Linux-Advocacy Digest #503, Volume #31 Tue, 16 Jan 01 06:13:03 EST
Contents:
Re: OS-X GUI on Linux? (Donn Miller)
Re: Red hat becoming illegal? ("David Brown")
Re: OS-X GUI on Linux? (Donn Miller)
Re: OS-X GUI on Linux? ("Tom Wilson")
Re: The Server Saga (Pete Goodwin)
Re: The Server Saga (Pete Goodwin)
Re: Why Hatred? (Nick Condon)
Re: Definition: Desktop, Workstation, Server. ("David Adams")
Re: Linux 2.4 Major Advance (Shane Phelps)
Re: Definition: Desktop, Workstation, Server. ("Tom Wilson")
Re: Linux 2.4 Major Advance ([EMAIL PROTECTED])
Re: Linux 2.4 Major Advance ("Tom Wilson")
Re: Kernel space? Who gives a @#$% (Shane Phelps)
----------------------------------------------------------------------------
From: Donn Miller <[EMAIL PROTECTED]>
Subject: Re: OS-X GUI on Linux?
Date: 16 Jan 2001 03:17:41 -0600
Tom Wilson <[EMAIL PROTECTED]> wrote:
> There should ALWAYS be the choice. I'm
> advocating a smaller, faster, Micky-Mouse Windowish GUI to placate those
> folks out there who bitch about such things. You'll notice that the desktop
> area is the only area that Linux isn't soundly trouncing Windows.
Such a GUI would also allow a consumer-oriented desktop OS to hit it big in
the consumer market, and further take a bite out of Windows sales. For
optimum performance, it'd probably be better to make the entire Windowing+GUI
system a monolith. This would enable you to build a Windows or MacOS -like
system, but with Linux running underneath. Plus, you could do other neat
things, like have the windowing system boot up before the kernel probes
occur. You'd see all the devices being probed in a special information
dialog, for example. Obviously, traditional Linux types would not like this.
But, it would allow you to use Linux, already a great OS, to construct an
entirely new desktop OS. It would not necessarily be Linux per se, but using
an existing Linux kernel to build a whole new OS around. If this idea hits
it big, you could always have some special compatibility layers that allow
you to run this new OS's apps under a traditional Linux with X running.
====== 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: "David Brown" <[EMAIL PROTECTED]>
Crossposted-To:
alt.destroy.microsoft,comp.os.ms-windows.advocacy,comp.os.ms-windows.nt.advocacy
Subject: Re: Red hat becoming illegal?
Date: Tue, 16 Jan 2001 10:23:31 +0100
Chad Myers wrote in message ...
>>
>> I've not heard about a TPC or TPD benchmark yet. I'm not sure
>> Linux will do well there, unless it runs on heavy hardware such
>> as an S/390, or perhaps a big multiCPU Sparc.
>
>I don't really consider Mindcraft or ZDNet major industry benchmarks,
>necessarily. While relevant, TPC and similar industry benchmarks
>are more reliable and standards based.
>
>c't is just FUD all around no matter what they're comparing.
>
The problem with benchmarks does not lie with Linux - it lies with the
commercial software suppliers. There is no point in doing comparisons on
database servers, or web servers, or whatever, as long as the commercial
suppliers (understandably, I might add - they have a reputation to maintain)
have such tight restrictions on what benchmark results can be published. No
one, not even ZD or Mindcraft or TPC, can actually run independant
benchmarks, so everyone is left guessing at benchmarks referring to "unnamed
commercial database 1" and the like.
------------------------------
From: Donn Miller <[EMAIL PROTECTED]>
Subject: Re: OS-X GUI on Linux?
Date: 16 Jan 2001 03:35:47 -0600
Charlie Ebert <[EMAIL PROTECTED]> wrote:
> What? It's inane and stupid to say if they offer OSX for Linux I want to
> try it?
> What's inane about that!
Because! Don't you know that if you run Linux, you've secretly signed a
special agreement never to try `evil' graphics systems like svgalib and GGI?
Anything you do *absolutely* has to be layered on top of X11! The 11th
commandment says "Thou shalt not run graphics in any shape or form on thy
Linux box other than X11!"
What the heck is wrong with trying alternative windowing/graphics systems on
Linux other than X11? Oh, right, running svgalib will create this enormous
security hole that will enable every world terrorist to log into your machine
simultaneously! Then, they will destroy the entire solar system, and you
will be sorry that you dared think of running any other graphics system other
than X11 on your Linux box.
So, I suppose the moral of this story is this: do NOT attempt to develop an
alternative graphics system to X11. Doing so will create a security hole the
size of some unknown galaxy (ironically, called "x11") that will cause every
terrorist known to man to log into your Linux box, and destroy the universe
as we know it. Then, X11, as well as all the matter in the known universe,
will be history! Then, you will be sorry that you committed the cardinal sin
of starting to think about running graphics systems other than X11 on your
Linux box.
Also, it's a little known fact, but the reason we Christians (assuming you are
one) are baptized at birth was to rid us of the extremely horrible sin of
plotting to start to think about running a graphics system other than X11 on
our Linux boxes. In fact, there will be extensive research in the medical
field in the near future of this magical pill that, when swallowed, will rid
mankind of evil thoughts of running any other graphics system other than X11
on your Linux box.
In fact, a lot of people probably panicked when W was being developed for
unix systems. "What???!! They want to put a windowing system on top of a
tty-driven CLI operating system? Will the universe end?" Of course, back
then, people probably abhorred the thought of ever running anything other
than curses on their UNIX boxes.
====== 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: "Tom Wilson" <[EMAIL PROTECTED]>
Subject: Re: OS-X GUI on Linux?
Date: Tue, 16 Jan 2001 09:45:53 GMT
"Donn Miller" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> Tom Wilson <[EMAIL PROTECTED]> wrote:
>
> > There should ALWAYS be the choice. I'm
> > advocating a smaller, faster, Micky-Mouse Windowish GUI to placate
those
> > folks out there who bitch about such things. You'll notice that the
desktop
> > area is the only area that Linux isn't soundly trouncing Windows.
>
> Such a GUI would also allow a consumer-oriented desktop OS to hit it big
in
> the consumer market, and further take a bite out of Windows sales. For
> optimum performance, it'd probably be better to make the entire
Windowing+GUI
> system a monolith. This would enable you to build a Windows or
MacOS -like
> system, but with Linux running underneath. Plus, you could do other neat
> things, like have the windowing system boot up before the kernel probes
> occur. You'd see all the devices being probed in a special information
> dialog, for example. Obviously, traditional Linux types would not like
this.
And I, on some levels wouldn't either hence my advocating keeping the X
model available.
What a lot of people can't seem to grasp is that an OS's success, outside
of scientific, IT, and academic circles, depends on how non-tech types
relate to it. That's Linux's Achilles Heel. Sure, some newbies get along
splendidly with Linux as it is now. They're exceptions to the rule, though.
Most people, and I hate to sound this way, are sheep who don't like to
blaze new trails. They're used to Windows and don't want anything else.
Even if they decide to try Linux, they'll carry this prejudice with them.
Linux's success on the desktop depends on winning this hard to please bunch
over. To do it requires a fast, game-friendly, and Windows-style desktop
with plenty of cohesive help wizards and simple configuration tools. You
add these things to Linux's proven speed and stability and you have a hands
down winner.
> But, it would allow you to use Linux, already a great OS, to construct an
> entirely new desktop OS.
I'm not sure I'd go quite that far... All we're talking about is an
interface API "makeover" not reconstructive surgery! :)
> It would not necessarily be Linux per se, but using
> an existing Linux kernel to build a whole new OS around. If this idea
hits
> it big, you could always have some special compatibility layers that
allow
> you to run this new OS's apps under a traditional Linux with X running.
An intriguing thought, but, It'd fly about as well as pigs and bricks do.
;)
The solid OS core is there already. We just need to dress it up nice and
pretty. <g>
--
Tom Wilson
Sunbelt Software Solutions
------------------------------
From: Pete Goodwin <[EMAIL PROTECTED]>
Subject: Re: The Server Saga
Date: Tue, 16 Jan 2001 09:30:20 GMT
In article <[EMAIL PROTECTED]>,
J Sloan <[EMAIL PROTECTED]> wrote:
> What planet are you living on Pete?
I was being silly.
> With Linux, things may not be immediately and intuitively
> obvious to a raw newbie, but there is a system and a way
> to learn where things are, and one you do it all makes sense.
Hmmm...
> Windows OTOH is black magic, and if it doesn't want to
> work, there's nothing you can do. I've got a ton of windows
> horror stories just from trying to set up a simple windows pc
> for my wife.
>From my perspective, this is how Linux looks.
--
---
Pete
Sent via Deja.com
http://www.deja.com/
------------------------------
From: Pete Goodwin <[EMAIL PROTECTED]>
Subject: Re: The Server Saga
Date: Tue, 16 Jan 2001 09:27:30 GMT
In article <[EMAIL PROTECTED]>,
"Aaron R. Kulkis" <[EMAIL PROTECTED]> wrote:
> Translation:
>
> "I, Pete Goodwin, am a complete idiot."
You mean to say you posted that MASSIVE response, including my original
post, and you say I'm an idiot? HAH!
--
---
Pete
Sent via Deja.com
http://www.deja.com/
------------------------------
From: Nick Condon <[EMAIL PROTECTED]>
Subject: Re: Why Hatred?
Date: Tue, 16 Jan 2001 09:59:46 +0000
Erik Funkenbusch wrote:
> Fine, I'll reword it just for you. Linux cannot replace Windows as the
> major OS of choice today.
(I won't even mention "choice"). Windows the major OS on the *desktop*. It has a
big a presence in the small-server space, whether or not it's the major OS there
is doubtful. It's made less progress in corporate data-centres, and makes a
pretty pathetic showing for Internet servers and academia. Here's a summary:
Sector: Is Windows The Major OS?
Desktop: Absolutely, and difficult to budge.
Workgroup Servers: Possibly, but much less so than the desktop, and they are
easily swapped for alternatives.
DC: It's there, but trailing.
Internet: It's there, but trailing.
Now there are lots more desktops than any other kind of computer, but they are
small and don't do much. If you compare something like workload, Windows pales
into the background. The OS that does most of the worlds data processing is
probably Solaris.
------------------------------
From: "David Adams" <[EMAIL PROTECTED]>
Subject: Re: Definition: Desktop, Workstation, Server.
Date: Tue, 16 Jan 2001 10:34:33 +0100
"mlw" <[EMAIL PROTECTED]> escribió en el mensaje
news:[EMAIL PROTECTED]...
> I don't know about you, but seeing "desktop" used interchangeably with
> "Workstation" seems a bit extreme. Perhaps I am all wrong, but here are
> definitions I usually associate with the terms.
>
> Server: A system and/or OS optimized or designed to provide optimal
> performance, at the expense of the user interface, for service level
> applications like file servers, SQL database servers, web (HTTP) server,
> FTP, etc. Low level performance and stability are the key measures of
> quality.
>
> Desktop: A system and/or OS optimized or designed to provide optimal
> user interface performance and simplicity, at the possible expense of
> stability and low level performance.
>
> Workstation: A system which weighs the advantages of both desktop and
> server and provides a reasonable compromise between the two. Stability
> and low level performance are very important, but usability is also
> important.
I think that the terms server and workstation are not in the same level
that desktop is. I mean, a desktop can be attached to both of them. For
me a workstation is a computer you use to achieve a job, be it reading
mail be it doing scientific calculations, so it depends on the task you are
doing, if you want/need a desktop or not. I dobasically agree in your
server definition.
------------------------------
From: Shane Phelps <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Linux 2.4 Major Advance
Date: Tue, 16 Jan 2001 21:20:54 +1100
Ayende Rahien wrote:
>
> "Shane Phelps" <[EMAIL PROTECTED]> wrote in message
> news:[EMAIL PROTECTED]...
> >
> >
> > J Sloan wrote:
> > >
> > > Chad Myers wrote:
> > >
> > > > Well, people who know more about this than I, including the defenders
> of Linux
> > > > were agreeing that the Tux used in SpecWeb99 was running in kernel
> mode.
> > >
> > > Nope, wrong again - the wintrolls claimed that it's kernel mode.
> > > I pointed out more than once what Tux architect Ingo Molnar
> > > said, and I'll say it once again for your benefit:
> > >
> > > The specweb tests were done with a user mode Tux.
> > >
> > > Tux can run in either mode -
> > >
> > > khttpd, which you confused with Tux, is an experimental
> > > in-kernel webserver (that takes about 26k of memory)
> > >
> > > jjs
> >
> >
> > The early assumption was that tux was largely running in kernel space,
> > and it took a while to clear this point up.
> >
> > Interestingly, it appears that "IIS 5" was actually "SWC 3" which is
> > apparently a kernel-mode cache (which can update time-stamps). This
> > also took some time to surface :-)
>
> Where did you hear that?
> Any sources to confim this? I'm asking, btw, not attacking.
On one of the threads on comna, actually.
I don't remember which one offhand (may even have been part of this
thread), but it shouldn't be too hard to find on comna or cola
I can't vouch for the accuracy of the postings, and I didn't check the
original material, but there were certainly some reasonable arguments
to suggest that most of the work was being done by SWC 3 rather than
IIS 5, and there's a comment in the spec benchmark source to show that
SWC 3 is running in kernel mode.
I happen to think that all this "mine is faster than yours" stuff
is silly anyway. Absolute speed in a single webserver is just
chest-beating anyway (or maybe beating something else instead :-)
Kernel-mode and ultra-lightweight web servers reek of benchmark specials
wherever they come from.
BTW, I got severely flamed by a thouroughly dishonest and unpleasant
winvocate
by the name of Conrad Rutherford, possibly for having a dig at benchmark specials.
I'm glad he's not representative of the general standard of poster.
------------------------------
From: "Tom Wilson" <[EMAIL PROTECTED]>
Subject: Re: Definition: Desktop, Workstation, Server.
Date: Tue, 16 Jan 2001 10:24:55 GMT
"David Adams" <[EMAIL PROTECTED]> wrote in message
news:9414eo$h4a$[EMAIL PROTECTED]...
>
> "mlw" <[EMAIL PROTECTED]> escribió en el mensaje
> news:[EMAIL PROTECTED]...
> > I don't know about you, but seeing "desktop" used interchangeably with
> > "Workstation" seems a bit extreme. Perhaps I am all wrong, but here are
> > definitions I usually associate with the terms.
> >
> > Server: A system and/or OS optimized or designed to provide optimal
> > performance, at the expense of the user interface, for service level
> > applications like file servers, SQL database servers, web (HTTP)
server,
> > FTP, etc. Low level performance and stability are the key measures of
> > quality.
> >
> > Desktop: A system and/or OS optimized or designed to provide optimal
> > user interface performance and simplicity, at the possible expense of
> > stability and low level performance.
> >
> > Workstation: A system which weighs the advantages of both desktop and
> > server and provides a reasonable compromise between the two. Stability
> > and low level performance are very important, but usability is also
> > important.
>
> I think that the terms server and workstation are not in the same
level
> that desktop is. I mean, a desktop can be attached to both of them. For
> me a workstation is a computer you use to achieve a job, be it reading
> mail be it doing scientific calculations, so it depends on the task you
are
> doing, if you want/need a desktop or not. I dobasically agree in your
> server definition.
I always considered a workstation to be a desktop tied to a server or
otherwise participating in a network.
Hmmm, anyone up for a silly semantics war ? :)
--
Tom Wilson
Sunbelt Software Solutions
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Linux 2.4 Major Advance
Date: Tue, 16 Jan 2001 10:39:33 GMT
Not all results are RAID 0 based, I have found a RAID 1 based Linux SpecWeb99
result:
http://www.spec.org/osg/web99/results/res2000q4/web99-20001127-00077.html .
It says: "8 disk software RAID1 mirror for fileset", and it's the fastest 2
CPU result. Sounds redundant enough to me.
Thomas
> >
> > <[EMAIL PROTECTED]> wrote in message
> > news:93ukgr$42t$[EMAIL PROTECTED]...
> > > Chad, I think both sides may have a point here. A comment in this
> thread
> > > suggested to read this URL:
> > > http://slashdot.org/interviews/00/07/20/1440204.shtml , it's a
> technical
> > > description of Tux. What appears to have happened is this: Tux serves
> > dynamic
> > > web pages (read: complex code) from userspace. It serves static web
> pages
> > > (read: simple stuff) directly from kernel-space. Tux also has the
> ability
> > to
> > > generate dynamic web pages in kernelspace, but this was not used in the
> > > SpecWeb99 test. From a security point of view this is a valid
> > architecture:
> > > the overwhelming majority of security problems originate from dynamic
> > code,
> > > serving static web pages is a well defined and simple thing.
> >
> > I wouldn't say that SPECWeb99 is the place to look for sensible web
> > arcitecture.
> > Those machines are tuned for speed, and speed only.
> > This is fine in a benchmark, in RL, any sane sysadmin would tune those
> > machines very differently.
> > (Who put RAID-0 on a production machine, frex?)
>
> Good point...
> RAID-5 would have been a bit more realistic.
>
> --
> Tom Wilson
> Sunbelt Software Solutions
>
>
Sent via Deja.com
http://www.deja.com/
------------------------------
From: "Tom Wilson" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Linux 2.4 Major Advance
Date: Tue, 16 Jan 2001 10:59:18 GMT
<[EMAIL PROTECTED]> wrote in message
news:9418d5$aun$[EMAIL PROTECTED]...
> Not all results are RAID 0 based, I have found a RAID 1 based Linux
SpecWeb99
> result:
> http://www.spec.org/osg/web99/results/res2000q4/web99-20001127-00077.html
.
> It says: "8 disk software RAID1 mirror for fileset", and it's the fastest
2
> CPU result. Sounds redundant enough to me.
>
> Thomas
Its' nice to see one that has a closer-to-real-world configuration.
I'm curious to see how RAID Level 5 fares since I'm partial to the
configuration for the protection it offers. I'm willing to sacrifice some
speed to achieve maximum reliability.
<snippage>
--
Tom Wilson
Sunbelt Software Solutions
------------------------------
From: Shane Phelps <[EMAIL PROTECTED]>
Subject: Re: Kernel space? Who gives a @#$%
Date: Tue, 16 Jan 2001 22:07:40 +1100
Conrad Rutherford wrote:
>
> "Peter Köhlmann" <[EMAIL PROTECTED]> wrote in message
> news:[EMAIL PROTECTED]...
> > Conrad Rutherford wrote:
> >
> > >
> > > and the point is?
> > >
> > > two nearly identical machines and they produced results within 3% of
> each
> > > other - so much for the "obvious" superiority of linux smoking the
> > > "bloated" GUI laden Windows...
> > >
> > >
> > What are you smoking?
> > Two nearly "identical" machines? Are you mad?
> > Reread what machines were used and then come back.
> >
>
> The ONLY difference is 15K rpm speed for some of the drives on the W2K
> server, everything else was exactly identical - reread yourself.
apart from the following section of Bones' post:
"It looks like the tuners felt there was an I/O bottleneck with disk access,
and the hardware is *not* the same with these two machines. The Linux box
^^^^^^^^^^^^^^^^^^^^^^^^^^^
uses the same five SCSI disks and configuration as I mentioned above. The
Windows machine uses *NINE* disks: One disk for the OS, two striped and
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
containing logs, and the remaining six drives have the data striped across
them. Eight of those nine are 15K RPM drives, unlike the 10K RPM drives
Linux is using. Apparently the solution to a Windows performance problem is
to get more powerful hardware involved (on the server anyway), not make the
OS more efficient. I am not surprised."
The NT disk layout seems to be better. DBAs do this all the time to keep
the redo logs on separate spindles from their tablespace, and also tune
tablespace allocation to reduce disk hotspots.
Do you *always* snip relevant data and quote out of context?
------------------------------
** 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
******************************
