Searching \ for '[OT]: Why MS$ Is Hated' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: techref.massmind.org/techref/index.htm?key=why+hated
Search entire site for: 'Why MS$ Is Hated'.

Exact match. Not showing close matches.
PICList Thread
'[OT]: Why MS$ Is Hated'
2003\06\27@115505 by Bob Axtell

face picon face
I realize that many younger guys don't understand why MS$ is so hated by so
many of us older guys.  Its because we saw, intimately, how MS$ managed to
flimflam so many people and get away with it.

A small gang of programmers disassembled Digital Research's CPM, tinkered
with it a little, then came up with IBM DOS 1.0. They then worked an odd
deal with IBM to compete against IBM with their own versions, that promised
to fix the problems they'd left in the first version. Once that pattern of
deception started, it became a pattern for the creation of vast wealth-
just create poor code with a little bit of razzle-dazzle, then blame the
resulting mess on everything and everyone else. It worked.

If you don't believe me, read the archives of early PC magazines. Watch how
MS$ handled the press, thru Win3.0, Win3.1, Win95, Win98, WinME (cringe)
then to Win2K and WinXP. Now, in all fairness, once the lawsuit was
enetered, MS$ finally began fixing the core, and Win2K and WinXP work
pretty well. But even Win2K with all shots (17 major and minor revisions)
doesn't match up to Linux on its best day.

I was forced by clients to endure Windows, and if I'd had the money I'd
shelled out for legal copies of Windows over the years, I'd have been able
to marry and divorce a few more wives.

So, for you youngsters that weren't there when it all happened, that's why
some of us hate MS$ so much.

--Bob


At 11:24 AM 6/27/2003 -0400, you wrote:
> > Oh you mean M$ has finally caught up with Borland, who have been
> > doing this
> > for years :))))

---------------
NOTICE

1. This account can accept email & attachments up to 10M in size.
2. Federal Monitors: At request of client, some attachments are encrypted.
Please DO NOT delay traffic; please reply with credentials for password.
--------------

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics

2003\06\27@134617 by T.C. Phelps

picon face
As well, a lot of other neat OSes might have been
around these days were it not for MS's death grip on
the computing market (thankfully Linux got in so
early, who knows otherwise?). Remember OS2, GEOS,
BEOS....

The thing I always found funny was their commercial
for Windows 95, where they showed four animated
symbols -- "In conventional OSes, each task operates
for only a small portion of time <showing each
animation running in turn for about a second> whereas
Windows 95 lets them all run simultaneously <shows all
the animations running concurrently>."

Of course Win95 still used timeslicing like Windows 3
and earlier used (at least I'd think so), except now
you couldn't change the length of the slice. "Since ya
can't see it, it must NOT be doing it!" Right? :)

- T.C.


--- Bob Axtell <spam_OUTcr_axtellTakeThisOuTspamYAHOO.COM> wrote:
{Quote hidden}

__________________________________
Do you Yahoo!?
SBC Yahoo! DSL - Now only $29.95 per month!
http://sbc.yahoo.com

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics

2003\06\27@141941 by Greg Miller

flavicon
face
On Fri, 27 Jun 2003 10:45:47 -0700 you wrote:

>Of course Win95 still used timeslicing like Windows 3
>and earlier used (at least I'd think so), except now
>you couldn't change the length of the slice. "Since ya
>can't see it, it must NOT be doing it!" Right? :)

    Win 3.x and below used cooperative multitasking, where each application ran for as long as it wanted to, then relinquished control back to the kernel.  Windows 95 and NT introduced the ability to interrupt an application before it relinquished control (preemtive multitasking).  But, of course, only one application is actually running at any given moment.

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics

2003\06\27@155125 by Andrew Warren

flavicon
face
Bob Axtell <.....PICLISTKILLspamspam@spam@mitvma.mit.edu> wrote:

> A small gang of programmers disassembled Digital Research's CPM,
> tinkered with it a little, then came up with IBM DOS 1.0.

   Um, no.  Tim Paterson wrote an 8086 DOS called QDOS (for "Quick
   and Dirty DOS"), distributed it briefly through the company he
   worked for (Seattle Computer), then sold it to Microsoft and went
   to work for them.  While there, he modified it to meet IBM's
   requirements for PC-DOS, and that's when it started to really
   look like CP/M.

   The notion that Paterson disassembled CP/M and "tinkered with it
   a little" to make QDOS is absurd; QDOS was written before the
   first IBM PC was built, and CP/M-86 wasn't available (as I
   recall) until after the PC-AT shipped.  In my opinion, the
   resemblance between PC-DOS and CP/M is really only due to a)
   IBM's desire for PC-DOS to have a familiar interface, and b) the
   fact that any OS that fit in 16K of RAM was probably going to
   look a lot like every other 16K OS, anyway.

> Now, in all fairness, once the lawsuit was enetered, MS$ finally
> began fixing the core, and Win2K and WinXP work pretty well.

   In all fairness, it wasn't the antitrust lawsuit that made Win2K
   and WinXP so good; it's the fact that they're built on WinNT,
   which was written by a team led by Dave Cutler, the architect of
   DEC's VMS operating system.

   If you want to accuse Microsoft of inappropriately using someone
   else's code, you could make a better case for it with WinNT than
   with PC-DOS 1.0.  Ever wonder why Windows NT supported the
   obscure DEC Alpha microprocessor as well as the Intel x86?
   Microsoft agreed to support the Alpha in return for DEC's
   agreeing not to sue over the Windows NT source code, much of
   which was, allegedly, a verbatim copy of DEC's VMS source.

   -Andy

=== Andrew Warren -- aiwspamKILLspamcypress.com
=== Principal Design Engineer
=== Cypress Semiconductor Corporation
===
=== Opinions expressed above do not
=== necessarily represent those of
=== Cypress Semiconductor Corporation

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics

2003\06\27@160200 by Dal Wheeler

flavicon
face
Hehe, I still run BEOS on one of my boxes; it's about the only thing that'll
run reasonably on that runt of a machine.

Anyway, one thing I noticed when I put out that link to a webserving port a
couple of days ago was that a full half of the people that looked at my
pictures weren't using the standard microsoft software.  Probably less than
half of the unique IP's were using windows incarnations and a fairly large
subset of that where using mozilla or opera.  There where a lot of Apple
OS9,OS X/safari/IE users--which kind of surprised me.  O'course there where
many linux users filling up the rest.  Anyway, I guess it just verifies that
the techies on this list are quite a bit more likely to use non-mainstream
OS/webtools than the greater computer user population.
-Dal

----- Original Message -----
From: "T.C. Phelps" <.....tcphelpsKILLspamspam.....YAHOO.COM>

> As well, a lot of other neat OSes might have been
> around these days were it not for MS's death grip on
> the computing market (thankfully Linux got in so
> early, who knows otherwise?). Remember OS2, GEOS,
> BEOS....

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics

2003\06\27@165523 by Peter Moreton

flavicon
face
It's true about the VMS = NT thing. I ran VMS systems for years, and it
never
ceases to amaze me how much of VMS appears in NT/2000/XP. Even the
pagefile
has the exact same name. Mind you, VMS v7.0 is still a more stable,
secure and
scalable O/S than 2K/XP. An OS that installs Kernal-Mode device drivers
on the
fly (when a plug and play event occurs) is never going to be truly
stable.


> {Original Message removed}

2003\06\27@170611 by T.C. Phelps

picon face
--- Dal Wheeler <EraseMEdwheelerspam_OUTspamTakeThisOuTINSIGHTEK.NET> wrote:
> Hehe, I still run BEOS on one of my boxes; it's
> about the only thing that'll run reasonably on that
> runt of a machine.

I really missed the boat on that one, it looked great
but I guess at the time (IE, when it was in
production) I figured I didn't need it.

<snip...>
> Anyway, I guess it just verifies that
> the techies on this list are quite a bit more likely
> to use non-mainstream OS/webtools than the greater
> computer user population.

Sure, it's nice to be able to change something if your
OS is misbehaving. I wonder how many of those
alternate-OSers were "converts", forced to use
something else at work or school and then finding they
liked the alternative. I had to switch from Windows 95
to Redhat when I first when to University, and I
wasn't overly fond of Linux at first(I was brought up
using simple things like DOS 3 etc, and I thought of
Linux as being a bit, well, ostentatious I suppose --
so many instructions, so many options). Here it is a
few years later, and whenever I decide to code
something totally new to me, I like going with Linux
because it eliminates one source of unknowns! Couldn't
imagine it under something like ME -- it crashed my
laptop so often I'd never know if it was my program or
the OS misbehaving.


- T.C.

__________________________________
Do you Yahoo!?
SBC Yahoo! DSL - Now only $29.95 per month!
http://sbc.yahoo.com

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics

2003\06\27@174344 by John Ferrell

face picon face
You make the best case I have heard for going to WinXP.
It is tough for me to believe that it can be as robust as VMS.

John Ferrell
6241 Phillippi Rd
Julian NC 27283
Phone: (336)685-9606
johnferrellspamspam_OUTearthlink.net
Dixie Competition Products
NSRCA 479 AMA 4190  W8CCW
"My Competition is Not My Enemy"

{Original Message removed}

2003\06\27@182349 by William Chops Westfield

face picon face
   It's true about the VMS = NT thing. I ran VMS systems for years, and
   it never ceases to amaze me how much of VMS appears in
   NT/2000/XP. Even the pagefile has the exact same name.

That's standard geek humor, though, and doesn't mean anything about whether
code was actually copied or not.  For instance, the cisco IOS operating
system source has file and function names that are "the same" as the DEC
tops20 operating system, because no few of the early cisco software
engineers were former DEC20 programmers.  Of course, the DEC code was
assembler, and the cisco code is C, so aside from a concept here and there,
not much was copied other than the names.  (Still, I sent out an April-1
joke "cisco licenses tops20 from DEC" some years ago, and had quite a few
people who appeared to believe it...)

I think fundamentally Microsoft is hated because they're held by the public
to be a major instance of "techie geek success story", but in fact much of
their success is due to marketing prowess (both moral and otherwise.)  Their
products are rarely "the best" technically, and frequenly don't originate
internal to microsoft (ie MSDOS), Microsoft just makes sure that they are
shoved down your throat, and/or eliminates competition.

Particularly obnoxious recent example - microsoft deciding to bow out of the
Macintosh web browser market.

BillW

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics

2003\06\27@182555 by Douglas Wood

picon face
Actually, Tim Paterson wrote QDOS to be a CP/M clone. The idea was to create
a more or less instant software base by giving Aston-Tate, Wordstar, etc. an
8080-to-8086 translator and an 8086 assembler. Presto! All of the CP/M
titles now available for the new 8086 platforms just starting to show up.

In fact, although Mr. Gates denied MS-DOS' heritage from CP/M, MS-DOS
continued to support the "call 0x0005" entry point well into the v3.x's.

The big thing that Microsoft did to QDOS was to make the "user areas" under
CP/M behave more like subdirectories under UNIX. But that change didn't
happen until v2.0.

Douglas Wood
Software Engineer
@spam@dbwoodKILLspamspamkc.rr.com
ICQ#: 143841506

Home of the EPICIS Development System for the PIC
http://epicis.piclist.com

{Original Message removed}

2003\06\27@194556 by Michael Davidson

flavicon
face
>Hehe, I still run BEOS on one of my boxes; it's about the only thing that'll
>run reasonably on that runt of a machine.

It's still my main OS on my desktop (ie. This machine) and it will be until
someone can give me an OS half way as user friendly. If you've ever
programmed for BeOS you'll know where I'm coming from as well.

It's really a pity that Be had to go bye-bye. Of interest to people who are
interested in BeOS is;
http://www.yellowtab.com <-- Licensed source from Be / PalmSource for what
was known as "5.1" / "Dan0" and are releasing a new incarnation known as
"Zeta"
http://www.openbeos.org <-- A bunch of guys writing a BeOS compatible open
source (BSD) replacement


--

When you're not looking at it, this fortune is written in FORTRAN.

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics

2003\06\28@051923 by Peter L. Peres

picon face
I don't think M$ is 'hated', its OS's just does not solve many people's
problems (the more technical their problems the less they solve them).

I was (and still am) shocked at the claims of Windows being a
'multitasking' OS. My first dawning to this came when I first ported a
honest BSD style forking server to Windows and I found out that there is
no fork() call and that the equivalent involves a lot of mucking with
threads and other abominations. Judging by the number of parameters each
library call requires I'd say they're paid by complexity or something like
that. Ever since I got familiar with that I understood why a good
programmer can turn out flaky windows code and fairly good *nix code. *nix
is not an innocent baby either (see file locking, shared memory and
signals f.ex.). But then I also read about the Windows sheduler (which
works like an Enigma encoder wrt determinisim of timing, only better), and
promptly gave up all my projects that used PC port timing directly (they
still work under Linux & Co where I have access to timing).

So I just think there are 'some' problems with applying W* to hardware
control and reasonably reliable software.

Peter

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.

2003\06\28@074704 by Peter Moreton

flavicon
face
'Standard geek humor'?  I relay only my experiences as a OpenVMS systems
programmer who moved on to NT 3.5.. 4.0 etc.  Dave Cutler probably did
not copy the OpenVMS code base, but it is undeniable that much of his
experience gained with RSX and then OpenVMS was used in the development
of OpenVMS. If you 'lift the hood' of both OpenVMS and NT, the
similarities are quite remarkable. This is fact, not geeky stuff.

Anyone who wishes to make comparisons, can read the VMS Internals
textbook by Helen Custer, and the 'Inside Windows NT' text by, err,
Helen Custer.

Rgds, Peter Moreton


> {Original Message removed}

2003\06\28@082112 by John Ferrell

face picon face
Windows could avoid a lot (most?) of the blue-screen criticism by providing
some indication of where the trouble occurred. It might take a little
software effort but as it is, it always appears that the culprit is windows.
I have (two cases) solved blue screens by reducing clock speed on CPU's. I
doubt the root problem was software.


John Ferrell
6241 Phillippi Rd
Julian NC 27283
Phone: (336)685-9606
KILLspamjohnferrellKILLspamspamearthlink.net
Dixie Competition Products
NSRCA 479 AMA 4190  W8CCW
"My Competition is Not My Enemy"

{Original Message removed}

2003\06\28@082725 by Michael Davidson

flavicon
face
>That's standard geek humor, though, and doesn't mean anything about whether
>code was actually copied or not.

You've gotta love geek humour. Below is an excerpt from "The BeBook" (The
BeOS programming bible);

"int32 is_computer_on(void)

Returns 1 if the computer is on. If the computer isn't on, the value returned
by this function is undefined."

and

"double is_computer_on_fire(void)

Returns the temperature of the motherboard if the computer is currently on
fire. If the computer isn't on fire,
the function returns some other value. "

Michael
--

You can make it illegal, but you can't make it unpopular.

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.

2003\06\28@091732 by Tal

flavicon
face
Regarding the VMS = NT thing, if you take the three letters VMS and
increments each by 1 you get exacly WNT. This is similar to the HAL/IBM
the the movie Odisea 2001. Read about some time ago but don't know if
this was intentional.

Tal

> {Original Message removed}

2003\06\28@094902 by Tal

flavicon
face
Peter,

Forking is a poor man multithreading. It was developed when in-process
light threads was rare in full fledge operating systems (as opposed to
real-time kernels). Forking has higher overhead because of the
replication of the entire data space of the project and requires more
complex synchronization between the threads because of the address space
separation.

In a modern OS, my first choose for multi threading will be in process
light threads.

BTW, I spent few years developing real time applications on Windows NT
(3.51 and then 4), and various Unix's. At least at that time, Windows NT
provided a way better set of primitives to develop multithreading
applications (not to mention a better development environment).

Tal

> {Original Message removed}

2003\06\28@095322 by Mike Singer

picon face
  These north-Americans who hate MS, who are they seem
to be from the overseas point of view?
Just stupid masochists, I'd suggest.
Tens milliards bucks are collected into north-American
economics from all over the Globe each year to let those
guys eat well, to be healthy, to let their mothers and
fathers feel safe when approaching retirement age and
so on...

Mike.

P.S. Don't tell me about moral side of the story.
At lest they don't bombard for not buying their stuff.

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.

2003\06\28@122502 by Bob Ammerman

picon face
fork() is far less efficient than Windows type threading.

Notice that unix based OS's have added threading packages, but I've not seen
anything like Windows trying to do fork().

Bob Ammerman
RAm Systems

{Original Message removed}

2003\06\28@130209 by D. Jay Newman

flavicon
face
> fork() is far less efficient than Windows type threading.

The system call that Windows has is closer to spawn(). Both are inefficient
compared to threading.

As I understand it Window really has two different types of threads, one
which more closely corresponds to processes and the other to threads.

> Notice that unix based OS's have added threading packages, but I've not seen
> anything like Windows trying to do fork().

I believe that threading started with unix, though not as an operating
system call, but rather a library. However, I'm not sure as to the dates.

> Bob Ammerman
> RAm Systems
--
D. Jay Newman          !
RemoveMEjayTakeThisOuTspamsprucegrove.com    ! Xander: Giles, don't make cave-slayer unhappy.
http://enerd.ws/~jay/  !

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.

2003\06\28@151116 by Mike Singer

picon face
MS is not my friend, I'm just trying to figure out where
Truth is located.

Peter L. Peres wrote:
> I don't think M$ is 'hated', its OS's just does not solve
> many people's problems"

Nothing is perfect. Definitely there are a lot of " people's
problems" they can't solve: cold weather for example.

> (the more technical their problems the less they solve
> them).

You meant, I suspect, "hardware" not "technical" problems.
"Technical" word can be applied to software as well.
Speaking of hardware Win is not positioned as RTOS. In fact
there is a pretty serious sector on software market where
real-time things doesn't matter.

> I first ported a honest BSD style forking server to Windows
> and I found out that there is no fork() call and that the
> equivalent involves a lot of mucking with threads and other
> abominations. Judging by the number of parameters each
> library call requires I'd say they're paid by complexity
> or something like that.

Your problems may not be common problems Windows is intended
to solve. Did you scan all types of commercial projects to
compare Win with other systems?

> So I just think there are 'some' problems with applying
> W* to hardware control and reasonably reliable software.

Nothing revealed. Windows is not intended to control
hardware. Software reliability problems are different to
hardware control problems.

Mike.

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.

2003\06\28@161206 by Tal

flavicon
face
> Nothing revealed. Windows is not intended to control
> hardware. ...
>
> Mike.

Now I understand why Windows refuses to control the new printer I just
installed.

;-)

Tal

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.

2003\06\28@195154 by Bob Ammerman

picon face
> The system call that Windows has is closer to spawn(). Both are
inefficient
> compared to threading.

spawn() is basically fork() followed by exec(), IIRC. On Windows it is
CreateProcess().

WIN32 does and always has supported true multithreading within one process
(address space). Note that for many years VMS did not support this (although
it did support 'AST's or Asynchronous System Traps, sorta like software
interrupts into a user application).

WIN32 has APC's, which are basically the same as ASTs!

Windows also now supports a concept called 'fibers', which are cooperatively
multitasked threads within a process. They are highly efficient and
particularly useful in a transaction processing environment.

Bob Ammerman
RAm Systems

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.

2003\06\29@041356 by William Chops Westfield

face picon face
>> fork() is far less efficient than Windows type threading.

Yeah, but a credible "multi-tasking" operating system ought to be able
to do both without excessive messiness.  Anything otherwise is denying
the programmer access to the capabilities used by the operating system
itself (ie to create new application instances.)  (This is assuming you
define "fork" approximately as "create a new process that doesn't share
the address space of the current process" and "thread" as "create a
process that does share at least part of the current address space.)

BillW

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email spamBeGonelistservspamBeGonespammitvma.mit.edu with SET PICList DIGEST in the body

2003\06\29@101421 by Bob Ammerman

picon face
> >> fork() is far less efficient than Windows type threading.
>
> Yeah, but a credible "multi-tasking" operating system ought to be able
> to do both without excessive messiness.  Anything otherwise is denying
> the programmer access to the capabilities used by the operating system
> itself (ie to create new application instances.)  (This is assuming you
> define "fork" approximately as "create a new process that doesn't share
> the address space of the current process" and "thread" as "create a
> process that does share at least part of the current address space.)

Well, then UNIX was not a credible "multi-tasking" operating system until
pthreads came along. The POSIX standard for pthreads came out in 1995.

Bob Ammerman
RAm Systems

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email TakeThisOuTlistservEraseMEspamspam_OUTmitvma.mit.edu with SET PICList DIGEST in the body

2003\06\29@145329 by Peter L. Peres

picon face
> Forking is a poor man multithreading. It was developed when in-process
> light threads was rare in full fledge operating systems (as opposed to
> real-time kernels). Forking has higher overhead because of the
> replication of the entire data space of the project and requires more
> complex synchronization between the threads because of the address space
> separation.

I understand what you say but imho forking is a *rich* man's
multithreading. Because when you fork you really want to. With light
threads the thread-starting application cannot exit and leave the thread
running usually. You usually fork to start a server in the background, or
some other heavy item that needs to run by itself.  Nowadays a fork
implies copying less since code is shared by concurrently running copies
and the data area is usually small at the time of the fork. And
lightweight threads are also available if you want them.

0.02 bits,

Peter

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email RemoveMElistservspamTakeThisOuTmitvma.mit.edu with SET PICList DIGEST in the body

2003\06\29@145334 by Peter L. Peres

picon face
> fork() is far less efficient than Windows type threading.

fork is a system call under *nix and is handled by the kernel. It is about
as efficient as you can get probably because there isn't a single server
daemon that doesn't use fork(). If there is I haven't seen it. Code is
shared between forked processes and usually the data area is small at the
time the fork occurs. This means very little copying.

> Notice that unix based OS's have added threading packages, but I've not
> seen anything like Windows trying to do fork().

Yes. The trick is that after you do _beginthread() in W* you can't exit
the caller application. The program structure is entirely different. A
fork() gives you a verbatim copy of the running application with full
rights to become a process group leader and to fork itself later, even if
the parent has stopped meanwhile. This can go on forever (almost). The
protection between the forked child and the parent is very good. It cannot
overwrite data or crash the parent.

Peter

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email listservEraseMEspam.....mitvma.mit.edu with SET PICList DIGEST in the body

2003\06\29@162734 by Bob Ammerman

picon face
> fork is a system call under *nix and is handled by the kernel. It is about
> as efficient as you can get probably because there isn't a single server
> daemon that doesn't use fork(). If there is I haven't seen it. Code is
> shared between forked processes and usually the data area is small at the
> time the fork occurs. This means very little copying.

1: "as efficient as you can get" is still not as efficient as a simple Win32
CreateThread(). Since you are creating a new address space, you have to
build the appropriate structures for them.

2: Actually, even data typically uses copy-on-write on modern systems, so
little if any copying goes on as a result of the fork() call.

> > Notice that unix based OS's have added threading packages, but I've not
> > seen anything like Windows trying to do fork().

> Yes. The trick is that after you do _beginthread() in W* you can't exit
> the caller application. The program structure is entirely different. A
> fork() gives you a verbatim copy of the running application with full
> rights to become a process group leader and to fork itself later, even if
> the parent has stopped meanwhile.

Win32 CreateProcess cal do all that fork() does, except pass the address
space. It can pass open handles, etc.

>  The
> protection between the forked child and the parent is very good. It cannot
> overwrite data or crash the parent.

Also true in Win32

In summary: Win32 gives you both choices. Unix tries to make everything
happen with fork(). Note that Linux now supports the powerful clone() kernel
call which can share any desired part of the calling process' context
(memory, handles, etc.). The Linux pthreads library is based on this call
instead of a traditional in-process signal-based pthreads implementation.



Bob Ammerman
RAm Systems

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email EraseMElistservspammitvma.mit.edu with SET PICList DIGEST in the body

2003\06\29@173150 by William Chops Westfield

face picon face
> Yeah, but a credible "multi-tasking" operating system ought to be able
   > to do both without excessive messiness.

   Well, then UNIX was not a credible "multi-tasking" operating system until
   pthreads came along. The POSIX standard for pthreads came out in 1995.

Yup.  I probably should have put "modern" in my assertion somewhere.

BillW (an original member of the unix-hater's mailing list.)

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email RemoveMElistservEraseMEspamEraseMEmitvma.mit.edu with SET PICList DIGEST in the body

More... (looser matching)
- Last day of these posts
- In 2003 , 2004 only
- Today
- New search...