made his stuff
work. On more than one occasion, I have given a client what he wanted on the
screen while he stood there and watched open-mouthed thinking of the
thousand-line X-Windows application he and his cohort had just spent several
months writing. So much for gooey GUIs.
Jim Collier
Dene Bebbington
>Windows NT is intended for the 486 and up --- it is pretty much used with
>Pentiums --- although its so-called Hardware Abstraction Layer makes it
>device independent in principle.
I wonder if the HAL acronym was accidental or deliberate.
Matt Dillon
:Jim Collier <pacc...@mindspring.com> wrote:
:>In article <Pine.SOL.3.96.981025...@uhura.cc.rochester.edu>,
:...
:>
:>> Well, I guess my main point is that I have had many bad
:>> experiences with Windows, and many good experiences with Unix (excluding
:>> Irix 5.3). I doubt you will be able to convince me of the superiority of
:>> NT, simply because experience tells me otherwise.
:>>
:>
:>And my company, CTI, will still be available to come and clean up the mess
:>when you're done. Generally, I find that one person can program in
:>Visual C++ and Basic in six months to replace what took two programmers to
:>do in X-Windows in two years. That figure of 6 person-months in exchange
:>for 48 person-months has been born out on a number of projects.
Oh please. Now you are confusing languages and operating systems with
windowing systems. Oh wait, I forgot... with Microsoft, it's all one
and the same! All integrated together so you couldn't port out of it
if your life depended on it. Maybe you should be thinking about how
easy it (apparently) was to port out of those environments and then start
thinking about how impossible it will be to port out of NT. Oh wait, I
forgot... MS and NT are going to last forever. Yah, right.
There are plenty of X API's, and X itself is pretty damn trivial if a
bit low level. That's why you use abstraction libraries. Doh.
:>what he's already done. He wouldn't have called us if he had made his stuff
:>work. On more than one occasion, I have given a client what he wanted on the
:>screen while he stood there and watched open-mouthed thinking of the
:>thousand-line X-Windows application he and his cohort had just spent several
:>months writing. So much for gooey GUIs.
:>
:>Jim Collier
Sounds like the programmer was an idiot to me, and here you are assuming
that everyone must be like that and, gee, it must always be like that and,
gee, anyone who uses UNIX *must* always do things like that and well I
guess that means you will be in business forever porting those stupid
unix programmers over to NT.
I've never heard more idiotic reasoning in my life. Microsoft has an
excellent programming environment, but it isn't leaps and bounds better
then anyone elses and, frankly, it's so highly integrated and proprietary
that I wouldn't touch it with a ten foot poll. There are plenty of people
who do... that's what a monopoly gives a company, but you won't find me
going to an environment like that and getting stuck paying MS maintenance
fees for the rest of my life!
Michael Kagalenko
]
]> ]The usual measure of real-time performance, Kagalenko, is thruput (in
]> ]units of samples/sec) for some specified error.
]>
]> No, it isn't. The usual measure is guaranteed responce time. The marginal
]> meaning is the one that you advocate.
]
]Kagalenko, it is pretty obvious why you are nothing but a closely
]supervised electronic tech without the haziest knowledge of signal
]processing.
Mr.Collier, it is quite obvious that you are classical pointy-haired boss,
ignorant and assholeish, disliked and disrespected by your unfortunate
charges. Do you think they read Usenet ? Do you imagine they snicker when
you walk by: "Here comes
the kook of the month" ? :-)
]By your criterion, two machines with the same processor could
]have software which gives one a slightly faster interrupt latency
]than the other, say 3 instruction cycles for one, and 4 for the other,
]making the first machine about 30% "better" than the other. Right?
Wrong. As I pointed out before, "faster" is not the same as "on time."
Learn to read already, Jimbo.
]Now suppose the latency time for either is much shorter than the
Jim Collier
>
> ]By your criterion, two machines with the same processor could
> ]have software which gives one a slightly faster interrupt latency
> ]than the other, say 3 instruction cycles for one, and 4 for the other,
> ]making the first machine about 30% "better" than the other. Right?
>
> Wrong. As I pointed out before, "faster" is not the same as "on time."
> Learn to read already, Jimbo.
Further, it has been pointed out for the past 15 years or
so that inerrupt latency becomes a complete non-issue in most
systems by buffering samples which were collected with a constant clock
and processing them in bursts. Zillions and zillions of tiny little
samples per interrupt. Nobody processes sampled data on a one-sample-
per-interrupt basis anymore, and even when they did, sampling was
performed by a fixed clock and occasional dropouts were tolerated.
Or not tolerated, thus spurring the use of technology which didn't
depend on interrupt latency.
Perhaps if I write this four or five more times, Mr. Kagalenko will
understand its significance. Then he'll have to find another issue
to make him miserable.
By the way, K., to give this RAB bickering some useful purpose, if
you were prescribed a desired frequency response for a digital filter,
say one with multiple passbands which aren't likely to be found
in tables, what method might you use to obtain IIR coefficients to
realize it? I'm *always* open to suggestions on this question. I'll
be (sigh) out of town for a few days, and probably won't have time to
log on for chitchat. Take as much time as you need.
Jim
Zoltan Kocsi
> Unix does not have preemptive scheduling, and given the same
> processor, could hardly manage two or more concurrent tasks as fast
> as NT where one requires real-time service... for instance, the
> periodic interrupts from a signal processing application.
Well, I think it might be very useful to you to look up the meaning
of 'preemptive' in some elementary books on multitasking.
The unix scheduler is, and has always been, preemptive.
It is not, and has never been, time deterministic.
Zoltan
--
+------------------------------------------------------------------+
| ** To reach me write to zoltan in the domain of bendor com au ** |
+--------------------------------+---------------------------------+
| Zoltan Kocsi | I don't believe in miracles |
| Bendor Research Pty. Ltd. | but I rely on them. |
+--------------------------------+---------------------------------+
Michael Kagalenko
]
]>
]> ]By your criterion, two machines with the same processor could
]> ]have software which gives one a slightly faster interrupt latency
]> ]than the other, say 3 instruction cycles for one, and 4 for the other,
]> ]making the first machine about 30% "better" than the other. Right?
]>
]> Wrong. As I pointed out before, "faster" is not the same as "on time."
]> Learn to read already, Jimbo.
]
]Further, it has been pointed out for the past 15 years or
]so that inerrupt latency becomes a complete non-issue in most
]systems by buffering samples which were collected with a constant clock
]and processing them in bursts. Zillions and zillions of tiny little
]samples per interrupt. Nobody processes sampled data on a one-sample-
]per-interrupt basis anymore, and even when they did, sampling was
]performed by a fixed clock and occasional dropouts were tolerated.
]Or not tolerated, thus spurring the use of technology which didn't
]depend on interrupt latency.
]
]Perhaps if I write this four or five more times, Mr. Kagalenko will
]understand its significance. Then he'll have to find another issue
]to make him miserable.
All this is irrelevant to the issue of real-time OSes, of which NT is not one.
So stick it somewhere.
]
]By the way, K., to give this RAB bickering some useful purpose, if
]you were prescribed a desired frequency response for a digital filter,
]say one with multiple passbands which aren't likely to be found
]in tables, what method might you use to obtain IIR coefficients to
]realize it? I'm *always* open to suggestions on this question. I'll
]be (sigh) out of town for a few days, and probably won't have time to
]log on for chitchat. Take as much time as you need.
My consulting rates are quite reasonable, but in your case, I would prefer
to collect fee in advance.
Jim Collier
> ]Further, it has been pointed out for the past 15 years or
> ]so that inerrupt latency becomes a complete non-issue in most
> ]systems by buffering samples which were collected with a constant clock
> ]and processing them in bursts. Zillions and zillions of tiny little
> ]samples per interrupt. Nobody processes sampled data on a one-sample-
> ]per-interrupt basis anymore, and even when they did, sampling was
> ]performed by a fixed clock and occasional dropouts were tolerated.
> ]Or not tolerated, thus spurring the use of technology which didn't
> ]depend on interrupt latency.
> ]
> ]Perhaps if I write this four or five more times, Mr. Kagalenko will
> ]understand its significance. Then he'll have to find another issue
> ]to make him miserable.
>
> All this is irrelevant to the issue of real-time OSes, of which NT is not one.
> So stick it somewhere.
It's only relevant if you want to process data in real time, something
in which you obviously have no experience.
Kagalenko, now that you have learned why interrupt latency is
unimportant, you don't have to posture any further. You may also
want to relate it back to your post of Friday in which you quoted a
FAQ fragment on NT as a preemptive interrupt handler... which has no
counterpart in Unix, unless of course you defeat Unix, as
unintentionally implied by your ill-informed Indian friend.
> ]
> ]By the way, K., to give this RAB bickering some useful purpose, if
> ]you were prescribed a desired frequency response for a digital filter,
> ]say one with multiple passbands which aren't likely to be found
> ]in tables, what method might you use to obtain IIR coefficients to
> ]realize it? I'm *always* open to suggestions on this question. I'll
> ]be (sigh) out of town for a few days, and probably won't have time to
> ]log on for chitchat. Take as much time as you need.
>
> My consulting rates are quite reasonable, but in your case, I would prefer
> to collect fee in advance.
I suppose a closely supervised tech at Northeastern does like to call
himself "a consultant" completely oblivious to the scorn in which
the term is held. Freelancers don't make any real money, Kagalenko,
although a job-shopper who keeps his nose clean can stay slightly
ahead of his full-time employed colleagues with their perks over the
long haul if he plans carefully. The money only starts after one takes
project responsibility, including manufacturing rights, for what one
designs. Nobody ever got rich from doing one-time projects for
clients and even poorer from being hired by the same client a
couple of years later to do a "new-and-improved" at half the cost
of the original job.
Jim
Michael Maxwell
> By the way (were we talking about books?), I read the usual pile of
> books in the past couple of months while I was on the road from here to
> Wazoo, the best of which I have referred to earlier: _Guns, Germs_ and
> Steel_ by Jared Diamond, which prompted me to read _The Bell Curve_,
> which is a sort of primer on racism through statistics. Studies
> show that if you kidnap your average joe on the Gambia River to
> America and keep the next 12 generations of his descendants caged
> like fierce animals, his 13th great-grandson will be a chip off the
> old block while the People of the Ice will go to the Ivies and
> react badly to loud rap coming from the car in the next lane.
[...]
Very nicely written. However, please check your headers; this article
has *nothing* to do with FreeBSD. Thanks.
--
drwho @ xnet.com, BOFH -- http://www.xnet.com/~drwho/
"If you don't stand for something, you'll fall for anything."
Michael Zeleny
>Michael Kagalenko wrote:
>>My consulting rates are quite reasonable, but in your case, I would
>>prefer to collect fee in advance.
>I suppose a closely supervised tech at Northeastern does like to call
>himself "a consultant" completely oblivious to the scorn in which
>the term is held. Freelancers don't make any real money, Kagalenko,
>although a job-shopper who keeps his nose clean can stay slightly
>ahead of his full-time employed colleagues with their perks over the
>long haul if he plans carefully. The money only starts after one takes
>project responsibility, including manufacturing rights, for what one
>designs. Nobody ever got rich from doing one-time projects for
>clients and even poorer from being hired by the same client a
>couple of years later to do a "new-and-improved" at half the cost
>of the original job.
What exactly is the point of your shameless preening? In the new
economy, no one will get rich from manufacturing rights to anything
that runs on the client side, thanks to the munificent practices of
your corporate patron. And since you have already demonstrated zero
awareness of server technology, you have nothing to look forward to
beyond job-shopping. Apropos of which, all the intellectual property
ownership in the world will get you zilch without an ability to make
deals in the marketplace, where the people skills you have evinced in
the process of securing the Kook of the Month title will foreclose any
prospects of significant accomplishment. Could it be that all your
impotent anger deserves to be aimed inwards, targeting yet another
superannuated and overeducated middlebrow Southern California snob?
Cordially -- Mikhail Zel...@math.ucla.edu * M...@ptyx.com ** www.ptyx.com
God: "Sum id quod sum." ** 7576 Willow Glen Road, Los Angeles, CA 90046
Descartes: "Cogito ergo sum." * 213.876.8234 (fon) * 213.876.8054 (fax)
Popeye: "Sum id quod sum et id totum est quod sum." **** www.alonzo.org
established on 2.26.1958 ** itinerant philosopher * will think for food
Dene Bebbington
>> ]By the way, K., to give this RAB bickering some useful purpose, if
>> ]you were prescribed a desired frequency response for a digital filter,
>> ]say one with multiple passbands which aren't likely to be found
>> ]in tables, what method might you use to obtain IIR coefficients to
>> ]realize it? I'm *always* open to suggestions on this question. I'll
>> ]be (sigh) out of town for a few days, and probably won't have time to
>> ]log on for chitchat. Take as much time as you need.
>>
>> to collect fee in advance.
>
>I suppose a closely supervised tech at Northeastern does like to call
>himself "a consultant" completely oblivious to the scorn in which
>the term is held. Freelancers don't make any real money, Kagalenko,
>although a job-shopper who keeps his nose clean can stay slightly
>ahead of his full-time employed colleagues with their perks over the
>long haul if he plans carefully.
Well, even though freelance contractors in this country may not
necessarily be rich, if we can keep finding contracts we're usually
quite a bit better off than our permanent employee colleagues. I'm now
grossing as a contractor about twice what I did in my last permanent job
last year, and the salary in that was on the upper side of average.
[rest snipped]
Lukas Wagner
Jim Collier <pacc...@mindspring.com> wrote:
>By the way, K., to give this RAB bickering some useful purpose, if
>you were prescribed a desired frequency response for a digital filter,
>say one with multiple passbands which aren't likely to be found
>in tables, what method might you use to obtain IIR coefficients to
>realize it? I'm *always* open to suggestions on this question. I'll
>be (sigh) out of town for a few days, and probably won't have time to
>log on for chitchat. Take as much time as you need.
>
>Jim
There exists no generic answer to this generically posed question.
For that handful of readers unacquainted with signal processing
algorithms, an engineering library might contain two to twenty
shelf feet of books filled with answers to JC's question.
Hamming's _Digital Filters_ would be one of the more readable,
though it's no longer state of the art.
Perhaps a book reference isn't what JC meant by useful purpose.
I suppose it's true what they say about hiring monkeys in silicon
valley if this is the way one looks for consultants.
--
-Lukas Wagner
wag...@pacific.mps.ohio-state.edu
Len Fehskens
> What Microsoft is doing is really nothing more then what they've
> always done: They throw out alpha software and use their customer
> base to 'test' it,
Without espousing the principle, I note that Microsoft is observing
the well established business principle that time to market matters
more than anything else. The unfortunate truth is that higher quality
products usually (though not necessarily) take longer to get to market,
and then face an uphill battle against entrenched though inferior
competition. And given two products of differing quality reaching the
market at the same time, the one with the more "in your face" marketing
will win, regardless of its quality. I may sound cynical, but the
evidence of this (the irrelevance of quality, not my cynicism) is
overwhelming.
> It never ceases to
> amaze me how MS can put out something that stinks, then put a slightly
> better release and have everyone singing its praises because it works
> better then the previous release.
"Objective" bean counters are taking over the world. There was a recent
news report about significant declines in customer perception of the
quality of customer service. But the bean counters proved right -- quality
customer service is not worth the investment if all you care about is the
"bottom line". Customers complain, but they still buy the products. More
than enough customers care more about price and early availability to make
up for any ill feelings about quality of service.
len.
Ted
> And yet, NT is slower and slower with each release. NT 5 will be slower yet.
>
slow. I Guess one will need a Zeon to run NT5.0.
Bart Lindsey
> Michael Kagalenko wrote:
> >
>
> Also not true. You need to learn details about modern operating
> systems, me friend. (I choose to answer your post only because
Modern? I recompiled my *whole* operating system a couple of weeks ago
from the *latest* updates to the FreeBSD source tree. You have to wait
6 months to a year to get the latest Service Pack from Micro$oft.
In any case, by what benchmark to you classify modern?
Did you know that the NT kernel was written by ex VMS developers and
this influenced its design? VMS is older than Unix!
If you mean support for modern devices, there are plenty of people out
there who are able to write their own drivers and share them with
others and these are often quickly ported to many free unices. It is
true to say that for most devices, the vendors supply Windoze drivers
with their products but many also supply drivers for SCO Unix and
increasingly Linux.
If you mean the graphical interface, I enjoy having the choice of the
many UIs available under X, some of which look remarkably like Windoze
if that's what you want. I must say that the KDE project's
(http://www.kde.org) GUI and file manager (explorer) was "web enabled"
years before WIN98 or NT5.0.
If you refer to sheer marketing power then I agree with you. It is
interesting though that the remarkable growth of Linux and *BSD has
taken place without the help of hundreds of millions of $$$ of
advertising. People still know a good thing when they see it.
>
> Unix does not have preemptive scheduling, and given the same
> processor, could hardly manage two or more concurrent tasks as fast
> as NT where one requires real-time service... for instance, the
> periodic interrupts from a signal processing application.
>
> There is a Unix work-around with preemption glued on. It's called
> QNX. Rather than lecture, I shall dangle it as a morsel before
> the truly curious.
>
> ObBook: William Stallings, _Operating Systems: Internals and Design
> Principles_ (3rd ed.)
Unix *does* preemptive multitask processes. Why don't you read about
it in the *many* books written about the design and philosophy of
Unix?
A good place to start is "Design & Implementation Of 4.4 Bsd
Operating" McKusick, Longman ISBN 0201549794
Such books have not been written about NT and NTs internals are a
closely guarded secret held by Micro$oft. NT would benefit greatly if
they were to open up their development process somewhat. Unix has been
evolving for over 20 years now under an open environment and has become
*very* efficient at *many* tasks as a result.
Microsoft themselves became involved with Unix through Xenix before
handing this to SCO. They must have had their reasons for discarding
Xenix, but I suspect that they felt they had less chance of cornering
markets and monopolising the computing industry.
Unix was designed to allow hundreds of people to have direct access to
ALL user programs on a computer. Each user, whether on a text based
terminal or on a remote graphical X session needn't be aware of any
other users on the system. NT is not capable of doing this without the
addition of Citrix Winframe or the like and not *all* programs will
work with it.
It is true that Unix is quite poor at real-time processing without the
types of modifications made in realtime OSes such as QNX.
I have heard that there is a real time version of NT (a cut down
version of NT, much as CE is a cut down look-a-like to Win95) in the
pipeline. This indicates to me that NT as it stands also lacks
realtime performance.
>
> Jim Collier
Bart.
Paul Ilechko
wrote:
>Did you know that the NT kernel was written by ex VMS developers and
>this influenced its design? VMS is older than Unix!
VMS + 111 = WNT
Think about it !
Paul.
***************
Paul Ilechko
http://www.transarc.com/~pilechko/homepage.htm
Ted Samsel
: By the way (were we talking about books?), I read the usual pile of
: books in the past couple of months while I was on the road from here to
: Wazoo, the best of which I have referred to earlier: _Guns, Germs_ and
: Steel_ by Jared Diamond, which prompted me to read _The Bell Curve_,
: which is a sort of primer on racism through statistics. Studies
: show that if you kidnap your average joe on the Gambia River to
: America and keep the next 12 generations of his descendants caged
: like fierce animals, his 13th great-grandson will be a chip off the
: old block while the People of the Ice will go to the Ivies and
: react badly to loud rap coming from the car in the next lane.
: I already knew that.
<stuff deleted>
: Diamond claims --- and here he draws on his experiences with New Guinea
: tribesmen --- that a distinguishing characteristic of all primitive
: farmers, people who are only slightly more advanced than their
: hunter-gatherer ancestors, is detailed, exquisite knowledge of crops
: around them. When they visit new places, the first thing they want to
: know is how the locals feed themselves, how they farm. Hunger has been
: a great motivator for a long, long time. If every adult was an
: agronomist in ancient times, then surely the population which could feed
: itself and have a surplus left over for its non-farming members was
: bound to dominate over the groups who didn't but were just as hungry.
: Hence by the sixteenth century, the marrauding Iberians bring down
: the Incan empire even though outnumbered 100 to 1 but possessing germs,
: steel swords and horses, all of which are overwhelming to the Quechuans.
Don't forget big cur dogs & mastiffs.
ObLastCurDogBreed: The Catahoula Hog Dog of Louisiana & East Texas
--
Ted Samsel....tejas@infi.net (or tbsa...@richmond.infi.net)
"do the boogie woogie in the South American way"
Rhumba Boogie- Hank Snow (1955)
Mina Kumar
People of the Ice will go to the Ivies and
> : react badly to loud rap coming from the car in the next lane.
In the Ivy I went to, it was the People of the Ice playing the loud rap,
thereby sending People of the Sun into deep discussion as to whether
Public Enemy or Dr. Dre had thereby been rendered inappropriate
car-music.
--
Mina Kumar
http://www.concentric.net/~Minak
Jim Collier
>
> Jim Collier <pacc...@mindspring.com> wrote:
> >Michael Kagalenko wrote:
>
> >>My consulting rates are quite reasonable, but in your case, I would
> >>prefer to collect fee in advance.
>
> >I suppose a closely supervised tech at Northeastern does like to call
> >himself "a consultant" completely oblivious to the scorn in which
> >the term is held. Freelancers don't make any real money, Kagalenko,
> >although a job-shopper who keeps his nose clean can stay slightly
> >ahead of his full-time employed colleagues with their perks over the
> >project responsibility, including manufacturing rights, for what one
> >designs. Nobody ever got rich from doing one-time projects for
> >clients and even poorer from being hired by the same client a
> >couple of years later to do a "new-and-improved" at half the cost
> >of the original job.
>
> What exactly is the point of your shameless preening?
The point, as was obvious to anybody involved in real-time
processing, is that Mr. Kagalenko isn't.
> In the new
> economy, no one will get rich from manufacturing rights to anything
> that runs on the client side, thanks to the munificent practices of
> your corporate patron.
Going companies don't have patrons, stupid, and if their management
wish to stay in business, they don't have a customer singular.
> And since you have already demonstrated zero
> awareness of server technology, you have nothing to look forward to
> beyond job-shopping.
ROTFL! You're into servers now, are you?
Jim
Michael Zeleny
>Michael Zeleny wrote:
>>Jim Collier <pacc...@mindspring.com> wrote:
>>>Michael Kagalenko wrote:
>>>>My consulting rates are quite reasonable, but in your case, I would
>>>>prefer to collect fee in advance.
>>>I suppose a closely supervised tech at Northeastern does like to call
>>>himself "a consultant" completely oblivious to the scorn in which
>>>the term is held. Freelancers don't make any real money, Kagalenko,
>>>although a job-shopper who keeps his nose clean can stay slightly
>>>ahead of his full-time employed colleagues with their perks over the
>>>long haul if he plans carefully. The money only starts after one takes
>>>project responsibility, including manufacturing rights, for what one
>>>designs. Nobody ever got rich from doing one-time projects for
>>>clients and even poorer from being hired by the same client a
>>>couple of years later to do a "new-and-improved" at half the cost
>>>of the original job.
>>What exactly is the point of your shameless preening?
>The point, as was obvious to anybody involved in real-time
>processing, is that Mr. Kagalenko isn't.
Ah, so. And here I got the impression of somebody pulling rank by
advertising himself as occupying a pipsqueak rung on the SBI scale.
The corporate hierarchy placement and hourly earnings that you have
volubly advertised in this forum, establish Kagalenko's ignorance of
real-time processing beyond all reasonable doubt. Silly me.
>>In the new
>>economy, no one will get rich from manufacturing rights to anything
>>that runs on the client side, thanks to the munificent practices of
>>your corporate patron.
>Going companies don't have patrons, stupid, and if their management
>wish to stay in business, they don't have a customer singular.
Fancy that. From where I sit, the impression is of your sucking up to
the high and mighty in return for favorable treatment, a nauseatingly
familiar scenario in the software developer community. As a point of
fact, numerous going companies subsist on revenue streams generated by
single clients, as Mike Morris would doubtless be happy to explain to
you. Concerning the rest of the lot, you may refer to the 80/20 rule,
right after you figure out the subtle difference between networking
and corporate clients, which appears to have interfered with your
understanding of my point.
>>And since you have already demonstrated zero
>>awareness of server technology, you have nothing to look forward to
>>beyond job-shopping.
>ROTFL! You're into servers now, are you?
I run a company that specializes in back end web software development.
Jim Collier
I don't claim to make an "hourly earning", stoopid, and therefore
the occasion for "volubly advertising" an hourly earning wouldn't have
arisen, would it now? I did say we do pretty well, "we" meaning
my company. Another student --- you are a full time student
yourself, I believe, Mr. Zeleny --- took the position that this
is unwarranted bragging. Who gives a rat's ass what the idiot thinks?
For that matter, who gives a rat's ass what any 22-year-old thinks?
Now, young man, try to follow this chain. I wrote that most Unix
systems are not preemptive. (For Zeleny: what is "preemptive" in this
context? Other readers: please hold off until Zeleny responds.) NT is
preemptive. (For Zeleny only: what advantage might preemption offer?)
I pointed out that for real-time processing, the main criterion
is thruput. (For Zeleny only: what is thruput? What is its
relationship with the so-called Nyquist criterion?) Kagalenko
volunteered, after somebody else suggested it first, that the only
meaningful measure of a real-time system's performance is interrupt
latency. (For Zeleny only: what is interrupt latency, and did
the original poster actually have something slightly different
in mind but closely related to interrupt latency?) I pointed out that
in practical systems, that periodically sampled data are processed in
bursts per interrupt, making the actual time of interrupt relatively
unimportant. (For Zeleny only: prove or disprove this with the
aid of an example.)
Kagalenko then argued without proof: nope, it's interrupt latency.
Then, it occurred to me that he hadn't been around real-time systems,
any more than say you or my cats have. I asked him a verification
question and as I expected he would do, he answered with an evasion,
as I expect you will do. You can surprise me. I like talent
in kids whom I had previously considered idiots and moral cowards.
Then, you said that I hadn't said anything about servers. Well, hell's
bells, what *do* you know about servers?
> I run a company that specializes in back end web software development.
Oh, dear me. Zeleny, I'm not talking about lemonade stands and
parents' spare bedrooms, and Java-programming (or whatever) projects
for a couple of friends that you met in the ASUCLA cafeteria.
You make the fourth or fifth juvenile I have encountered just today
who writes Web software. Nyquist, watch out!
But Zeleny, a server is related to the issues of preemption and
real-time processing roughly the way that coconut palms are related
to R-2R ladders. You are indeed appropriately named.
Jim Collier
Michael Zeleny
antipodean namesake. Truly of a piece with compelling apodeixis aimed
at Dene Bebbington in http://www.dejanews.com/getdoc.xp?AN=400117732:
DB:
##The main problem I have with Microsoft is
##that the Operating Systems such as Win95 and NT are so resource
##demanding in comparison to something like UNIX or Linux.
JC:
#Unix requires a virtual memory (and uses it!) and NT requires
#a virtual memory. What's the big deal? This is 1998 and memory
#is hardly worth the price of discussion. I can buy a 2-gigabyte
#hard drive, an obsolete size, for what I make in an hour, or
#33 megabytes/minute. Really, modern operating systems aren't large
#enough to take advantage of available resources.
Compare your witty disclaimer infra:
>I don't claim to make an "hourly earning", stoopid, and therefore
>the occasion for "volubly advertising" an hourly earning wouldn't have
>arisen, would it now?
Concerning lemonade stands and parents' spare bedrooms, and giving a
rat's ass what any 22-year-old thinks, I suppose it is flattering to
be mistaken for a callow youth at my advanced age. Alas, niggling
scruple compels me to aver having worked in the software industry for
over half my life, since 1977. Regrettably, I cannot take sides in
your most entertaining signal processing controversy, since real-time
computing does not fit within my core competence. However, based on
too many years of managerial and entrepreneurial experience, I hereby
offer my heartfelt condolences to all of your employees and business
associates. The original reason I brought up servers and scalability
was to countermand your silly prediction of Windows NT driving Unix
out of the marketplace. (Remember Hotmail?) And since my modest
hourly earnings can buy an entire 8 gigabyte hard drive, you are quite
welcome to prostrate yourself before my jackboots, in accordance with
your manifest standard of discursive solicitude.
That thing you do with epenthetic epithets and evident eponyms --
is it your second childhood or congenital poverty of imagination?
xxxooo
-- MZ
Cordially -- Mikhail Zel...@math.ucla.edu * M...@ptyx.com ** www.ptyx.com
Dene Bebbington
[snips]
>Now, young man, try to follow this chain. I wrote that most Unix
>systems are not preemptive. (For Zeleny: what is "preemptive" in this
>context? Other readers: please hold off until Zeleny responds.) NT is
>preemptive. (For Zeleny only: what advantage might preemption offer?)
I don't have the article still on my newsreader, but IIRC you actually
said that UNIX does not have preemptive scheduling. Thus your asking for
the context is redundant since you originally referred to scheduling
thus specifying the context you had in mind.
Jim Collier
>
> DB:
> ##The main problem I have with Microsoft is
> ##that the Operating Systems such as Win95 and NT are so resource
> ##demanding in comparison to something like UNIX or Linux.
>
> JC:
> #Unix requires a virtual memory (and uses it!) and NT requires
> #a virtual memory. What's the big deal? This is 1998 and memory
> #is hardly worth the price of discussion. I can buy a 2-gigabyte
> #hard drive, an obsolete size, for what I make in an hour, or
> #33 megabytes/minute. Really, modern operating systems aren't large
> #enough to take advantage of available resources.
>
> Compare your witty disclaimer infra:
>
> >I don't claim to make an "hourly earning", stoopid, and therefore
> >the occasion for "volubly advertising" an hourly earning wouldn't have
> >arisen, would it now?
>
Again for the poster's benefit in the event he should realize a
recurring profit in years to come, I am *not* paid by the hour.
Nor are most business owners.
> Concerning lemonade stands and parents' spare bedrooms, and giving a
> rat's ass what any 22-year-old thinks, I suppose it is flattering to
> be mistaken for a callow youth at my advanced age. Alas, niggling
> scruple compels me to aver having worked in the software industry for
> over half my life, since 1977.
Point taken there, Zeleny. I should have written "Internet-posting
22-year-olds and 42-year-olds with lemonade stands.
Jim Collier
Ken MacIver
wrote:
>zel...@oak.math.ucla.edu (Michael Zeleny) writes:
>
>> >I don't claim to make an "hourly earning", stoopid, and therefore
>> >the occasion for "volubly advertising" an hourly earning wouldn't have
>> >arisen, would it now?
>>
>> Concerning lemonade stands and parents' spare bedrooms, and giving a
>> rat's ass what any 22-year-old thinks, I suppose it is flattering to
>> be mistaken for a callow youth at my advanced age.
>
>Any bets on how long before Collier starts complaining about rampant
>age discrimination in the tech industry?
Perhaps sympathy is in order, for he seems more bitter than the
subject would seem otherwise to warrant.
ObSaying: "You can find sympathy in the dictionary between shit and
syphllis." Captain Moffatt, Georgetown, Maine (looking at his boat
FOGGY DAZE, which sunk at the mooring during a northeaster).
Ken
Ken MacIver
wrote:
>nan...@tiac.net (Ken MacIver) writes:
>
>> On 31 Oct 1998 10:53:53 -500, Rajappa Iyer <raj...@mindspring.com>
>> wrote:
>> >Any bets on how long before Collier starts complaining about rampant
>> >age discrimination in the tech industry?
>>
>> Perhaps sympathy is in order, for he seems more bitter than the
>> subject would seem otherwise to warrant.
>
>Perhaps. Although `bitter' is not the word that springs to mind
>before `insecure.'
I dunno. It's contextual, depends on whether something is perceived
to have happened in the past(hence, bitter) or whether it's
anticipatory (hence, insecure).
K.
Michael Zeleny
>Michael Zeleny writes:
>>DB:
>>##The main problem I have with Microsoft is
>>##that the Operating Systems such as Win95 and NT are so resource
>>##demanding in comparison to something like UNIX or Linux.
>>
>>JC:
>>#Unix requires a virtual memory (and uses it!) and NT requires
>>#a virtual memory. What's the big deal? This is 1998 and memory
>>#is hardly worth the price of discussion. I can buy a 2-gigabyte
>>#hard drive, an obsolete size, for what I make in an hour, or
>>#33 megabytes/minute. Really, modern operating systems aren't large
>>#enough to take advantage of available resources.
>>
>>Compare your witty disclaimer infra:
>>
>>>I don't claim to make an "hourly earning", stoopid, and therefore
>>>the occasion for "volubly advertising" an hourly earning wouldn't have
>>>arisen, would it now?
>
>Again for the poster's benefit in the event he should realize a
>recurring profit in years to come, I am *not* paid by the hour.
>Nor are most business owners.
>
>>Concerning lemonade stands and parents' spare bedrooms, and giving a
>>rat's ass what any 22-year-old thinks, I suppose it is flattering to
>>be mistaken for a callow youth at my advanced age. Alas, niggling
>>scruple compels me to aver having worked in the software industry for
>>over half my life, since 1977.
>
>
>Point taken there, Zeleny. I should have written "Internet-posting
>22-year-olds and 42-year-olds with lemonade stands.
I see that Collier has been attending classes in Plausible Deniability
at the Mike Godwin Institute of Usenet Demagoguery. The fact remains
that his superannuated enterprise has been outclassed in profitability
by my modest lemonade stand. Perhaps Kagalenko's Konsulting Klique can
help our hapless hero measure up to the competition. Are you still
offering your KoTM special, Michael?
Jim Collier
>
> nan...@tiac.net (Ken MacIver) writes:
>
> > On 31 Oct 1998 10:53:53 -500, Rajappa Iyer <raj...@mindspring.com>
> > wrote:
> > >Any bets on how long before Collier starts complaining about rampant
> > >age discrimination in the tech industry?
> >
> > Perhaps sympathy is in order, for he seems more bitter than the
> > subject would seem otherwise to warrant.
>
> Perhaps. Although `bitter' is not the word that springs to mind
> before `insecure.'
Well, little man, after all is said, I think it is pretty well
established now that you aren't involved in real-time processing any
more than Kagalenko or Zeleny. So you had nothing to contribute to
this latest brawl from the start except performing the noisy yip-yip
function. Why don't you and the TV chihuahua go do some serious
battle for a Taco Bell treat?
Jim Collier
Jim Collier
> The fact remains
> that his superannuated enterprise has been outclassed in profitability
> by my modest lemonade stand.
Zeleny, I quite seriously doubt that your parttime spare bedroom
lemonade stand does $10 million a year in business. Assuming you have
formed a Subchapter "S" corporation, I would even be surprised to learn
that you net more than enough to cover depreciation taken as Section 179
deductions, like the great majority of part time "consultants" who
work all their hours outside their day jobs to avoid paying a couple
of extra bucks to Sam and get a shiny new PC in the process... not to
mention being able to pass out business cards to the corner butcher,
baker, and candlestick maker, and anyone willing to listen to
a quarter-hour of bullshit.
Jim Collier
Jim Collier
> Fancy that. From where I sit, the impression is of your sucking up to
> the high and mighty in return for favorable treatment, a nauseatingly
> familiar scenario in the software developer community.
What drivel. Many a small-time, no-discount supplier with sun-bleached
merchandise in the window of his 750 sq. ft. storefront and a shared loo two
doors down would love to meet this guy's money and him. Or if both are not
available, then his money without him.
Now, Zeleny, if you're going to tell people that you have a business, the
first thing you have to know about customers is that the *only* thing they
care about is getting what they pay you for, which is probably not too
different from the way you feel about places where you shop.
> As a point of
> fact, numerous going companies subsist on revenue streams generated by
> single clients, as Mike Morris would doubtless be happy to explain to
> you.
Small-timers sometimes think that, up until the day after their clients change
suppliers. (The change may occur through no fault of the supplier. A
department head changes, a buyer drops dead on the weekend, the client company
is sold. Happens to all business relationships sooner or later unless the
parties are immortal.) Any bizman who gives a single customer an exclusive
should explain to that customer that he has to reciprocate by putting the
entrepeneur on the company payroll.
Jim Collier
-----------== Posted via Deja News, The Discussion Network ==----------
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
Jim Collier
>
> In article <71e886$ia0$1...@carroll.library.ucla.edu>,
>
> ends," my reaction was "Of *course* he does."
Y'know I looked at that phrase for 90 seconds and finally thought
to myself, no, that would be a cheap shot.
Jim Collier
Jim Collier
>
> You are mistaking gross revenues for profitability, a common failing
> among the innumerate.
>
You are mistaking a spare bedroom lemonade stand for a business.
> And my hard drive is still bigger than yours.
If I know you're in the neighborhood, I'll be sure to call in the
cats.
Jim Collier
Jim Collier
>
> On Sat, 31 Oct 1998
> 00:14:46 -0800, Jim Collier <pacc...@mindspring.com> wrote:
> >Now, young man, try to follow this chain. I wrote that most Unix
> >systems are not preemptive. (For Zeleny: what is "preemptive" in this
> >context? Other readers: please hold off until Zeleny responds.) NT is
> >preemptive. (For Zeleny only: what advantage might preemption offer?)
>
> That's what the OED says and, in the context of Microsoft, it brings to
> mind Tacitus comment on those in the Senate, loudly praising a new Emperor
> without any hope of advantage, "but merely for love of servility".
>
> >I pointed out that for real-time processing, the main criterion
> >is thruput. (For Zeleny only: what is thruput? What is its
> >relationship with the so-called Nyquist criterion?) Kagalenko
>
>
> ---------------------------------
> Victor Yodaiken
> Department of Computer Science
> New Mexico Institute of Mining and Technology
> Socorro NM 87801
> Homepage http://www.cs.nmt.edu/~yodaiken
> PowerPC Linux page http://linuxppc.cs.nmt.edu
> Real-Time Page http://rtlinux.org
Actually, the http://rtlinux site is pretty decent.
Jim Collier
Mike Godwin
Michael Zeleny <zel...@oak.math.ucla.edu> wrote:
>>ROTFL! You're into servers now, are you?
>
>I run a company that specializes in back end web software development.
When I read that the famously homophobic Zeleny specializes in "back
ends," my reaction was "Of *course* he does."
--Mike
--
----------------------------------------------------------------------------
"I speak the password primeval .... I give the sign of democracy ...."
--Walt Whitman
yoda...@chelm.cs.nmt.edu
Michael Zeleny
>Michael Zeleny wrote:
You are mistaking gross revenues for profitability, a common failing
among the innumerate. And my hard drive is still bigger than yours.