Linux-Advocacy Digest #85, Volume #31 Wed, 27 Dec 00 20:13:02 EST
Contents:
Re: linux does NOT suck (oh yes it does) ("John W. Stevens")
Re: Newbie: "Linux has come so far only to seem so far away" (Form@C)
Re: linux does NOT suck (oh yes it does) (Gary Hallock)
Re: Operating Systems? Where would you go next? (Harlan Grove)
Re: linux does NOT suck (oh yes it does) ("John W. Stevens")
Re: Conclusion ("Adam Ruth")
Re: Why Advocacy? (Gary Hallock)
Re: if linux is good, why is it so easy to freez it with netscape? (Arthur Frain)
Re: linux does NOT suck (oh yes it does) ("John W. Stevens")
Re: ATA RIPOFF! ALERT! (Generic PC CPRM copy control) (Tesla Coil)
Re: What's the sed line mean? ("John W. Stevens")
Re: An Entire Day With Linux (Yukkkkk!!!) (Peter Hayes)
Re: Since this is an Advocacy.... (Arthur Frain)
Re: Newbie: "Linux has come so far only to seem so far away" (pip)
----------------------------------------------------------------------------
From: "John W. Stevens" <[EMAIL PROTECTED]>
Crossposted-To: alt.linux.sux
Subject: Re: linux does NOT suck (oh yes it does)
Date: Wed, 27 Dec 2000 17:10:58 -0700
Kyle Jacobs wrote:
>
> I think your missing the point in this post.
>
> More often then not, users have THE EXACT SAME problem with their Linux
> experiences. Their hardware is just not supported 100%, features and
> functionality are MISSING or not implemented yet.
But this is also true of Windows.
1) We've had this discussion before: I could not successfully
install Win98 on a Linux box, because Windows did not support
my hardware.
2) Windows comes pre-installed, and pre-configured on hardware
chosen specifically to run Windows, while Linux is all to often
installed on Windows-hardware by a newbie.
Catch the apples to oranges comparison there . . .
3) Windows can support some hardware that Linux cannot, and will
never support, because the hardware specs, protocols, or interfaces
are not open to the public.
> It's a typical example of Linux playing "catch-up" with Windows
No, it's the usual "apples to oranges" comparsions that newbies make,
without even realizing how unreasonable that comparison is.
Buy a pre-built, pre-installed, pre-configured Linux box, use it
for a while, *THEN* tell us how much it sucks, and you might actually
be listened to seriously by the few remaining readers who haven't
already kill filed you.
> (on the
> desktop). Most people don't buy HARDWARE to accommodate their SOFTWARE .
Wrong. Most people *DO* buy hardware to accomodate their software . . .
that's why they buy pre-installed Windows boxen.
> It's the other way around, and Linux may claim "support" but more often then
> not, true functionality is either not yet available under Linux, or is
> available in "beta form" in either test-kernel modules, Beta releases of
> XFree86 or some other questionable piece of software not included "out of
> the box".
Hardware manufacturers will send pre-release specs, hardware and
information to commercial organizations like MS . . . they won't
release such information to the Linux developers, so your complaints
are equivalent to saying that, in a 100 yard dash race where one
runner is required to wait ten seconds after the starting gun
before being allowed to start, that that runner is the slowest of
all.
It's not that Linux is slower, it's that the race is rigged.
> Most people evaluating Linux aren't in the mood to either learn the minutia
> of information needed just to UPGRADE the components (let alone compile
> them).
They don't need to learn this . . . all they need to do is buy their
Linux box exactly the same way they buy a Windows box . . .
pre-installed
and pre-configured on hardware that was assembled to run their
software.
> But instead of creating a GUI based compiler which also tracks
> dependencies to compiled projects, Linux programmers just chalk the whole
> thing up to "learning curve", and do nothing about it.
It is superior to *TEACH* a man to fish, than to give him a fish.
> Clearly you can claim the "functionality exists, you just don't know how to
> do it." Fine, claim it. Do they (The community) do anything to resolve it
> (see aforementioned recommendation)? No.
Actually, the proper answer is "yes". The Linux Documentation
Project is "something done to resolve this problem".
--
If I spoke for HP --- there probably wouldn't BE an HP!
John Stevens
[EMAIL PROTECTED]
------------------------------
Subject: Re: Newbie: "Linux has come so far only to seem so far away"
From: [EMAIL PROTECTED] (Form@C)
Date: Thu, 28 Dec 2000 00:13:48 GMT
pip <[EMAIL PROTECTED]> wrote in
<[EMAIL PROTECTED]>:
<snip>
>> >Linux is _NOT_ DOS. DOS is just not comparable in *any* aspect!
>> >
>>
>> Sorry, it *is* comparable - just not favourably, that's all!
>
>Linux is a modern scalable 32-64 bit smp multitasking multi-user
>multi-platform posix like os.
>I am sorry you can't see the difference. In fact it is REALLY insulting
>to most of us Linux types that people use this comparison. DOS, the
>16-bit cp/m rip off is not in the same ball-park. DOS has no security
>and no memory protection. It is the relic of a past generation and
>serves
>no useful purpose other than for disk target practice on my darts board.
>
<mick dives for cover behind teletype>
WOAH! don't shoot me! The following bit of my reply was the important bit
and you agreed with that!
_On operating systems (much shrunk to fit in a reasonable post)_
DOS, CP/M, unix and Linux are/were all operating systems. All were
originally designed to be run from a command line. That is where the
similarity ends. I already know that. They are all from different periods
in the tangled history of operating systems. The first two share a common
heritage as do the second two. (Just to wind you up, IIRC DOS and unix are
also distantly related!). For goodness sake, don't feel insulted if Linux
is even slightly compared to "old fogey" OSs! I fully acknowledge the
superiority of Linux over DOS. I also realise that Linux is only possible
with the memory management systems built into microprocesors from the 386
onward. Before then the restricted address space made it completely
impossible (ok, somebody bring up bank switching now...). Only mainframes,
where unix was born, had big enough memories to *load* multiple processes,
never mind *run* them. Once the miroprocessor appeared CP/M came into being
(Control Program for Microprocessors). Much later there was an attempt to
get multi-user DOS accepted. It failed, and rightfully so. Early attempts
at multi-user operation relied on task-switching with varying degrees of
success to make it appear that several users were running at once. This
carried on for many years.
_On multi-tasking (once again, much shrunk...)_
Since the 386 came along we have had the capability of true multi-tasking
by allocating each task its own protected memory area with (what appears to
the program to be) its own processor. Each of these is a virtual machine.
They do not interfere with each other as they cannot see each other. This
is why I suspect that a GUI does not run on a virtual machine - if it did
it couldn't see the other processes. Therefore I suspect that the processor
is switched out of vm mode while the GUI program has access and back into
it at other times. OK, if I'm wrong I hope someone will correct me - in a
reasonable manner. It is only in vm mode that processes are truly
independent so as soon as a GUI is added the system must become less
stable. (sorry pip, no java!)
>
>> Exactly my point. If you want a server, run a server OS. If you want a
>> desktop, run a desktop OS. If you want to run games, run a games OS.
>> None of the current crop of OSs are fully capable of all tasks. IMHO a
>> "truly great" OS *would* be capable ao all these and more! It would
>> also be easy to configure, unconditionally stable and intuitive in
>> use. At least Linux is on the way there...
>
>Well this I can agree with!
>
--
Mick
Olde Nascom Computers - http://www.mixtel.co.uk
------------------------------
Date: Wed, 27 Dec 2000 19:14:16 -0500
From: Gary Hallock <[EMAIL PROTECTED]>
Subject: Re: linux does NOT suck (oh yes it does)
Ayende Rahien wrote:
>
>
> Please note that TUX is possibly the worst web server you can choose.
> It run in kernel space, if it goes down, you system goes down as well.
> It can't handle dynamic content.
>
This is not entirely true. TUX can run in kernel space, but it does not have
to. Running TUX is user space is possible and still provides performance
improvements over Apache for static pages. It can pass dynamic content to
Apache.
Gary
------------------------------
From: Harlan Grove <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux,comp.os.os2.advocacy,comp.os.os2.apps
Subject: Re: Operating Systems? Where would you go next?
Date: Thu, 28 Dec 2000 00:09:40 GMT
In article <[EMAIL PROTECTED]>,
p@kkkkkk <[EMAIL PROTECTED]> wrote:
>one thing I liked about NTFS, is that I can turn compression on when
>I format the disk, and I can get almost twice as much capacity
>for the same disk, as I can with say ext2.
...
>why does not linux file systems support this feature? it does
>not seem to have side effect on applications on windows.
Compression requires processor time when reading from or writing to
disk. On a single-user machine or any machine with relatively little
disk I/O, the additional virtual disk storage may be worth the
performance drag.
Linux is primarily a server OS, and in that capacity there's usually a
great deal of disk I/O. The performance drag in such circumstances may
not be worthwhile.
However, with Linux you do get the full OS source code, so it's
possible (I'm not saying practical) to rewrite the file system to
perform on-the-fly compression.
One last thing. Data recovery from crashed compressed drives would be
more difficult than from uncompressed drives.
Sent via Deja.com
http://www.deja.com/
------------------------------
From: "John W. Stevens" <[EMAIL PROTECTED]>
Crossposted-To: alt.linux.sux
Subject: Re: linux does NOT suck (oh yes it does)
Date: Wed, 27 Dec 2000 17:22:23 -0700
Chris Ahlstrom wrote:
>
> Kyle Jacobs wrote:
> >
> > 6.The "Ports" collection. Visit the name of your desired program, type
> > "make" and "make install", and ALL the program's DEPENDENCIES and LIBRARY
> > files are COMPILED, INSTALLED and CONFIGURED automatically. Try this in
> > Linux, it's a PROJECT not a task. This makes component upgrades easier.
>
> Actually, RPM's seem to handle this quite well. But they don't compile.
Actually, RPM's *DO* compile . . . if they are source RPM's, you can
download 'em, then rebuild 'em. Same is true for Debian packages, too.
> I thought OpenBSD was better for security.
It is.
--
If I spoke for HP --- there probably wouldn't BE an HP!
John Stevens
[EMAIL PROTECTED]
------------------------------
From: "Adam Ruth" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.advocacy,comp.os.ms-windows.nt.advocacy
Subject: Re: Conclusion
Date: Wed, 27 Dec 2000 17:26:01 -0700
> Which means you have the inability to prove that it's wrong, or right. I
> didnt' say it was impossible to prove, I said that nobody here seems to
have
> the ability to do so.
No, not so. Anyone can prove it's wrong, no one has. I'm asking you to
prove it's wrong and not continue with the mantra, "nobody knows absolutely,
so let's assume it's wrong". You claimed it was wrong, so you have the
burden of proof. No one can prove, absolutely, that's it's correct.
However, in all cases I am aware of, it is correct, so what reason should I
use to discount the numbers? The theoretical possibility that there may be
some sites that could prove to possibly be causing a statistically
significant error? I don't think so.
> > I'm familiar with quite a few sites
> > (that I or friends have worked on) that have all have shown to be
correct.
>
> Such as?
www.intercation.com
www.evolocity.com
www.webpowerpak.com
www.nvisioncorp.com
www.runreports.com
www.firstsecuritybank.com
www.ckinet.com
www.shytei.com
www.sirspa.com
These are just a few off the top of my head that I know I've checked at one
time or another in the past.
You said in an earlier post that you've worked on sites with complex
firewall schemes that could cause problems with Netcraft's numbers. Care to
enlighten us on those sites?
> The mere fact that things CAN interfere, and even netcraft admits this, is
> enough. That means the only way to know if a number is correct is to know
> the sites infrastructure. You can't presume it's correct until proven
> wrong.
No it's not. It's a matter of how much they interfere. All research has
some errors, the Heisenberg Uncertainty Principle is one reason why. 100%
of the sites I work with show correct uptimes (of the ones I checked, I
didn't test them all, admittedly). 0 sites have been demostrated to be
incorrect. It would seem that if there is a possibility of interference,
it's also probably quite small.
> I've already shown they're incorrect for systems with firewalls, something
> that netcraft itself agrees with. So, unless you can prove that a site
has
> no firewalling or other factor, you don't know if the number is true or
not.
No you didn't. You showed that on your firewall Netcraft get's no numbers.
That's not incorrect. Did or did not Netcraft say that they couldn't get
uptime numbers from your site? Was Netcraft able to get uptime numbers for
your site? It would appear that what they claim is what really happened.
That, as far as I know, is the deffinition of "correct". For the numbers to
be shown incorrect, Netcraft must first publish some numbers.
With all the sites you've claimed to have worked with, can't you show that
one of them is incorrect?
Adam Ruth
------------------------------
Date: Wed, 27 Dec 2000 19:35:00 -0500
From: Gary Hallock <[EMAIL PROTECTED]>
Subject: Re: Why Advocacy?
Pete Goodwin wrote:
>
>
> Or why not use a STARTUP script. OpenVMS uses a startup script and a
> shutdown script.
>
The problem with a startup script is that it must be modified by each program
being installed or removed. Using separate scripts for each program with
symbolic links makes it less likely that things will get screwed up as changes
are made.
> What you seem to be advocating is a set of three distinctly different ways
> of doing GUI on a desktop. It's a bit of mess because you can't drag and
> drop between those three styles, you can't expect standard dialogs to even
> _look_ the same (take a look at gv for instance) let alone give the user a
> chance to expect things in similar places.
>
As of KDE 2, you can now drag and drop between KDE and Gnome apps.
Gary
------------------------------
From: Arthur Frain <[EMAIL PROTECTED]>
Subject: Re: if linux is good, why is it so easy to freez it with netscape?
Date: Wed, 27 Dec 2000 16:39:37 -0800
mike@nowhere wrote:
Looks like another Steve psuedonym to me -
making a big deal out of a little annoyance.
> Why is it that Linux seems to be the only OS (at least from the
> ones I tried) that the Netscape browser can bring it to its knees?
> so easy to do. I remove my ethernet cable from the interface,
> then bring up Netscape. It hangs. X hangs. the mouse does
> not respond, can't access an xterm to type anything becuase I
> can't move the mouse.
I've found most word processors aren't usable
when I unplug the keyboard too. It's a helluva
thing :)
(This kinda reminds of the old joke:
patient: Doctor, it hurts when I do this.
doctor: Then don't do that.)
> Only way is to kill X
Actually it's not. But if you can kill X, Linux
is still running fine.
> Happens everytime.
I'd believe that. My guess is that it happens
because you're misconfigured somehow - like
trying to reach the LAN for DNS rather than
using the 'net. It's not an interesting enough
problem to be worth much thought.
> Never happens on Solaris or Windows.
Don't know if I believe that - NS wasn't
too good on Win95 either.
> (Netscape must be the most powerfull software in the world to be
> able to hang the computer becuase it can't access the network).
> And, after all these years, no one is able to fix this problem
> with netscape? the source is out there right? and I hear that
> hundereds of hackers are working on it, and yet, no one knows
> how to prevent Netscape from locking up the computer if it is
> able to connect to network?
Actually, I've been using Mozilla (both 0.6 and
the Milestone 17 release) lately and neither seem
to have similar problems (of course I'm so used to
Netscape's quirks I may be missing something)
> On Linux, this means, if I want to read an HTML file on my local
> disk, I must be connected to the network to do that, becuase if
> I try to launch netscape while not connected, too bad. Here goes
> my whole desktop session down.
Nope, doesn't mean that at all. You can use Mozilla
or Konqueror, for example, or even KFM. Nice to
have choices. I've always thought KFM was a lot easier
for reading local HTML files. (re: Konqueror - KDE 2.0.1
is perfectly stable; KDE 2.0.0 isn't IME)
> we send a man to the moon, but we can't figure how to stop a browser
> from locking X down. amazing.
I used to be more willing to defend Netscape (I don't
find it looks bad at all on my systems, and it works
for most everything I do) but it has been annoying
enough over the years that I don't feel much like
defending it anymore. But it's simply that - annoying.
First, your title is way off. Linux, in fact, *is* so
good that NS *doesn't* freeze it. The problem is actually
within X, or maybe even Motif. My guess is that NS
hijacks the event loop somehow. I haven't tried
disconnecting the ethernet cable (seems silly to me),
but trying to visit a non-existent URL will sometimes
do the same thing (I'm assuming you have the same
problem I have). The fact that you can CTRL-ALT-BKSPC
and kill X demonstrates that there is no problem with
Linux itself (Linux is still running fine).You can
also CTRL-ALT-F2 to a CLI a then "killall -9 netscape"
and then CTRL-ALT-F7 to return to X. Sometimes the
mouse will still take a few seconds to come back.
Lastly, you can telnet in (although not, I suppose,
with the ethernet cable disconnected) and kill NS.
Generally, though, I can usually get to an xterm
and kill NS, which is the easiest of all. Keep an
xterm open and try ALT-TAB to switch to it if the
mouse doesn't work (you don't have to use the
mouse, y'know).
Some people may find the fact that Linux continues
unaffected to be nit-picking, but in my case, I
usually only run NS when I have some other useful
process running as well, so the fact that the
problem is recoverable and does not represent a
problem with the OS itself is significant to me.
I've never had a NS problem that wasn't easily
(i.e - less than 30 seconds) recoverable, and
never suffered any data loss or had to reboot.
Lastly, my experience has been that what appears to
be a "hang" is in fact an *extremely* long timeout
(like a few minutes). I'd agree this *is* picking
nits, because I usually kill NS rather than waiting
for the timeout to expire. But it's worth the
wait if you have something else important going on.
Anyway, the problem doesn't occur that often with
the ethernet cable attached - maybe once a month.
It isn't a problem at all if you have the slightest
knowledge of Linux (which Steve - er - Mike doesn't
have, apparently).
I've had Win9x machines that wouldn't even boot
if they were set up for networking and the ethernet
cable was disconnected (or worse, intermittent).
Arthur
------------------------------
From: "John W. Stevens" <[EMAIL PROTECTED]>
Crossposted-To: alt.linux.sux
Subject: Re: linux does NOT suck (oh yes it does)
Date: Wed, 27 Dec 2000 17:39:34 -0700
Kelsey Bjarnason wrote:
>
> "Chris Ahlstrom" <[EMAIL PROTECTED]> wrote in message
> news:[EMAIL PROTECTED]...
> > Kelsey Bjarnason wrote:
> > >
> > You can use hdparm to get them detected. Windows and Linux install
> > the most conservative configuration for hard-drive parameters.
> > You can tweak them using hdparm, then make them permanent by editing
> > /etc/sysconfig/harddrives (under RedHat, at least).
>
> Sorry, I'm playing typical end-user here. What menu option is hdparm
> selected by?
Right Click on Root --> XTerm
type:
man hdparm
then construct the appropriate command.
> Not true at all; I learned that Linux was more work for less result.
Really? Do you really not see the simple fact that you set yourself
up to have these problems? That you, in effect, set yourself up to
fail?
> Sorry,
> some of us use the machines for fun, work and other things... and that's
> real work, doing real tasks, rather than catering to a half-functional OS.
I almost never spend time "catering" to my OS, and I do use it for
fun, work, and other things.
You set yourself up to fail, then blame Linux . . .
> > My Linux box caters to me.
>
> Good. Mine doesn't. So I turfed it.
No, that's not true. You chose your hardware so that it wouldn't
run Linux, installed Linux on it, then complained bitterly that Linux
wouldn't support your hardware.
Spot the self-delusion in this! ;-)
> I can only go by my experience;
Not True.
> I can't go by yours, or the guy next door's.
False. You *CAN* go by my experience, or the guy next doors . . .
Step 1: Buy Linux compatible hardware. When you're done with that,
come back and talk to me again.
I can give you a hardware list that, if you purchased this hardware,
you will be able to install a complete, working Linux system in under
a hour, and then just start using it.
> My experience says that Linux is a waste.
You made it be a waste.
My experience is that Windows 98 is a waste (see my posting about
re: Windows would not even INSTALL on my box).
> When I try it again in another 2
> years, that might be different;
No, it won't be. You'll still be able to set yourself up
to fail.
--
If I spoke for HP --- there probably wouldn't BE an HP!
John Stevens
[EMAIL PROTECTED]
------------------------------
From: Tesla Coil <[EMAIL PROTECTED]>
Subject: Re: ATA RIPOFF! ALERT! (Generic PC CPRM copy control)
Date: Wed, 27 Dec 2000 19:47:50 -0600
> Can anyone find a main stream news organisation that is
> covering this story? Everything feels eerily silent (perhaps
> its just the Christmas season).
Indeed, however...
"New technology could help squelch digital music piracy"
By John Borland, Staff Writer, CNET News.com
December 27, 2000, 3:05 p.m. PT
http://news.cnet.com/news/0-1005-200-4292282.html
------------------------------
From: "John W. Stevens" <[EMAIL PROTECTED]>
Subject: Re: What's the sed line mean?
Date: Wed, 27 Dec 2000 17:42:59 -0700
"G.P. Hwang" wrote:
>
> Yeah,thank you very much.
> Through your explanation and man pages I had myself understood. ;-)
> Ehh,can you tell me why you change sed/awk to perl? I have not any
> idea of perl for I doesn't ever use it.
NOOOO!!!!
Don't turn towards the dark side, Luke!
Use the Python! Use the Python!
;->
Python: Because life's to short to learn the same language six times.
--
If I spoke for HP --- there probably wouldn't BE an HP!
John Stevens
[EMAIL PROTECTED]
------------------------------
From: Peter Hayes <[EMAIL PROTECTED]>
Crossposted-To: alt.linux.sux
Subject: Re: An Entire Day With Linux (Yukkkkk!!!)
Date: Thu, 28 Dec 2000 00:54:47 +0000
Reply-To: [EMAIL PROTECTED]
On Wed, 27 Dec 2000 21:28:35 +0200, JM <[EMAIL PROTECTED]> wrote:
> On Sun, 24 Dec 2000 04:57:22 GMT, in comp.os.linux.advocacy,
> ("Kyle Jacobs" <[EMAIL PROTECTED]>) wrote:
>
> >If your dial up connection is so anemic that you can't afford to download an
> >extra 5kilobytes of data, you need to get a better job, or give up the
> >internet.
>
> If everyone does it then that's a lot of bytes.
And your ISP has to store all these superfluous bytes, adding to costs
passed onto you, their customer, as well as the rest of us. Or they'll
expire news quicker to save space. Either way it costs.
Peter
------------------------------
From: Arthur Frain <[EMAIL PROTECTED]>
Subject: Re: Since this is an Advocacy....
Date: Wed, 27 Dec 2000 17:01:47 -0800
John W. Stevens wrote:
> [EMAIL PROTECTED] wrote:
>> On Wed, 20 Dec 2000 22:27:09 -0500, "Colin R. Day"
>> <[EMAIL PROTECTED]> wrote:
>> I already know how it works. In fact I learned that way back in 1979
>> or so and it is old news. Now I concentrate on USING it instead of
>> having to figure out how it works every time I add a new piece of
>> hardware or software.
> Excuse, but are you trying to claim you can't get the same kind of
> problem under Windows, that you can get under Linux (having to tweak
> it to work with new hardware)?!
> If so, you're either extremely lucky, or extremely deluded.
>> Windows 2000 allows me that luxury and Linux doesn't.
> Wrong on both counts. There are indeed bits of hardware that Windows
> 2K will barf on, and there are many pieces of hardware you can simply
> stick into a Linux box, and have it work.
Sorry to bring this up in response to your post, John,
but I just bought my wife an HP 2250 Inkjet. It's a
really excellent printer, and it works with Linux
right out of the box (being Postscript and PCL 5
capable so no drivers required).
My wife still likes to use Windows for photos (it's
a little more work under Linux), so she booted Windows
and installed the printer software HP supplied. It
not only doesn't work - any time she clicks anything
related to the printer, the screen goes black and
she has to hit the reset button (even CTRL-ALT-DEL
won't work). I suggested maybe HP had updated drivers
available, but her response was "Don't bother - I'll
just use Linux".
Now I have the luxury of not having to get the printer
working under Windows.
Arthur
------------------------------
From: pip <[EMAIL PROTECTED]>
Subject: Re: Newbie: "Linux has come so far only to seem so far away"
Date: Thu, 28 Dec 2000 01:04:20 +0000
"Form@C" wrote:
> <mick dives for cover behind teletype>
and well you might :-)
> _On operating systems (much shrunk to fit in a reasonable post)_
> DOS, CP/M, unix and Linux are/were all operating systems. All were
> originally designed to be run from a command line. That is where the
> similarity ends. I already know that. They are all from different periods
> in the tangled history of operating systems. The first two share a common
> heritage as do the second two. (Just to wind you up, IIRC DOS and unix are
> also distantly related!).
Just out of my curiosity - how is this? Was the cp/m guy an ex Unix guy
or something of that sort?
>For goodness sake, don't feel insulted if Linux
> is even slightly compared to "old fogey" OSs! I fully acknowledge the
> superiority of Linux over DOS. I also realise that Linux is only possible
> with the memory management systems built into microprocesors from the 386
> onward. Before then the restricted address space made it completely
> impossible (ok, somebody bring up bank switching now...). Only mainframes,
> where unix was born, had big enough memories to *load* multiple processes,
> never mind *run* them.
Yes true. In fact Linux (I think) does not use all the all the options
available on modern processors (ix86 I think has 4-5 rings of processor
protection and I am sure Alpha and powerpc are similar).
>Once the miroprocessor appeared CP/M came into being
> (Control Program for Microprocessors).
I used cp/m on my Amstrad cpc6128 - now that was a while ago!
>Much later there was an attempt to
> get multi-user DOS accepted. It failed, and rightfully so. Early attempts
> at multi-user operation relied on task-switching with varying degrees of
> success to make it appear that several users were running at once. This
> carried on for many years.
> _On multi-tasking (once again, much shrunk...)_
> Since the 386 came along we have had the capability of true multi-tasking
> by allocating each task its own protected memory area with (what appears to
> the program to be) its own processor. Each of these is a virtual machine.
Ok, I see where you are coming from...
> They do not interfere with each other as they cannot see each other. This
> is why I suspect that a GUI does not run on a virtual machine - if it did
> it couldn't see the other processes. Therefore I suspect that the processor
> is switched out of vm mode while the GUI program has access and back into
> it at other times. OK, if I'm wrong I hope someone will correct me - in a
> reasonable manner. It is only in vm mode that processes are truly
> independent so as soon as a GUI is added the system must become less
> stable.
Process can communicate with each other via various means such as
signals
and pipes or sockets (as in Gnome which uses CORBA). I feel that it is
_very_ unlikely that x would run in kernel space as it is a user-space
program. I think that the only kernel space operations are through the
video drivers (which can cause the whole system to crash). I would
not bet a fortune on this, but I believe it to be correct. Otherwise,
as I stated before, if x were to crash, you would not be able to end
the process as it would be in kernel space - i.e. your system would
be cream cheese and you couldn't find the knife.
>(sorry pip, no java!)
and I love Java! :-p
------------------------------
** 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 by posting to comp.os.linux.advocacy.
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
******************************
