Linux-Advocacy Digest #997, Volume #27 Wed, 26 Jul 00 19:13:03 EDT
Contents:
Re: BASIC == Beginners language (Was: Just curious....
Re: Would a M$ Voluntary Split Save It? ([EMAIL PROTECTED])
Re: Would a M$ Voluntary Split Save It?
Re: Would a M$ Voluntary Split Save It? (Seán Ó Donnchadha)
Re: I had a reality check today :( ("John W. Stevens")
Re: I had a reality check today :( ("John W. Stevens")
Re: Would a M$ Voluntary Split Save It? (Seán Ó Donnchadha)
Re: Will Linux Dominate the Desktop Future? (T. Max Devlin)
Re: I had a reality check today :( ("John W. Stevens")
Re: Would a M$ Voluntary Split Save It? (Marty)
Re: I had a reality check today :( ("John W. Stevens")
Re: Tinman digest, volume 2451750 (Still more Tholen Lies and ([EMAIL PROTECTED])
Re: I had a reality check today :( ("John W. Stevens")
Re: I had a reality check today :( ("John W. Stevens")
----------------------------------------------------------------------------
From: <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: BASIC == Beginners language (Was: Just curious....
Date: Wed, 26 Jul 2000 14:46:49 -0700
Reply-To: <[EMAIL PROTECTED]>
John Sanders <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> [EMAIL PROTECTED] wrote:
>
> >
> > None of that is in question. The question was about WRITING an OS in
BASIC,
> > not using a BASIC interpreter written in another language AS the OS.
That
> > is the source of your confusion.
>
> Sure. You write your OS in BASIC, and never run it, you can always
> claim that you did a great job on it and BASIC is the _ideal_ tool for
> OSes!
Yea, reminds me of a certain shop. They used a computer since the mid 50's.
Every few years they replaced their computer with modern state of the art
model. They still wanted to run software for theold computer so they would
purchase an emulator for the new computer that would run the programs of the
previous computer. By the time that I encounterd the shop in the late 70's,
they were running their old programs through as many as five layers of
emulation programs. Talk about slow!
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To:
comp.os.os2.advocacy,comp.os.ms-windows.nt.advocacy,comp.sys.mac.advocacy
Subject: Re: Would a M$ Voluntary Split Save It?
Date: Wed, 26 Jul 2000 16:52:35 -0500
On 26 Jul 2000 13:42:36 -0500, [EMAIL PROTECTED] (Leslie Mikesell)
wrote:
>>Not at all true - MacOS in particular has consumers that like it and
>>ask for it; this forum is proof of that. (CSMA). CUSA has demand
>>for Macs; I suggest there is little or no demand for Linux, and that's
>>why they don't carry machines with it.
>
>That is a rather odd conclusion if you have actually been to
>one of the stores recently. The ones I've seen have a very
>large Linux software section which doesn't mesh well with a
>claim of 'no demand'. Perhaps they are still under one
>of those 'per-CPU' contracts with MS...
They sell Macs, and they sell their own "house" brand. Why not stick
Linux on a few and see how they sell?
------------------------------
From: [EMAIL PROTECTED] ()
Crossposted-To:
comp.os.ms-windows.nt.advocacy,comp.os.os2.advocacy,comp.sys.mac.advocacy
Subject: Re: Would a M$ Voluntary Split Save It?
Date: Wed, 26 Jul 2000 21:55:20 GMT
On Wed, 26 Jul 2000 16:17:53 -0400, Seán Ó Donnchadha <[EMAIL PROTECTED]> wrote:
>On Wed, 26 Jul 2000 20:12:57 GMT, [EMAIL PROTECTED] (Bob Hauck) wrote:
>
>>
>>And most Linux distros comes with at least three browsers (kfm,
>>netscape, lynx).
>>
>
>That makes perfect sense to me. So why can't Windows come with even
>one browser?
Except that isn't what's going on.
Several interchangable modules.
versus
one and only one module that just happens to make
itself inseparable from the OS.
--
Unless you've got the engineering process to match a DEC,
you won't produce a VMS.
You'll just end up with the likes of NT.
|||
/ | \
------------------------------
From: Seán Ó Donnchadha <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.ms-windows.nt.advocacy,comp.os.os2.advocacy,comp.sys.mac.advocacy
Subject: Re: Would a M$ Voluntary Split Save It?
Date: Wed, 26 Jul 2000 17:55:53 -0400
On Wed, 26 Jul 2000 21:06:34 GMT, [EMAIL PROTECTED] (Bob Hauck) wrote:
>>>
>>>And most Linux distros comes with at least three browsers (kfm,
>>>netscape, lynx).
>>
>>That makes perfect sense to me. So why can't Windows come with even
>>one browser?
>
>If it came with three I think MS would have had less trouble with the
>DoJ.
>
Excuse me? "Less trouble with the DoJ"?! Come on now; either Microsoft
broke a law or they didn't, right? Or are you saying that this law is
subject to creative interpretation by prosecutors as well as judges
who then get to call their take on it "the findings of fact"?
Also, why should Microsoft (or anyone else) be required to distribute
a third party's software with their own? I mean, why stop at three?
Why not force Microsoft to include *ALL* browsers - current and
future? And why stop at browsers? If browsers are bundled, isn't it
only fair that *ALL* applications are?
Look, I'm not asking for much. I just want to hear one sensible reason
why Windows shouldn't include a browser when every other desktop
platform does. Just give me something that makes *SENSE*! How could a
browserless Windows possibly hope to compete in the Internet age
against platforms that come with Web browsers?
------------------------------
From: "John W. Stevens" <[EMAIL PROTECTED]>
Subject: Re: I had a reality check today :(
Date: Wed, 26 Jul 2000 16:02:03 -0600
"Aaron R. Kulkis" wrote:
>
> Erik Funkenbusch wrote:
> > Ever heard of OS/2, written by Microsoft in the late 80's?
>
> IBM, you moron. Some of my classmates WORKED on OS/2 during
> their co-op semesters.
As far as I know, OS/2 was a joint (cooperative) development project, so
it is correct to say that both IBM and MS wrote OS/2.
> You asked "And you can do this without stopping the kernal process?"
>
> Since the Unix kernal is NOT a process, there is no way to "stop the
> kernal process" as such a process does not exist.
>
> Now...how do you copy the process table? Simple: From root, you
> signal each prosses with SIGSTOP, which puts ALL of them to sleep.
> Then, you copy the process table from with whatever debugger program
> you are using, and then re-awaken with SIGCONT (which means "continue
> if stopped").
Thanks, Aaron, I hadn't thought of that way of doing it . . . 'course,
you'd have to have your debugger running underneath a process that
accepts the STOP signal, but doesn't pass it on to the debugger, right?
Otherwise, your debugger would stop, too! :-)
> In other words, it's quite simple. The only difficulty is finding
> the memory location which holds the address of the process table.
And this requires:
1) The ability to run a dynamic memory allocator that allocates the
right kind of memory from within the kernel, OR to allocate that memory
by hand.
2) A kernel that accesses the process table through indirection.
3) Don't forget to turn off the timer interrupt while you are swapping
the process table.
Obviously, you'd have to run the debugger on a serial or character based
UI system.
--
If I spoke for HP --- there probably wouldn't BE an HP!
John Stevens
[EMAIL PROTECTED]
------------------------------
From: "John W. Stevens" <[EMAIL PROTECTED]>
Subject: Re: I had a reality check today :(
Date: Wed, 26 Jul 2000 15:53:45 -0600
Erik Funkenbusch wrote:
>
> "Aaron R. Kulkis" <[EMAIL PROTECTED]> wrote in message
> news:[EMAIL PROTECTED]...
> > > That's called Multitasking. Even on systems without the concept of a
> task
> > > (like earlier Unix systems).
> >
> > OK
> > Unix Microsoft
> > Multi-tasking 1970 1995
> > Multi-processing 1980 1997
>
> Wrong again. Ever heard of Xenix? Written by Microsoft in the early 80's?
Xenix == Unix for the purposes of this discussion. MS == Windows.
Which you knew, of course.
> Ever heard of OS/2, written by Microsoft in the late 80's?
OS/2 was co-written. MS did not write OS/2 all by it's self.
> Ever heard of Windows NT, released by MS in 1993 (which did both
> multi-processors and multitasking)?
But didn't do multi-processing.
> Please explain to me how you can copy the process table without stopping the
> scheduler.
"Stopping the scheduler" does not equal "rebooting the operating
system", since the scheduler can be restarted after the changes have
been made.
> If you don't stop the scheduler, then processes can begin or end
> between the time you start the copy and when you change the pointer,
Hence the need for running a high priority process that has kernel
memory access privileges . . .
> totally
> corrupting your system, not to mention that there is more to increasing the
> table size than just allocating new memory. You also need to modify
> constants that define the size in numerous places in the kernel image
Wrong. Typically, the size of the process table is not stored in
"numerous places in the kernel image", it is stored in a single
variable. Change that variable, and you are done.
Remember, when running in USER SPACE, the KERNEL is not running!
> (something that you can't do without the debugger stopping the process).
"Stopping" does not equal "rebooting". And you do not need to use a
debugger to do this.
> > write to [EMAIL PROTECTED] for details.
>
> I don't have to. I know it can't be done without stopping the machine.
"Stopping the machine" happens 100 times a second on Unix.
Note: there certainly are things that you can only modify by rebooting,
but most of the kinds of changes that NT requires rebooting after, do
not require rebooting on Unix systems. And that was the original
contention, Erik.
> > Considering that for the undergraduate students in the School of
> > Electrical Engineering, this machine was their PRIMARY account, and
> > that many had no other accounts, AND that there were 100 students
> > logged in and working at the time....
> >
> > ...YES, this *IS* something you can do on a production machine.
>
> Forgive me if I don't take you word for it. Provide some proof, other than
> telling me to talk to someone whoe very well could be you.
I'll tell you it can be done, then.
Before the advent of complex scheduler technology, some pretty amazing
things could be done using
kernel memory access and a "high priority" process.
> > Even thoug it was a "beta test" for Gould, Purdue Electrical
> > Engineering considered it to be a PRODUCTION MACHINE.
>
> Uh huh.
So, is that your fall back position? When you are proven wrong, you
simply refuse to admit it, by implying that Aaron is lying?
There goes your credibility . . .
> > in other words, you win ONE POINT out of what...35?
>
> Forgetting already that you said Windows didn't have full multi-user
> capabilities
Considering what Aaron meant by "full multi-user capabilities", he was
correct.
> > > Hmm.. GCC runs on it. I guess GCC is an abomination then, so does gdb.
> Oh,
> > > and Unix can run on NT as well (since it's a pseudo micro-kernel).
>
> Again, no answer.
Neither GCC nor GDB run on NT.
They run on a POSIX compatibility layer . . . in essence, they don't run
on NT.
> Sounds like an admission to me.
And I repeat: discussion by definition is a poor substitute for a real
discussion.
X + WM + GUI Apps == a GUI.
> > You're the one claiming that DirectX is superior. Make your case.
> > Burden of proof is upon you, not me.
>
> I made no such claim. I simply said they were doing it first, not that it
> was superior to anything.
By definition, of *COURSE* they were doing it first! DirectX is a MS
product. However, the engineering pattern that DirectX is but a single
implementation of, was implemented in Unix X servers before Windows
added DirectX.
--
If I spoke for HP --- there probably wouldn't BE an HP!
John Stevens
[EMAIL PROTECTED]
------------------------------
From: Seán Ó Donnchadha <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.ms-windows.nt.advocacy,comp.os.os2.advocacy,comp.sys.mac.advocacy
Subject: Re: Would a M$ Voluntary Split Save It?
Date: Wed, 26 Jul 2000 18:06:14 -0400
On Wed, 26 Jul 2000 21:55:20 GMT, [EMAIL PROTECTED] () wrote:
>>>
>>>
>>>And most Linux distros comes with at least three browsers (kfm,
>>>netscape, lynx).
>>
>>That makes perfect sense to me. So why can't Windows come with even
>>one browser?
>
> Except that isn't what's going on.
>
Of course it is. The DoJ and Judge Jackson have been trying to get IE
out of Windows for years now.
>
> Several interchangable modules.
>
> versus
>
> one and only one module that just happens to make
> itself inseparable from the OS.
>
Microsoft doesn't force you to use IE as your primary Web browser. You
can use it to download Navigator and never use again, if that's what
you want. So what's the problem?
------------------------------
From: T. Max Devlin <[EMAIL PROTECTED]>
Crossposted-To: alt.destroy.microsoft
Subject: Re: Will Linux Dominate the Desktop Future?
Date: Wed, 26 Jul 2000 18:11:05 -0400
Reply-To: [EMAIL PROTECTED]
Said <[EMAIL PROTECTED]> in alt.destroy.microsoft;
>In alt.destroy.microsoft Jacques Guy <[EMAIL PROTECTED]> wrote:
>:
>: "I this WERE to happen". "Were" is the subjunctive, expressing a
>: contrary-to-fact
>: condition. Understood is: "but it will not happen". An example in
>: plainer English
>: "If I were a rich man" implying: "I am not".
>
>This is a ridiculous play with words. You take 'were' here to imply
>'that is not currently the case'. Then you make the ridiculously HUGE
>inference from that that it will NOT BE THE CASE. Max Devlin is talking
>about the future, you are talking about the present/past. In fact your
>first interpretation of 'were' is wrong. 'If this WERE to happen' is
>merely speculating about the future, and says NOTHING about the present
>situation.
Thank you. But he was referring to the original text by M. Gartenberg,
not my re-write, in an attempt to show that the considerations of that
"white-paper" were valid. The original text was published in September
of 1999, so some consternation over past/present tense is expected,
either way. Needless to say, his contentions are trivial.
>: Quoting again:
>:
>: Despite the press hype, we believe that Linux deployments for desktops
>: will not usurp OS dominance from Microsoft.
>:
>: "USURP". I'll just copy the definition out of the Collins Cobuild,
>: about the best
>: dictionary of modern English there is: "If you usurp a job, role,
>: title, or
>: position, you take it from someone, especially when you have no right
>: to
>: do this."
>
>Do you mean here that Microsoft has NO right to OS dominance?
No, he means that MS *has* the "right" to dominate, and that Linux would
be "usurping" Windows' 'proper place' as most popular product.
>The 'right' is probably meant to mean: 'given Microsoft's financial
>size and historical dominance' as supposed to 'given Microsoft's fine
>workmanship'.
Again, that text was by the FUD writers, not me. In my parody, I simply
deleted the word "not". The FUD authors did intend, as Jacques
indicates, that Linux being adopted on desktops would "usurp" OS
dominance from Microsoft. Being a good piece of FUD to begin with, the
text does not indicate whether it is fine workmanship or
anti-competitive business practices which placed MS in that position or
maintains it.
--
T. Max Devlin
-- Such is my recollection of my reconstruction
of events at the time, as I recall. Consider it.
Research assistance gladly accepted. --
====== Posted via Newsfeeds.Com, Uncensored Usenet News ======
http://www.newsfeeds.com - The #1 Newsgroup Service in the World!
======= Over 80,000 Newsgroups = 16 Different Servers! ======
------------------------------
From: "John W. Stevens" <[EMAIL PROTECTED]>
Subject: Re: I had a reality check today :(
Date: Wed, 26 Jul 2000 16:12:25 -0600
Erik Funkenbusch wrote:
>
> "Aaron R. Kulkis" <[EMAIL PROTECTED]> wrote in message
> news:[EMAIL PROTECTED]...
> > You asked "And you can do this without stopping the kernal process?"
> >
> > Since the Unix kernal is NOT a process, there is no way to "stop the
> > kernal process" as such a process does not exist.
>
> In effect, it does.
And in spite of the spin, irrelevant. The original contention was re:
rebooting, not temporarily freezing the kernel.
> > Now...how do you copy the process table? Simple: From root, you
> > signal each prosses with SIGSTOP, which puts ALL of them to sleep.
> > Then, you copy the process table from with whatever debugger program
> > you are using, and then re-awaken with SIGCONT (which means "continue
> > if stopped").
>
> Funny how it doesn't put the debugger to sleep, now does it?
If you are't careful to properly wrap your debugger . . . it *WILL*
freeze your debugger! But there is a way around that problem.
> And again, you can't just go stopping processes on a production machines.
Yes you can. Most production machines don't even need *soft* real time,
to say nothing of about hard real time.
> Many machines depend on continued access in order to function properly.
And that access will not be unduly affected by the split seconds it
takes to do this.
> For
> example, machine control software. Oh, but not living in the real world you
> wouldn't understand about production needs.
Ah . . . more "discussion by definition" . . . it's not "production",
unless it is "real time"?
That's just pitiful, Erik.
> Oooh.. 2 minutes of a stopped machine.
Actually, it can be done in much less time. The scheduler only needs to
be turned off long enough
to swap the table and modify the process table size variable . . . about
one missed timer interrupt, if that much, for an automated system.
A few seconds, at most, by hand.
> Why not just reboot?
Because while the scheduler is temporarily stopped, the machine can
still receive and process other interrupts.
> Don't you
> think the users are going to notice that their systems aren't responding for
> 2 minutes and start issuing breaks, or simply turinging off their terminals
> and trying to relog in?
If the operator turns off the scheduler before it is really neccessary,
the users will notice, I guarantee it, but that isn't strictly
neccessary.
> > Not a problem, see above.
>
> This wasn't what you were saying before, but even so. It's still not
> useable on a production machine.
Yes it is.
In fact, some HA machines have a built-in support mechanism now to do
just that. Newer kernels, of course, don't have a static process table.
> I suppose commands like PS and top don't need to know how BIG the table is.
> If you increase it's size, you have to change the constants which tell
> everything how big it is... Idiot.
PS and TOP don't require recompiling just 'cause you modify the size of
a process table.
--
If I spoke for HP --- there probably wouldn't BE an HP!
John Stevens
[EMAIL PROTECTED]
------------------------------
From: Marty <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.ms-windows.nt.advocacy,comp.os.os2.advocacy,comp.sys.mac.advocacy
Subject: Re: Would a M$ Voluntary Split Save It?
Date: Wed, 26 Jul 2000 22:18:27 GMT
Chris Wenham wrote:
>
> Actually, if I sprayed Deja with magic Invective-GoWay 99.9% of your
> posts would vanish!
Interestingly enough that seems to have happened on my news server at work. I
am not employing a killfile, yet the internal IBM news server that I use
during the day did not have a single one of Ed's posts. He's the only person
that I've noticed seems to be getting filtered by them for some reason (aside
from the spam postings which they also do an excellent job of filtering).
------------------------------
From: "John W. Stevens" <[EMAIL PROTECTED]>
Subject: Re: I had a reality check today :(
Date: Wed, 26 Jul 2000 16:13:57 -0600
"Anthony D. Tribelli" wrote:
>
> Aaron R. Kulkis <[EMAIL PROTECTED]> wrote:
>
> > DR-DOS implemented true Multi-processing in 1985, yet Microsoft
> > didn't accomplish the same task for another 8 years.
>
> Bad guess. Microsoft offered a Unix implementation called XENIX, IIRC -
> and later sold to SCO, and also OS/2.
MS didn't accomplish this in Windows for another 8 years.
The context was: Unix vs. MS (Windows), not Unix vs. MS. Aaron simply
used some short hand, because for the purposes of this discussion, Xenix
== Unix.
--
If I spoke for HP --- there probably wouldn't BE an HP!
John Stevens
[EMAIL PROTECTED]
------------------------------
Crossposted-To:
comp.sys.mac.advocacy,comp.os.os2.advocacy,comp.os.ms-windows.nt.advocacy
From: [EMAIL PROTECTED]
Subject: Re: Tinman digest, volume 2451750 (Still more Tholen Lies and
Reply-To: [EMAIL PROTECTED]
Date: Wed, 26 Jul 2000 22:30:17 GMT
Jim "little boy" Stuyck writes:
> [EMAIL PROTECTED] wrote:
Typical invective. Still engaging in the lowest level of discussion,
I see. No surprise there.
>> Jim "little boy" Stuyck writes:
>>> [EMAIL PROTECTED] wrote:
>> Typical invective. Still engaging in the lowest level of discussion,
>> I see. No surprise there.
> If the shoe fits...
The shoe is on your foot, Stuyck (little boy).
> It was YOU that brought up the discussion of this "Mai Fong" person,
It was YOU that investigated the "cast of characters" at Chauvet,
Stuyck (little boy).
> the "lowest level of discussion" in your own words.
On the contrary, those are your words, Stuyck (little boy).
JS] Date: 1996/07/08
JS] Message-ID: <4rr1nl$t0h1@IRIS>
JS]
JS] There is a hierarchy of subjects for conversation, discussion and debate.
JS] It goes something like this:
JS]
JS] o highest level -- ideas
JS] o middle level -- things
JS] o lowest level -- people
>>>> Ask Stuyck about his investigation involving Mai Fong, for example.
>>> What "investigation involving Mai Fong" do you refer to?
>> Yours, Stuyck (little boy).
> Prove it, if you think you can.
Here are just a couple of recent references:
Message-ID: <[EMAIL PROTECTED]>
Message-ID: <[EMAIL PROTECTED]>
> Otherwise, still another argument by evasion.
Your evasion of the truth, Stuyck (little boy).
>>> This is the first such allegation I've ever seen.
>> Read some of your own postings, Stuyck (little boy).
> Show us this specific "investigation of Mai Fong." Otherwise, still another
> example of your argument by evasion.
Your evasion of the truth, Stuyck (little boy).
>>> You didn't write "alleged investigation," so your blanket assertion
>>> that there was an "investigation" is written as if it were fact.
>> It is a fact, Stuyck (little boy). Or were you lying about all your
>> detective work in your previous postings on the subject?
> Show us this specific "investigation of Mai Fong." Otherwise, still
> another example of your argument by evasion. No surprise there.
Your evasion of the truth, Stuyck (little boy).
>>> There was no such "investigation involving Mai Fong,"
>> Balderdash, Stuyck (little boy). Have you also forgotten about
>> Doug Lantini?
> What's "Doug Lantini" got to do with "Mai Fong?"
According to you, they're the same person, Stuyck (little boy).
> Are you asserting that they are one in the same?
You're the one who did the investigation, Stuyck (little boy). I have
not taken any position on whether they are "one in the same".
> Are you asserting that a "Doug Lantini" really exists?
Having more reading comprehension problems, Stuyck (little boy)?
>>> so I'm curious as to why you post yet another "lie and
>>> misrepresentation."
>> How ironic, coming from the person lying and misrepresenting things.
>> You must be getting senile, Stuyck (little boy), to have forgotten
>> about your investigation of the alleged "cast of characters" at
>> Chauvet.
> Not at all.
Then try to explain your responses, Stuyck (little boy).
> You DID find my results -- this "'cast of characters' at Chauvet"
> -- accurate, didn't you?
Irrelevant, Stuyck (little boy). I merely noted the existence of your
investigation. I said nothing about the accuracy or inaccuracy of
that investigation. You do have a history of doing lousy research,
however; recall your failure to find the Pluto-Charon book at the
Space Science Series web site.
> I knew you did.
Liar.
> How about "FREESPEECH?" How about "webrider?"
How about them, Stuyck (little boy)? You claim that they are the same
person, thus you must admit the possibility that Glatt's collection
of comments does not represent as many unique individuals as comments.
> Why was it that, at the time, you -- you who claims to be all for
> fairness, blah, blah -- didn't tell these "characters" that their
> presence was not welcome?
On what basis do you speak for what I might find welcome or unwelcome,
Stuyck (little boy)?
> Tell us, wedgie-boy,
Typical invective, as expected from someone who lacks a logical argument,
preferring instead to continue engaging in the lowest level of discussion.
> what did you think of the Tim Martin appeal for more members of Warp
> City so that he could purchase a surfboard?
I saw no such appeal, Stuyck (little boy). Recall that I've told you
on multiple occasions that I do not read everything in this newsgroup.
> What, with all that Chauvet largess he talks about, you'd think he'd
> simply ask one of the "Chauvet's" for a gift, instead of all that
> champagne, wouldn't you?
What I would think is irrelevant, Stuyck (little boy). Meanwhile, I
see that you're trying to divert attention away from the subject at
hand, namely your investigation, one that you seem to be backing away
from now. Why is that?
------------------------------
From: "John W. Stevens" <[EMAIL PROTECTED]>
Subject: Re: I had a reality check today :(
Date: Wed, 26 Jul 2000 16:23:56 -0600
Erik Funkenbusch wrote:
>
> "Steve Mading" <[EMAIL PROTECTED]> wrote in message
> news:8ld8vs$cji$[EMAIL PROTECTED]...
> > You seem to be operating under the false impression that sockets on
> > localhost are just as slow as sockets on the network, and therefore
> > constitute a performance hit, which is a bad thing. They don't.
> > Localhost sockets are basically nothing more than moving bytes around
> > in memory. So, yes, it's true the Unix didn't corrupt the kernel
> > by infecting it with GDI code (which, since it is device-driver
> > dependant, can be glitchy and crashy). But this is not a bad thing.
>
> No, I understand that local sockets are much faster than using a network.
> They di carry more overhead than your basic API call, which is a matter of
> stuffing a few values on a stack and invoking an interrupt in most modern
> OS's.
Local sockets didn't have *much* more overhead, and some of that
overhead was reduced through shared memory segments.
> In a modern Unix, yes. Go back to, say 1993 when NT was released, or 1989
> when NT was started and you'll notice that kernel recompiles were much more
> common for things like adding device drivers in most Unices.
Wrong. Kernel's weren't recompiled to add device drivers . . . they
were basically relinked.
But after relinking, you did have to reboot.
Since Unix is more modular, however, the addition of new devices and new
device drivers is a relatively rare thing, compared to "configuration
changes".
> Well, Unix doesn't have full remote administration anyways. There are often
> tasks which need to be carried out on the console, usually in something like
> single-user mode.
The only thing I can think of that would require "single user mode" is
performing a by-hand repair of a root file system. . . in which case, if
that is your definition of "full", then no OS will ever have "full
remote administration capability", because you will *always" have to
have the administrator within arms length of the system to fix hardware
problems.
By any realistic definition of "full remote administration", Unix has
it, and even now, has more of it than NT does (NT needs more up and
functioning systems to support remote administration).
> Also consider a kernel recompile which causes the kernel
> to panic or simply hangs upon reboot.
Only developers do kernel recompiles, and the ability to relink a kernel
in such a way as to cause it to panic does not stop you from
administering it remotely . . . as the new kernel is installed in a
try-and-revert fashion . . . after the panic or HPMC completes, the
machine boots off of the original kernel.
'Course, you cannot do that with PC's, as they lack the intelligent
firmware to support this capability.
> You have to go to the console for
> that.
No, you don't need to be in front of the console to panic or hang a
machine.
> I liken X to the windows GDI, or the OS/2 GPI. That's not quite correct
> though, since X also includes some things from the windows USER. No, X
> doesn't include your wigits, and it doesn't include your window managers.
> But it does include the basic GUI functionality.
Now, what do you define as "basic GUI functionality"?
--
If I spoke for HP --- there probably wouldn't BE an HP!
John Stevens
[EMAIL PROTECTED]
------------------------------
From: "John W. Stevens" <[EMAIL PROTECTED]>
Subject: Re: I had a reality check today :(
Date: Wed, 26 Jul 2000 16:25:51 -0600
Erik Funkenbusch wrote:
>
> Serial ports don't have more than a few hundred feet of distance on them at
> best.
So . . . my telephone really *DOESN'T* connect me to people that are on
the other side of the planet?
--
If I spoke for HP --- there probably wouldn't BE an HP!
John Stevens
[EMAIL PROTECTED]
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list (and comp.os.linux.advocacy) via:
Internet: [EMAIL PROTECTED]
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Advocacy Digest
******************************
