John Doe and Linux (Re: Linux vs FreeBSD)
Tor Slettnes
>>>>> "Curtis" == Curtis D Levin <cdl...@dania.dialisdn.com> writes:
Curtis> Noone wants to split the market more than MS. The plain
Curtis> truth is that MS stuff is easier to use.
Initially. There is nothing as goodlooking as a fresh Windows install,
with as much "fun" built into GUI components like the control panel.
With the insane amount of processor power available nowadays, it also
starts up relatively quickly.
The registry gets a little out of whack, but what the heck, it is also
"fun" reinstalling Windows. A few BSOD's occur - initially they are a
novelty ("hey cool!"), and soon you learn how to avoid most of them
anyway. You install a few fonts and applicatons, and the machine gets
a little slower, but you don't notice because it is a gradual process.
No, sir, easier to use it is not. Even the average John knows that
the Macintosh has that "advantage" over Windows. Then again, that
alone is not what counts.
People also want power. Microsoft has people believe that Windows NT
is as powerful as UNIX, but with an "intuitive" interface. A couple
of quotes from <http://www.microsoft.com/ie/unix/devs.htm>:
...crossing the industry equivalent of the Atlantic Ocean from the
Old World command-line traditions of UNIX to the GUI New World of
Windows, confronted by a host of strange new priorities:
intuitiveness, discoverability, usability
"It's amazing to me how far UNIX has to go today to catch up to
NT," says Dawson. "Take, just for one example, threading support.
UNIX still has benefits, but NT is just a lot more full-featured."
I kindof want to puke.
X-Windows was there way before MS Windows 3 (Windows 2 was hardly a
GUI, running in text mode), with features like remote windows (MS
Windows still lacks it), multiple displays, replacable Window Manager
(Windows actually has that too, but not very many people know it - and
there is not really any alternatives to Microsoft's anymore). Etc.
Threads are something Windows needed because of other deficiencies,
such as very inefficient task spawning, lack of a select() call on
input streams, lack of IPCs such as shared memory, semaphores, message
queues. Nevertheless, there are POSIX threads which almost all UNIXen
support by now.
And UNIX is rock solid, despite rapid development by thousands of
people. Windows is simply alpha quality software marketed way beyond
its capability. It deserves its inevitable fate - when the current
Linux-aware corporate techies gets "promoted" into manager positions
they will remember still having to deal with Windows (due to current
policies in many companies), and turn away with a vengeance.
Curtis> Redhat Linux makes great use of this by using the GUI to
Curtis> some extent to ease the burden.
Also RedHat Linux is doing well among an increasing number of morons.
You know - those people who are brought into this world to complain
that everything is not served on a plate and fed with a spoon,
forgetting that this is free software and that they are (a) not forced
to use it, and (b) lucky to have it.
FreeBSD has more "snob appeal", as Alan Calawhatever so eloquently
said it. (Debian just has more technical appeal, and I am a bigot :-)
To put it this way, Linux "support" groups are seing an increasing
number of AFAQ (annoyingly frequently asked questions) such as "How do
I get the RedHat CD out of my box and into my drive?" or "I have a
400Mhz PII-MMX CPU, what IRQ is that?" Failure to respond to the
most obscure and illegible questions usually results in a "you guys
suck, Linux is doomed to failure" post.
Yes, RedHat is making money. Whether the Linux community as a whole
is well served by too sudden of an influx is another matter. Though
it should be noted that a larger market means higher availability of
commercial software.
Curtis> Put that into perspective when dealing with all types of
Curtis> unixes. Then understand the end user who wants everything
Curtis> to work. Understand the secretary who is on the phone to
Curtis> the IT dpt every time the system does something strange.
Curtis> Then you should have a better perspective on why there is
Curtis> so much MS.
Unfortunately, there is also here a very large discrepancy between
marketing and reality. By definition, if a server does "something
strange" (including BSOD) chances are it is running Windows NT.
Curtis> 2) Advertise. On web pages, on yahoo, etc. Market unix
Curtis> like MS does win products.
No!
First of all, free UNIXen are not supported by the amount of money
available in Microsoft's advertising budget.
More importantly, the success of free software has to do with its
"grassroot appeal" and "paradigm shift" - not that same trite old
marketing. Commercial advertisements are by definition targeted
towards the "lowest common IQ"; free software appeals to an audience
that appreciates sophistication. That also means they want to
discover the truth for themselves, and are rather unimpressed by
something that aims at insulting their intelligence like marketing
usually does.
I will also tell you this: The first one of FreeBSD and Debian to
extensively engage in stupid advertisements is the first OS off my
machine. (I currently keep both on to keep some software of mine
portable - I also use my Sun station at work for that).
Curtis> 3) Better auto detection of hardware on installation. Make
Curtis> installation easier. More people will install it.
That is very true. Score.
Curtis> 4) Develop, develop, develop.
Yes.
Curtis> And charge charge charge for your efforts.
No. Then we are back to same old, same old. The appeal of free
software, and Linux in particular, is the openness and sharing that
takes place. No worries about licenses, violations, copy bans.
At the very least - charge for something like manuals, support
etc. instead. Just don't hinder redistribution of software. That
would make our Great New World of Freedom become like that Crummy Old
Windows World with "I Agree To Become Bill's Towel-boy by opening this
package" software and Lawyer Gestapos.
Curtis> 5) Smile.
Yeah! Second score!
What RMS tried for fourteen or so years Linus Torvalds did in one or
two. The difference? He smiles.. :-)
-tor
Rajat Datta
On 03 Jul 1998 00:35:08 -0700, Tor Slettnes <t...@netcom.com> wrote:
>Yes, RedHat is making money. Whether the Linux community as a whole
>is well served by too sudden of an influx is another matter. Though
>it should be noted that a larger market means higher availability of
>commercial software.
Also, more non-commercial software. There is a tremendous amount of
development going on in the free software world that has been fueled
by Linux and Linux-folk. A lot of experimentation with X is going on
and I think that's going to be for the good. A lot of the projects
are clearly bogus, but I use tools everyday that was developed
primarily on Linux, and the FreeBSD distribution carries them. Some
of the tools, like GIMP, show enormous promise.
So what if there are a lot of less savvy people getting interested
in Linux (and maybe other Unix clones)? There will be some who are
genuinely interested in projects that cater to them, and you might
get something like KDE (it's not to my taste, but I can appreciate
what they've accomplished). It doesn't detract from those who are
more at home with CLI and find it more powerful (like myself). But
maybe, catering to those computer "users" brings about more tools
like the GIMP, like LyX, etc.
As far as Linux vs. FreeBSD, what's good for Linux will be good for
FreeBSD. Unless, of course, most FreeBSD-ers really do want to see
FreeBSD "take over the world". There do seem to be quite a few
people in Linux-land who really have that as a mission, and good
hunting to them. What they bring about in terms of more open
information about hardware for device drivers, more commercial
software, etc. will benefit all the free OSes.
rajat
Christopher B. Browne
On 3 Jul 1998 08:29:25 GMT, Rajat Datta <ra...@goteborg.netcom.com> posted:
>>is well served by too sudden of an influx is another matter. Though
>>it should be noted that a larger market means higher availability of
>>commercial software.
>
>development going on in the free software world that has been fueled
>by Linux and Linux-folk. A lot of experimentation with X is going on
>and I think that's going to be for the good. A lot of the projects
>are clearly bogus, but I use tools everyday that was developed
>primarily on Linux, and the FreeBSD distribution carries them. Some
>of the tools, like GIMP, show enormous promise.
>
>So what if there are a lot of less savvy people getting interested
>in Linux (and maybe other Unix clones)? There will be some who are
>genuinely interested in projects that cater to them, and you might
>get something like KDE (it's not to my taste, but I can appreciate
>what they've accomplished). It doesn't detract from those who are
>more at home with CLI and find it more powerful (like myself). But
>maybe, catering to those computer "users" brings about more tools
>like the GIMP, like LyX, etc.
More "non-hackers" has three results:
a) It doesn't directly result in any code being written.
Unfortunately, those people that most want things like "Word for Linux"
tend to also be a group made up of people that are likely to be the
*least* capable to actually create it.
b) If non-coders send money in the direction of coders, this may
result in the creation of new code that the non-coders are interested
in seeing.
Of course, at $10 per person, it takes a whole *PILE* of of non-coders
to add up into anything that could cover the costs of supporting a
coder.
c) Representation by population.
Those incapable of coding nonetheless make up an increasing proportion
of the community. And therefore have an increasing expectation of
influencing the direction of the community.
If they are at least contributing *somehow,* perhaps financially,
perhaps by working on documentation, or testing, or something of the
sort, it may be reasonable to have an expectation of having influence.
If they aren't, then that isn't particularly reasonable.
The point at which things fracture is when non-developers start figuring
that they can tell developers what to do, whilst developers figure that
since they write the code, it is their right to call the shots.
I strongly suspect that *this* is what will eventually result in Linux
developers moving on to other systems, and will be the impetus for the
next such system. For instance, kernel hackers might get sick of the
"I want another GDI driver" pleas, and move to working on Hurd instead.
>As far as Linux vs. FreeBSD, what's good for Linux will be good for
>FreeBSD. Unless, of course, most FreeBSD-ers really do want to see
>FreeBSD "take over the world". There do seem to be quite a few
>people in Linux-land who really have that as a mission, and good
>hunting to them. What they bring about in terms of more open
>information about hardware for device drivers, more commercial
>software, etc. will benefit all the free OSes.
To the extent to which Linux efforts result in:
a) Code that is at least *somewhat* portable,
b) The release of specifications,
c) An increase of interest in UNIX-like systems,
Linux activities are "good" for FreeBSD. (And for NetBSD. And OpenBSD.
And Hurd. And perhaps whatever the new OS is that John Dyson is
working on. And the list can probably go on...)
--
Those who do not understand Unix are condemned to reinvent it, poorly.
-- Henry Spencer <http://www.hex.net/~cbbrowne/lsf.html>
cbbr...@hex.net - "What have you contributed to Linux today?..."
jedi
On Sun, 05 Jul 1998 06:17:31 GMT, BR <ben...@mailhost.ind.ameritech.net> wrote:
>Bloody Viking wrote:
><small snippage>
>
>> The "insane" computing power of modern PCs means that anyone with a credit
>> card can buy a computer with the power of the earliest Cray
>> supercomputers. This is what I find exciting. Yep, even postal workers can
>> have a cray! Crays use UNIX too.
>
>Just a small interjection here. If it wasn't for the duo poly known as
>MS and Intel. The consumer probably wouldn't have as much power as they
>have in such a short time frame.
Before M$ and Intel brought us the blight that was the
PC, it's competitors were more than capapble of delivering
systems (even then in 1981) of similar capacity of those today.
The i86's were just playing catch up to the 68k's for a
considerable amount of time.
--
Hardly. Microsoft has brought the microcomputer OS to
the point where it is more bloated than even OSes from |||
what was previously larger classes of machines altogether. / | \
This is perhaps Bill's single greatest accomplishment.
fl...@interport.net
On Sun, 05 Jul 1998 06:06:07 GMT, Kevin Huber <khu...@yuck.net> wrote:
>someone wrote:
>>> A lot of experimentation with X is
>>> going on and I think that's going to be for the good.
>
>precautions.
It helps if you drink enough water, especially if you're dancing. Don't
go home with strangers if it's your first time. Etc.
--
Ben
root@localhost (hi spammers!)
jedi
>
>I'm not arguing your above points. My point was,for good or bad.
>Feature-itus and bloat,put us on the upgrade path. Do you think that we
>would have pentiums if it wasn't for the above? Now all a person has to
Yup. We had Pentium class microprocessors and their equivalents
through multiprocessing a considerable time before Pentiums were
available. Bloatware is more a justification than a cause. A
pentium-requiring wordprocessor is a perfect excuse to get
consumers to buy pentiums and certainly help unimaginitive
software vendors.
>do is run more efficient software on that pentium,viola instant cray. :)
--
Hardly. Microsoft has brought the microcomputer OS to
the point where it is more bloated than even OSes from |||
what was previously larger classes of machines altogether. / | \
This is perhaps Bill's single greatest accomplishment.
Michele Marie Dalene
I started out with a Commodore 64 as well. I also used a PDP-11 minicomputer
, Ti 99/4a, Commodore Pets, Commodore Vic-20 (not good fur much) and my
very first computer was a Timex Sinclair 1000 with a measly 2k of ram
I too leared how to use the Command line exclusively. Being that
I have a visual disability, I cannot manage a GUI very well. Thus a
Commodore Amiga, Macintosh or Windows (any flavor) is trying for me.
I also btw type Unix commands in dr-dos 7.02. Too bad one
cannot alias dos commands to unix names like you can in bash.
for example cp=copy.
I went for a short time back to dos to run a bbs. People loved
the bbs but I did not like dos. within two months I returned back to
a full linux only system. Now I have terminals hooked up to this thing
(a Wyse 50, an Old XT, and a DEC 100 Rainbow. My system is old, its only
a 486DX2/66mhz box with 16mb, originally it had only CGA. but being I
only used text. it never mattered much.. Tried Xwindows. but on this thing
the constant SWAPPING in/out was too much to bear. plus I had a hard time
using the mouse anyway. Linux/Unix allows you to choose. either CLI or GUI.
Plus I prefer VI to Microsofts editor or even Caldera's DR-dos 7.02
editors. I learned VI on my Tandy Color computer 2 running Os9 Level 1.
Linux works well on my system. even with only 600mb spread over
two 380MB and one 40mb SCSI drives. and occassionally I bring my Xts
Bernoulli box over to do a major file copy session. Sometimes I even fool
my adaptec SCSI card to boot the dos bernoulli cartridge to do installs
for the XT as it only has 360K floppies and this sucker has a 1.44 3 1/2".
Linux is great for me, as I often get lots of old equipment. I
now have TWO tape drives on this. a Iomega 800 Ditto and a WAngtek 5150es
(scsi) drive. Does anyone have any DC6120 or DC6250 cartridges for it?
As you are aware. Linux does not need me to fuss with EMM386 or
himem.sys. it just works all on its own to manage its ram. I like also
the fact that I have mine set up for 3 text virtual consoles. I took
the other 3 that slackware gave me out as 3 was enough for most uses.
In article <6nmjmc$n6f$1...@hirame.wwa.com>, Bloody Viking wrote:
>started computering in the Commodore 64 days, I got used to the idea of a
>CLI. The "OS" was a kernel with a BASIC interpreter for a shell. When I
>first bought a PC in 1993, it took about 2 weeks to find out how to get to
>DOS, and QBASIC, in which case, I felt at home again. I always thought of
>a CLI as "real" computering, as that was what I first saw. With Linux,
>it's a CLI unless you light off X.
>
>The "insane" computing power of modern PCs means that anyone with a credit
>card can buy a computer with the power of the earliest Cray
>supercomputers. This is what I find exciting. Yep, even postal workers can
>have a cray! Crays use UNIX too. The fun question of Linux being UNIX
>
>This Joe Average is quite happy with Linux. That's why I use it, though at
>the beginning, I used it becuse of software costs. I've gotten used to it.
>When I'm at a DOS prompt, I end up typing in UNIX commands, only to get
>the error that DOS can't understand UNIX commands. Linux is fun! Linux,
>like beer, is one of those acquired taste things. Is Linux ready for the
>Joe Average? Maybe, maybe not, but this Joe Average is happy with it!
--
B'ichela
When a Klingon goes to Linux. It means one must respect the Klingons
pride and honorable decision to work their computers to their fulllet
potential.
Linux is Proudly displayed to all my friends. and I help get
others going on this powerful alternative to the watered down mush
served by Microslop.
Steinar Haug
[jedi]
| >Just a small interjection here. If it wasn't for the duo poly known as
| >MS and Intel. The consumer probably wouldn't have as much power as they
| >have in such a short time frame.
|
| Before M$ and Intel brought us the blight that was the
| PC, it's competitors were more than capapble of delivering
| systems (even then in 1981) of similar capacity of those today.
Bull. I used 68k based Sun systems in the middle of the 80s. There's
no way they could compete with a well equipped PC of today (assuming
the PC runs a decent operating system).
Steinar Haug, Nethelp consulting, sth...@nethelp.no
BR
jedi wrote:
>
> On Sun, 05 Jul 1998 06:17:31 GMT, BR <ben...@mailhost.ind.ameritech.net> wrote:
> >Bloody Viking wrote:
> ><small snippage>
> >
> >> card can buy a computer with the power of the earliest Cray
> >> supercomputers. This is what I find exciting. Yep, even postal workers can
> >> have a cray! Crays use UNIX too.
> >
> >MS and Intel. The consumer probably wouldn't have as much power as they
> >have in such a short time frame.
>
> Before M$ and Intel brought us the blight that was the
> PC, it's competitors were more than capapble of delivering
> systems (even then in 1981) of similar capacity of those today.
>
> considerable amount of time.
>
> --
> Hardly. Microsoft has brought the microcomputer OS to
> the point where it is more bloated than even OSes from |||
> what was previously larger classes of machines altogether. / | \
> This is perhaps Bill's single greatest accomplishment.
I'm not arguing your above points. My point was,for good or bad.
Feature-itus and bloat,put us on the upgrade path. Do you think that we
would have pentiums if it wasn't for the above? Now all a person has to
do is run more efficient software on that pentium,viola instant cray. :)
--
************************
* Enjoy the pane-Run NT*
************************
Ted Mittelstaedt
BR <ben...@mailhost.ind.ameritech.net> wrote in article
<359F19BA...@mailhost.ind.ameritech.net>...
> Bloody Viking wrote:
> <small snippage>
>
> > The "insane" computing power of modern PCs means that anyone with a
credit
> > card can buy a computer with the power of the earliest Cray
> > supercomputers. This is what I find exciting. Yep, even postal workers
can
> > have a cray! Crays use UNIX too.
>
> Just a small interjection here. If it wasn't for the duo poly known as
> MS and Intel. The consumer probably wouldn't have as much power as they
> have in such a short time frame.
>
Intel, yes. MS- no. In the history of the PC, Microsoft improvements in
software have always come _AFTER_ complimentary improvements in the
hardware.
The real users driving the Intel chips have been the gamers running under
straight DOS. (and the server people, to a far lesser extent) These people
always buy the fastest, most powerful chips despite the expense.
Ted
Alexander Viro
In article <slrn6q0am0....@pinkrose.snet.net>,
Michele Marie Dalene <mda...@snet.net> wrote:
[snip]
> I went for a short time back to dos to run a bbs. People loved
>the bbs but I did not like dos. within two months I returned back to
>a full linux only system.
FreeBSD. Why bother with DOS?
[snip]
>--
> B'ichela
>When a Klingon goes to Linux. It means one must respect the Klingons
>pride and honorable decision to work their computers to their fulllet
>potential.
> Linux is Proudly displayed to all my friends. and I help get
>others going on this powerful alternative to the watered down mush
>served by Microslop.
Lady, it's all nice and dandy, but you'ld better watch the
newsgroups list ;-) I'm reading it c.u.b.freebsd.misc (albeit from my
Linux box; FreeBSD one is in the middle of _big_ compile).
Cheers,
Al
Followups set.
--
"You're one of those condescending Unix computer users!"
"Here's a nickel, kid. Get yourself a real computer" - Dilbert.
jedi
On 05 Jul 1998 21:45:59 GMT, Steinar Haug <sth...@nethelp.no> wrote:
>[jedi]
>
>| >Just a small interjection here. If it wasn't for the duo poly known as
>| >MS and Intel. The consumer probably wouldn't have as much power as they
>| >have in such a short time frame.
>|
>| PC, it's competitors were more than capapble of delivering
>| systems (even then in 1981) of similar capacity of those today.
>
>no way they could compete with a well equipped PC of today (assuming
>the PC runs a decent operating system).
Unless you're D2, it really shouldn't matter.
That's the whole point. Except for better games,
the bloat parade really hasn't devlivered anything
that's really better. We just now have Wordprocessors
that need ungodly system resources just to run...
--
Hardly. Microsoft has brought the microcomputer OS to
the point where it is more bloated than even OSes from |||
what was previously larger classes of machines altogether. / | \
This is perhaps Bill's single greatest accomplishment.
Steinar Haug
[jedi]
| >Bull. I used 68k based Sun systems in the middle of the 80s. There's
| >no way they could compete with a well equipped PC of today (assuming
| >the PC runs a decent operating system).
|
| Unless you're D2, it really shouldn't matter.
|
| That's the whole point. Except for better games,
| the bloat parade really hasn't devlivered anything
| that's really better. We just now have Wordprocessors
| that need ungodly system resources just to run...
You're entitled to your point of view, of course. I can even symphatize,
once in a while. But most of the time I'm *much* happier running emacs
on a P-166 than a 16 MHz Sun-3/50 - and I'm also quite certain that I
am more productive today than I was with a Sun-3/50. Not ten times as
productive, though...
Toon Moene
vi...@steklov.math.psu.edu (Alexander Viro) wrote:
> Lady, it's all nice and dandy, but you'ld better watch the
> newsgroups list ;-) I'm reading it c.u.b.freebsd.misc (albeit from my
> Linux box; FreeBSD one is in the middle of _big_ compile).
Hey, you don't want to claim that FreeBSD is a single tasking OS, don't you ?
[ I never understand these young wippersnappers - have 20 times as much
compute power as I do and *still* can't run more than one task on one
computer ]
... while typing this in foreground having our Numerical Weather Prediction
system running in the background ...
--
Toon Moene (mailto:to...@moene.indiv.nluug.nl)
Saturnushof 14, 3738 XG Maartensdijk, The Netherlands
Phone: +31 346 214290; Fax: +31 346 214286
g77 Support: mailto:for...@gnu.org; NWP: http://www.knmi.nl/hirlam
Alexander Viro
In article <6nratq$843$3...@sucker.nl.uu.net>,
Toon Moene <to...@moene.indiv.nluug.nl> wrote:
>vi...@steklov.math.psu.edu (Alexander Viro) wrote:
>
>> Lady, it's all nice and dandy, but you'ld better watch the
>> newsgroups list ;-) I'm reading it c.u.b.freebsd.misc (albeit from my
>> Linux box; FreeBSD one is in the middle of _big_ compile).
>
>Hey, you don't want to claim that FreeBSD is a single tasking OS, don't you ?
Sure it isn't. Ever heard of -j option of make? ;-) But to be serious:
rebuilding gcc and X server isn't a nice experience on a box with 12M RAM.
And if I have another box sitting near - what's the problem with moving my
ass two feets left?
>[ I never understand these young wippersnappers - have 20 times as much
> compute power as I do and *still* can't run more than one task on one
> computer ]
Oh, yeah... PDP-11 with 32K, 8 users and sucking OS (RSX). 3 TKBs
in parallel and everybody are deep in space ;-)
Cheers,
Al
jedi
[IOW: I need the cpu cycles to run my legacy bloatware.]
You've got to be kidding? Emacs is one of the bloatware KINGS.
Although, unlike it's little cousin (MS-Word) it doesn't
vendor-lock more austere solutions out of the marketplace.
Jed users can use a 16Mhz machine and not worry about being
locked out of the world inhabited by Emacs users...
--
Hardly. Microsoft has brought the microcomputer OS to
the point where it is more bloated than even OSes from |||
what was previously larger classes of machines altogether. / | \
This is perhaps Bill's single greatest accomplishment.
John Fieber
In article <slrn6q2f4...@dementia.mishnet>,
je...@dementia.mishnet (jedi) writes:
> You've got to be kidding? Emacs is one of the bloatware KINGS.
Was. I'm afraid it has dropped quite far down the list if you
include Windows applications. With competition like Netscape
Communicator or any of the office suites, emacs doesn't stand a
chance in the bloatware competition, although XEmacs is still in the
running.
-john
jedi
--
Hardly. Microsoft has brought the microcomputer OS to
the point where it is more bloated than even OSes from |||
what was previously larger classes of machines altogether. / | \
This is perhaps Bill's single greatest accomplishment.
Ben Sandler
Rajat Datta wrote:
>
> As far as Linux vs. FreeBSD, what's good for Linux will be good for
> FreeBSD. Unless, of course, most FreeBSD-ers really do want to see
> FreeBSD "take over the world". There do seem to be quite a few
> people in Linux-land who really have that as a mission, and good
> hunting to them. What they bring about in terms of more open
> information about hardware for device drivers, more commercial
> software, etc. will benefit all the free OSes.
I have no experience with *BSD whatsoever (I got a BSDi account at work
two weeks ago, but I don't really use it). Is there binary
compatibility in between Linux and *BSD? If so, that certainly argues
your point.
- Ben
>
> rajat
--
Ben Sandler
email me: sandler at ymail dot yu dot edu
"Windows is an operating system, not a religion."
- Ted Waitt, chairman of Gateway
Steve Sheldon
jfi...@fallout.campusview.indiana.edu (John Fieber) writes:
>In article <slrn6q2f4...@dementia.mishnet>,
> je...@dementia.mishnet (jedi) writes:
>> You've got to be kidding? Emacs is one of the bloatware KINGS.
>Was. I'm afraid it has dropped quite far down the list if you
>include Windows applications. With competition like Netscape
>Communicator or any of the office suites, emacs doesn't stand a
>chance in the bloatware competition, although XEmacs is still in the
>running.
Hmm, I used to work with an app called Arc/Info. On our DECstations, it
took up around 500-600 Meg of harddrive space. It was a bit better on the
DEC Alphastation's with shared libraries, only 250-300 Meg.
Of course the Alpha binaries consumed more memory once run. A base machine
needed 64 Megs, and really needed 128 Megs to work semi comfortably. That
was in 1994 when 128 Megs of RAM cost a lot of money, as did 1 Gig drives.
But it definately utilized the FPU of that Alpha!
I haven't worked with very many applications since quite so large. But
then I also am no longer working with GIS software. :)
Ondra Koutek
Ben Sandler wrote:
>
> I have no experience with *BSD whatsoever (I got a BSDi account at work
> two weeks ago, but I don't really use it). Is there binary
> compatibility in between Linux and *BSD? If so, that certainly argues
> your point.
--
Ondra Koutek and...@sh.cvut.cz
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
This is the end...
Richard Lyon
>>once in a while. But most of the time I'm *much* happier running emacs
>>on a P-166 than a 16 MHz Sun-3/50 - and I'm also quite certain that I
>>am more productive today than I was with a Sun-3/50. Not ten times as
>>productive, though...
>
>[IOW: I need the cpu cycles to run my legacy bloatware.]
>
>
Err ... Like using an old 68K Apollo to compile code for an embedded
project. A full
compile and link use to take 6 hours. On a 200MHz Pentium it takes maybe
15 minutes irrespective of whether the host OS is FBSD or WINNT. I seem to
remember
SPARC 10 was pretty slow also.
Give me a 200MHz MMX, 64 MEG RAM and FBSD any day.
jedi
>>>You're entitled to your point of view, of course. I can even symphatize,
>>>once in a while. But most of the time I'm *much* happier running emacs
>>>on a P-166 than a 16 MHz Sun-3/50 - and I'm also quite certain that I
>>>am more productive today than I was with a Sun-3/50. Not ten times as
>>>productive, though...
>>
>>[IOW: I need the cpu cycles to run my legacy bloatware.]
>>
>> You've got to be kidding? Emacs is one of the bloatware KINGS.
>>
>
>
>Err ... Like using an old 68K Apollo to compile code for an embedded
>project. A full
>compile and link use to take 6 hours. On a 200MHz Pentium it takes maybe
>15 minutes irrespective of whether the host OS is FBSD or WINNT. I seem to
...and this is just how relevant to John Doe?
Besides, this is the sort of thing that is what
nice and batch queues were made for if you're
not running in a production enviroment.
>remember
>SPARC 10 was pretty slow also.
>
>Give me a 200MHz MMX, 64 MEG RAM and FBSD any day.
>
>
--
Hardly. Microsoft has brought the microcomputer OS to
the point where it is more bloated than even OSes from |||
what was previously larger classes of machines altogether. / | \
This is perhaps Bill's single greatest accomplishment.
Richard Tobin
>> You've got to be kidding? Emacs is one of the bloatware KINGS.
>Was. I'm afraid it has dropped quite far down the list
Indeed. "Eight Megabytes And Constantly Swapping" doesn't have the ring
to it that it did ten years ago.
-- Richard
--
Because of all the junk e-mail I receive, all e-mail from .com sites is
automatically sent to a file which I only rarely check. If you want to mail
me from a .com site, please ensure my surname appears in the headers.
User Rdkeys Robert D. Keys
> I have no experience with *BSD whatsoever (I got a BSDi account at work
> two weeks ago, but I don't really use it). Is there binary
> compatibility in between Linux and *BSD? If so, that certainly argues
> your point.
BSDI is not exactly like Linux or FreeBSD but more like SCO.
They are not exactly binary compatible, but they are source portable
compatible, and you can run some binaries between them.
FreeBSD and Linux are somewhat like apples and oranges. The BSD is
more of the ``real'' unix (whatever that is anymore), and Linux is
more of the clone unix (whatever that is anymore). To the end user,
they are very much similar until you get down deep in the nooks and
crannies. In my hands, over the past 5 years, FreeBSD has always
been a bit more stable (subjectively and in practice) than Linux.
(I can't comment on BSDI, since I have never run that variant.)
Linux always has a few more bells and whistles. Things port back
and forth between the two rather easily. Most of what one has is
on the other. I look at it this way --- if stability counts, then
I use FreeBSD (or one of the other BSD's if on other platforms), else,
I use either that has what I need to work with. Me, I like the
structure of FreeBSD over all the others. Also, it builds up and
adds to with a nicer degree of consistency than all the rest.
I would rate FreeBSD at the 99th percentile, and Linux at the 95th
percentile, with the other BSD's at the 90th percentile, and the
rest below. That is my feelings on the matter.
> - Ben
Ben.... all are better than that other OS that always nukes my wife's
machine..... such is the joy of unix.....(:+}}....
RDK
User Rdkeys Robert D. Keys
>> The "insane" computing power of modern PCs means that anyone with a credit
>> card can buy a computer with the power of the earliest Cray
>> supercomputers.
> Just a small interjection here. If it wasn't for the duo poly known as
> MS and Intel. The consumer probably wouldn't have as much power as they
> have in such a short time frame.
Yes and no. Instead of intel toys we would have 68000 based toys and
all be running unix with X instead of the ms clone with toy windoz.
Maybe the Z8000 series would have been competitive, instead. The end
result would have been much the same. By now comsumers would have had
some sort of powerbox-cpu with a different OS base (probably unix based).
GUI's would still have happened. Would CP/M 98 have taken over?
Interesting twist of fate.....RIP Gary....
RDK
User Rdkeys Robert D. Keys
> You're entitled to your point of view, of course. I can even symphatize,
> once in a while. But most of the time I'm *much* happier running emacs
> on a P-166 than a 16 MHz Sun-3/50 - and I'm also quite certain that I
> am more productive today than I was with a Sun-3/50. Not ten times as
> productive, though...
> Steinar Haug, Nethelp consulting, sth...@nethelp.no
Yes, but rather than a 16mhz Sun box, it would probably be a 500mhz
Sun box, and the difference then would be minimal. Still, even a
16mhz Sun box or dos box has less bloat than a 100mhz Intel box
loaded for bear on NT. A lot depends upon the software bloat overhead.
Emacs is nice but a bloater, for sure, compared to a zippy little vi.
Vi on the 166 flat flies.
RDK
User Rdkeys Robert D. Keys
> Hardly. Microsoft has brought the microcomputer OS to
> the point where it is more bloated than even OSes from |||
> what was previously larger classes of machines altogether. / | \
> This is perhaps Bill's single greatest accomplishment.
Very interesting comparison.... A couple of days back I brought up
a V5 unix kernel in an emulator on a dos box and the kernel was only
25K in size. Bloat is, unfortunately, almost everywhere these days.
RDK
Steve Sheldon
>Very interesting comparison.... A couple of days back I brought up
>a V5 unix kernel in an emulator on a dos box and the kernel was only
>25K in size. Bloat is, unfortunately, almost everywhere these days.
Bah, I run CP/M 2.2 on a machine with 64K of RAM, and have 56K or so
available for applications.
25K? for an O/S kernel? Bah, talk about bloat!
void
>
>I have no experience with *BSD whatsoever (I got a BSDi account at work
>two weeks ago, but I don't really use it). Is there binary
>compatibility in between Linux and *BSD? If so, that certainly argues
>your point.
FreeBSD runs Linux binaries nicely. I don't know about the other BSDs. I
don't believe Linux runs BSD binaries, however.
Kevin Huber
Steve> Bah, I run CP/M 2.2 on a machine with 64K of RAM, and have 56K
Steve> or so available for applications.
Steve> 25K? for an O/S kernel? Bah, talk about bloat!
I was reading an old Dr. Dobb's Journal (yes, my life is that
exciting) from a couple years ago and they had a kernel for an
embedded system which supported distributed processing in less than
250 bytes. I rather suspect it was a little more simple-minded than
Beowulf. I/O abstractions are for weenies.
-Kevin
Fewtch
The term "bloatware" is very subjective, especially these days. For
example, the "average" new computer sold these days has a 7.2 gig hard
drive, 48-64 megs RAM and a 300 MHz processor. On a machine like
that, what would be considered "bloatware" on a Pentium 120 with a 1.2
gig hard drive and 16 megs RAM is pretty lightweight. In my opinion,
even products like MS Office aren't really bloated on a machine with
64 megs RAM and a 7 gig hard drive. It's all in how you look at it.
Say you have a 300 MHz machine with 128 megs RAM and an 8 gig HD. You
want to run only what you "objectively" consider to be non-bloatware.
It's likely that 6 gigs of that hard drive will go to waste and most
of the RAM and processor power go unused. So what's wrong with
running heavyweight software on heavyweight hardware?
On Tue, 07 Jul 1998 05:32:46 GMT, she...@visi.com (Steve Sheldon)
wrote:
>jfi...@fallout.campusview.indiana.edu (John Fieber) writes:
>
>>In article <slrn6q2f4...@dementia.mishnet>,
>> je...@dementia.mishnet (jedi) writes:
>
>>> You've got to be kidding? Emacs is one of the bloatware KINGS.
>
>>Was. I'm afraid it has dropped quite far down the list if you
>>include Windows applications. With competition like Netscape
>>Communicator or any of the office suites, emacs doesn't stand a
>>chance in the bloatware competition, although XEmacs is still in the
>>running.
>
>
> Hmm, I used to work with an app called Arc/Info. On our DECstations, it
>took up around 500-600 Meg of harddrive space. It was a bit better on the
>DEC Alphastation's with shared libraries, only 250-300 Meg.
>
> Of course the Alpha binaries consumed more memory once run. A base machine
>needed 64 Megs, and really needed 128 Megs to work semi comfortably. That
>was in 1994 when 128 Megs of RAM cost a lot of money, as did 1 Gig drives.
>
> But it definately utilized the FPU of that Alpha!
>
> I haven't worked with very many applications since quite so large. But
>then I also am no longer working with GIS software. :)
--
Spam Avoidance: Remove "dont_spam_me_" in my Email address to reply.
My website is at http://www.serv.net/~fewtch/index.html - something for everyone
there, take a look if you have the time! My tribute to the dearly loved cat of
19 years I recently lost is at http://www.serv.net/~fewtch/cats/index.html.
User Rdkeys Robert D. Keys
> "User Rdkeys Robert D. Keys" <rdk...@seedlab1.cropsci.ncsu.edu> writes:
>>Very interesting comparison.... A couple of days back I brought up
>>a V5 unix kernel in an emulator on a dos box and the kernel was only
>>25K in size. Bloat is, unfortunately, almost everywhere these days.
> Bah, I run CP/M 2.2 on a machine with 64K of RAM, and have 56K or so
> available for applications.
> 25K? for an O/S kernel? Bah, talk about bloat!
Yeah, I still run a 7K CP/M 2.2 kernel now and again, but I was
referring to a real unix kernel...... the earliest one that still
survives, from Dennis Ritchie..... Modern unix kernels tend to
500K and up, so that dinky little V5 25K kernel is tiny by comparison.
It still runs in 128K. The file system image is only 2.5 megs in size,
so it, the emulator, and dos to boot it up all fit on a 10meg HD from
an ancient PC running along at 4.77 mhz. Such is the joy of unix.....
All I need now is a model 37 to hook up as console.....(:+}}....
RDK
jedi
>
>The term "bloatware" is very subjective, especially these days. For
>example, the "average" new computer sold these days has a 7.2 gig hard
>drive, 48-64 megs RAM and a 300 MHz processor. On a machine like
>that, what would be considered "bloatware" on a Pentium 120 with a 1.2
>gig hard drive and 16 megs RAM is pretty lightweight. In my opinion,
>even products like MS Office aren't really bloated on a machine with
>64 megs RAM and a 7 gig hard drive. It's all in how you look at it.
Why bother with all that when most people can't demonstrate
use beyond what a $50 old used ST or Amiga could get them?
What's the point of a 'pentium' version of WordWriter?
--
Fewtch
I know this post was tongue-in-cheek, but look at it this way: CP/M
2.2 is pretty much featureless as far as OS's go. Does it multitask?
Multithread? Support any networking whatsoever? Etc, etc :-)
I'm not sure of the size of the MS-DOS 5.0 kernel (io.sys +
msdos.sys)... anyone here have any idea? Around 45-50k or so? Then
again, one might argue that MS-DOS doesn't even qualify for OS status,
since it really doesn't control any hardware whatsoever and lets the
BIOS handle just about everything... unless you'd consider himem.sys
and EMM386.EXE to be "part of the kernel" ;-)
On Tue, 07 Jul 1998 17:49:19 GMT, she...@visi.com (Steve Sheldon)
wrote:
>"User Rdkeys Robert D. Keys" <rdk...@seedlab1.cropsci.ncsu.edu> writes:
>
>>Very interesting comparison.... A couple of days back I brought up
>>a V5 unix kernel in an emulator on a dos box and the kernel was only
>>25K in size. Bloat is, unfortunately, almost everywhere these days.
>
> Bah, I run CP/M 2.2 on a machine with 64K of RAM, and have 56K or so
>available for applications.
>
> 25K? for an O/S kernel? Bah, talk about bloat!
Fewtch
...or even four years ago... !
On Tue, 7 Jul 1998 14:04:56 GMT, ric...@cogsci.ed.ac.uk (Richard
Tobin) wrote:
>Indeed. "Eight Megabytes And Constantly Swapping" doesn't have the ring
>to it that it did ten years ago.
>
>-- Richard
Fewtch
Naah, actually I think Office 97 will "run" on an 8 meg 486... if you
want to wait 10 minutes between cursor blinks while the system swaps
everything else out but the base OS kernel just to get the cursor to
blink on, then back off again.... :-) Type a key on the keyboard, and
you might wait an hour for it to show up on the screen... but it DOES
RUN... ;-) Aintcha never heard of swap files before?
On Tue, 7 Jul 1998 09:20:01 +0100, Jose Marques
<no...@nohow.demon.co.uk> wrote:
>On Mon, 6 Jul 1998, jedi wrote:
>> You've got to be kidding? Emacs is one of the bloatware KINGS.
>
>But it still works on an 8MB 486 - Word left that hardware behind how many
>years ago?
>
>[Snip]
Fewtch
Heheheh... that's my exact equipment as well. I find that when
compiling a new kernel, "make depend" takes approximately 30 seconds,
and "make" about 4 minutes. Under 5 minutes total to recompile the
kernel (so much for the old documents raving about it being "one of
the most time consuming and tedious tasks the Unix administrator has
to face"). I'd like to try it on a PII-400 :-)
On Tue, 7 Jul 1998 19:12:51 +1000, "Richard Lyon"
<c16...@email.mot.com> wrote:
>Give me a 200MHz MMX, 64 MEG RAM and FBSD any day.
Fewtch
So are you advocating that people not waste their money on new
hardware, and instead go buy an old Commodore 64 (or get one donated
to them free) and get productive? I'm not quite sure what you're
trying to say here.
On Tue, 7 Jul 1998 12:11:43 -0700, je...@dementia.mishnet (jedi) wrote:
>On Tue, 07 Jul 1998 18:27:58 GMT, Fewtch <fewtch@dont_spam_me_serv.net> wrote:
>>
>>The term "bloatware" is very subjective, especially these days. For
>>example, the "average" new computer sold these days has a 7.2 gig hard
>>drive, 48-64 megs RAM and a 300 MHz processor. On a machine like
>>that, what would be considered "bloatware" on a Pentium 120 with a 1.2
>>gig hard drive and 16 megs RAM is pretty lightweight. In my opinion,
>>even products like MS Office aren't really bloated on a machine with
>>64 megs RAM and a 7 gig hard drive. It's all in how you look at it.
>
> Why bother with all that when most people can't demonstrate
> use beyond what a $50 old used ST or Amiga could get them?
>
> What's the point of a 'pentium' version of WordWriter?
Fewtch
>I was reading an old Dr. Dobb's Journal (yes, my life is that
>exciting) from a couple years ago and they had a kernel for an
>embedded system which supported distributed processing in less than
>250 bytes. I rather suspect it was a little more simple-minded than
>Beowulf. I/O abstractions are for weenies.
So is reading old copies of Dr. Dobb's Journal (hehehe.. kidding...)
Fewtch
> Why bother with all that when most people can't demonstrate
> use beyond what a $50 old used ST or Amiga could get them?
Actually, that applies *only* to the most basic uses of a computer
(word processing, data entry, light spreadsheet work). Try any kind
of serious work with digital audio or image processing on that $50 ST
or Amiga, and let me know how far you get... How about mastering a
writable CD on that $50 Atari ST? Or working with a 1024x768 24-bit
color image? Or even browsing a web site with a few small Java
applets or one or two pictures? If you think "most people" only use
their computers for word processing and data entry these days, you're
still living in the dark ages.
Oh, I forgot to mention games... they've gotten pretty amazing lately
thanks to the latest hardware (in fact, computer gaming is one of the
primary driving forces behind hardware improvements, maybe even more
of a driving force than MS's increasingly larger apps and OS's).
John Fieber
fewtch@dont_spam_me_serv.net (Fewtch) writes:
> In my opinion,
> even products like MS Office aren't really bloated on a machine with
> 64 megs RAM and a 7 gig hard drive. It's all in how you look at it.
I tend to think in terms of utility provided for resources consumed,
not resources consumed in relation to resources available. Measuring
bloat in terms of resources available only serves to power the insane
upgrade treadmill.
I don't think any programmer should be unleashed on a modern PC until
they have put in a year or two in a nice confined environment--say a
Commodore 64--where they can learn the skills of simplicity and
efficiency in programming.
Of course, if bloat stopped the hardware industry would be thrown into
a state of crisis.
-john
Dave Kenny
: Before M$ and Intel brought us the blight that was the
: PC, it's competitors were more than capapble of delivering
: systems (even then in 1981) of similar capacity of those today.
:
: The i86's were just playing catch up to the 68k's for a
: considerable amount of time.
You mean Intel caught up?
I've spent many years programming in assembly language and have always
disliked Intel's "architectural" values from the 386 back through
the 8086, the 8080, the 8008, the 4040, and the 4004.
Intel likes special purpose everything. IIRC in every Intel processor
I programmed (those listed above) up through the 80286, EVERY SINGLE
REGISTER was unique. That is for any given register, there was
some operation you could perform ONLY using that register.
The PDP11, the IBM 360 etc, the Dec VAX, the Motorola 680x0 processors,
along with some others I've used, all had reasonably general registers,
although most of them split the register set in half between data
and address registers.
The thing I hated most though was the 8088/8086/80x86 segmented memory
model. <shudder> <yuck!> The best thing they introduced (in the 80386)
as a flat memory space like normal computers.
jedi
>
>So are you advocating that people not waste their money on new
>hardware, and instead go buy an old Commodore 64 (or get one donated
>to them free) and get productive? I'm not quite sure what you're
>trying to say here.
If an ST would suffice, then it should be able to instead
of the current cascading vendor-lock that occurs in the
present market whenever anyone with a little free will
wants to exchange data.
Bill has priced more people out of the market than Steve,
ironically enough...
>
>On Tue, 7 Jul 1998 12:11:43 -0700, je...@dementia.mishnet (jedi) wrote:
>
>>On Tue, 07 Jul 1998 18:27:58 GMT, Fewtch <fewtch@dont_spam_me_serv.net> wrote:
>>>
>>>The term "bloatware" is very subjective, especially these days. For
>>>example, the "average" new computer sold these days has a 7.2 gig hard
>>>drive, 48-64 megs RAM and a 300 MHz processor. On a machine like
>>>that, what would be considered "bloatware" on a Pentium 120 with a 1.2
>>>gig hard drive and 16 megs RAM is pretty lightweight. In my opinion,
>>>even products like MS Office aren't really bloated on a machine with
>>>64 megs RAM and a 7 gig hard drive. It's all in how you look at it.
>>
>> Why bother with all that when most people can't demonstrate
>> use beyond what a $50 old used ST or Amiga could get them?
>>
>> What's the point of a 'pentium' version of WordWriter?
>
>
>--
>Spam Avoidance: Remove "dont_spam_me_" in my Email address to reply.
>My website is at http://www.serv.net/~fewtch/index.html - something for everyone
>there, take a look if you have the time! My tribute to the dearly loved cat of
>19 years I recently lost is at http://www.serv.net/~fewtch/cats/index.html.
jedi
>In comp.unix.advocacy jedi <je...@dementia.mishnet> wrote:
>: ...
>: Before M$ and Intel brought us the blight that was the
>: PC, it's competitors were more than capapble of delivering
>: systems (even then in 1981) of similar capacity of those today.
>:
>: The i86's were just playing catch up to the 68k's for a
>: considerable amount of time.
>
>You mean Intel caught up?
If you're not using M$ OSes, a 386 ~ a fast 68000.
Although the 80x86's are still incredibly annoying
as a cpu family go...
>
>I've spent many years programming in assembly language and have always
>disliked Intel's "architectural" values from the 386 back through
>the 8086, the 8080, the 8008, the 4040, and the 4004.
>
>Intel likes special purpose everything. IIRC in every Intel processor
>I programmed (those listed above) up through the 80286, EVERY SINGLE
>REGISTER was unique. That is for any given register, there was
>some operation you could perform ONLY using that register.
>
>The PDP11, the IBM 360 etc, the Dec VAX, the Motorola 680x0 processors,
>along with some others I've used, all had reasonably general registers,
>although most of them split the register set in half between data
>and address registers.
>
>The thing I hated most though was the 8088/8086/80x86 segmented memory
>model. <shudder> <yuck!> The best thing they introduced (in the 80386)
>as a flat memory space like normal computers.
>
jedi
>On Tue, 7 Jul 1998 12:11:43 -0700, je...@dementia.mishnet (jedi) wrote:
>
>> Why bother with all that when most people can't demonstrate
>> use beyond what a $50 old used ST or Amiga could get them?
>
>(word processing, data entry, light spreadsheet work). Try any kind
>of serious work with digital audio or image processing on that $50 ST
>or Amiga, and let me know how far you get... How about mastering a
>writable CD on that $50 Atari ST? Or working with a 1024x768 24-bit
>color image? Or even browsing a web site with a few small Java
>applets or one or two pictures? If you think "most people" only use
>their computers for word processing and data entry these days, you're
>still living in the dark ages.
If you think the bulk of Jon Doe is working with MegaPixel
truecolor images, you're giving Jon too much credit. Although,
as a mastering system a Falcon may yet still run circles around
your PC (optimization). Then there's Java: gauranteed to cripple
most people's net pipes and bring the dominant browsers down as
well. While the ST isn't up to MegaPixel webbrowsing, some of
it's successors are as are most of the subsequent Ameobas.
>
>Oh, I forgot to mention games... they've gotten pretty amazing lately
>thanks to the latest hardware (in fact, computer gaming is one of the
>primary driving forces behind hardware improvements, maybe even more
>of a driving force than MS's increasingly larger apps and OS's).
Yup... once lemming users managed to get over 'playing
games' on their PC's they managed to give birth to the
3K game machine.
Although it is at least SOMETHING to soak up those
mips with.
Personally, I'm still looking for a suitable replacement
for my commie Epxy 500XJ. I've found them for consoles
but not PC's. So much for bully markets...
Toon Moene
> On Mon, 6 Jul 1998, jedi wrote:
> > You've got to be kidding? Emacs is one of the bloatware KINGS.
> But it still works on an 8MB 486 - Word left that hardware behind how many
> years ago?
One learns nice things about Linux almost every day. Today I talked to a
colleague of mine about the newest plans to give every employee at my
Institute the opportunity to save 1500 Dutch Guilders (at an exchange rate of
1 DG <-> 0.50 $) on the acquisition of a PC.
Because he's already using Linux since 93?, 94, I asked him whether the
Institute-approved PC's would be able to run Linux.
Of course, he replied - in the beginning it was hard, because Linux used much
more memory than MSDOS, but now with Windows 95, that wasn't much of a
problem.
I didn't get this at first, because I was mostly interested in the answer to
the question whether the "Institute approved PC" would be comply to the
Hardware Compatibility List :-)
--
Toon Moene (mailto:to...@moene.indiv.nluug.nl)
Saturnushof 14, 3738 XG Maartensdijk, The Netherlands
Phone: +31 346 214290; Fax: +31 346 214286
g77 Support: mailto:for...@gnu.org; NWP: http://www.knmi.nl/hirlam
Christopher Browne
>On 7 Jul 1998, John Fieber wrote:
>[Snip]
>> I don't think any programmer should be unleashed on a modern PC until
>> they have put in a year or two in a nice confined environment--say a
>> Commodore 64--where they can learn the skills of simplicity and
>> efficiency in programming.
>
>the "hand written" listing - I couldn't afford a printer). The
>development tools were on two floppies - so I had to keep swapping disks
>all the time. It took 20 minutes to compile and link my 1000 line
>program. I had to give it up when the executable produced got so big that
>there wasn't any memory left for it to be able to do anything.
And now, you probably have more *cache* on your motherboard than there
was memory and disk on the C=64.
The increased resources have the substantial benefit that many things
that could not before be done now can be. Unfortunately, it also means
that we have programmers that have no concept of how algorithms work and
how resources are consumed because the environment so hides it from
them.
There are some useful things that can be done in 64K; I wouldn't go so
far as to force people into using dramatically obsolescent equipment.
It would be more appropriate to have people do some work on more
recent "limited" platforms.
Writing software for PalmOS (3Com PalmPilot) would be a reasonable case
in point. It has some visible resource limitations, but is nonetheless
"big" enough that it is visibly useful.
Actually, I suspect that there would be some real merit to building
UNIX-based tools to manipulate PalmOS "databases," and making those
tools "open source." It may not be a "full, relational" database system,
but it there's certainly enough there to support some diverse PalmPilot
applications, which suggests that it's powerful enough to support
reasonably interesting applications atop a UNIX.
--
"Problem solving under linux has never been the circus that it is under
AIX." (By Pete Ehlke in comp.unix.aix)
cbbr...@ntlug.org- <http//www.hex.net/~cbbrowne/lsf.html>
Arthur Corliss
>In article <35a2672d...@news.serv.net>,
> fewtch@dont_spam_me_serv.net (Fewtch) writes:
>
>> even products like MS Office aren't really bloated on a machine with
>> 64 megs RAM and a 7 gig hard drive. It's all in how you look at it.
>
>not resources consumed in relation to resources available. Measuring
>bloat in terms of resources available only serves to power the insane
>upgrade treadmill.
>
>they have put in a year or two in a nice confined environment--say a
>Commodore 64--where they can learn the skills of simplicity and
>efficiency in programming.
>
>a state of crisis.
>
>
>-john
Well said, and my sentiments exactly. My first home computer had 2K, and I
remember paying close attention to compact and efficient coding. We've gotten
spoilt by the capabilities of modern machines, but not everyone has forgotten.
:-) Check out what some of our counterparts are doing with 32K on a Z80,
sitting in a T-85 calculator, for instance.
--Arthur Corliss
"Live Free or Die--the Only Way to Live" (NH State Motto)
jedi
>On Tue, 7 Jul 1998 14:33:54 -0700, je...@dementia.mishnet (jedi) wrote:
>
>>On 7 Jul 1998 20:54:17 GMT, Dave Kenny <d...@parka.winternet.com> wrote:
>>>In comp.unix.advocacy jedi <je...@dementia.mishnet> wrote:
>>>: ...
>>>: Before M$ and Intel brought us the blight that was the
>>>: PC, it's competitors were more than capapble of delivering
>>>: systems (even then in 1981) of similar capacity of those today.
>>>:
>>>: The i86's were just playing catch up to the 68k's for a
>>>: considerable amount of time.
>>>
>>>You mean Intel caught up?
>>
>> If you're not using M$ OSes, a 386 ~ a fast 68000.
>> Although the 80x86's are still incredibly annoying
>> as a cpu family go...
>
Actually, your comments support my claim in more
detail than I myself was willing to present.
>two cycles vs four on the 68000, and the 386 was pipelined to
>boot. At 16MHz, the 386 supported a maximum bus bandwidth
>of 32MB/sec, vs 8MB/sec for the 68000 (4 bytes every 2 clocks
>instead of 2 bytes every 4 clocks.) The 68030 on the other
>hand offered these features and more, but alas, the general
>purpose 680x0 family is effectively dead. Some popular
>derivatives exist like the 68302 and 68360, but even these
>are candidates for replacement with MPC860s (PowerPCs).
>
>>>
>>>I've spent many years programming in assembly language and have always
>>>disliked Intel's "architectural" values from the 386 back through
>>>the 8086, the 8080, the 8008, the 4040, and the 4004.
>>>
>>>Intel likes special purpose everything. IIRC in every Intel processor
>>>I programmed (those listed above) up through the 80286, EVERY SINGLE
>>>REGISTER was unique. That is for any given register, there was
>>>some operation you could perform ONLY using that register.
>>>
>>>The PDP11, the IBM 360 etc, the Dec VAX, the Motorola 680x0 processors,
>>>along with some others I've used, all had reasonably general registers,
>>>although most of them split the register set in half between data
>>>and address registers.
>>>
>>>The thing I hated most though was the 8088/8086/80x86 segmented memory
>>>model. <shudder> <yuck!> The best thing they introduced (in the 80386)
>>>as a flat memory space like normal computers.
>
>If only you knew Multics and its children from the 64-bit mainframe
>world. Paged segmented memory subsystems offered some unique advantages
>which paged only systems have yet to match.
>
>>--
>>Hardly. Microsoft has brought the microcomputer OS to
>>the point where it is more bloated than even OSes from |||
>>what was previously larger classes of machines altogether. / | \
>>This is perhaps Bill's single greatest accomplishment.
>
>-Steve
r.e.b...@usa.net
ben...@ameritechREMOVETHIS.net wrote:
> jedi wrote:
> > On Sun, 05 Jul 1998 06:17:31 GMT, BR <ben...@mailhost.ind.ameritech.net>
wrote:
> > >Bloody Viking wrote:
> > ><small snippage>
> > >
> > >> The "insane" computing power of modern PCs means that anyone with a
credit
> > >> card can buy a computer with the power of the earliest Cray
can
> > >> have a cray! Crays use UNIX too.
Actually, a BeoWulf Linux cluster can outperform all but about 300 of the
worlds fastest super computers. It takes a bit of doing (not for junior
programmers), but you too can have a 7 gigaflop cluster.
> > >Just a small interjection here. If it wasn't for the duo poly known as
> > >MS and Intel. The consumer probably wouldn't have as much power as they
> > >have in such a short time frame.
Microsoft's only "REAL" contribution was a rip-off of DEC BASIC which was
ported to the Intel 8080 based Altair. Under today's laws, Gates would have
been doing time (the copyright laws were changed in 1976, prior to Gate's
"creation" of Micro-Soft BASIC.
MS-DOS, was intentionally brain-dead to keep Gary Kildall from releasing
MP/M-86, which was a direct threat to the IBM Series I and System 3X markets.
Intel used everything it could to keep the 8086 away from Zilog, whose Z-8000
processor wasn't limited by that silly little 1 megabyte memory limit.
While Microsoft touted it's PC-DOS/MS-DOS 2.0 file-system as "unix-like"
(8.3 filenames and all :-), the REAL UNIX had found it's way to Zilogs,
several 68k machines (including SUN), and several "industrial strength"
machines. IBM tried to protect it's mainframe market from UNIX while
AT&T tried to take back control from BSD. Meanwhile, Richard Stallman
was quietly formalizing his General Public License and promoting what is
now called "Open Source Software".
The UNIX boys were having a field day with TCP/IP, and Bill Joy (now of Sun)
was getting a good feel for what would eventually be come to be known as the
Internet. All this time, Microsoft was at "arm's length" with Novell,
and IBM had it's own ideas of a "Lan Manager" network.
When the Mac came out, I went into the computer room to look at a Sun 1.
At $30,000 per workstation, it was a bit steep for my pockets, but it didn't
take much imagination to see that THIS was the system of the 21st century.
It had graphics, animations, realtime data charting, 3D modelling, and could
do it ALL at once. Bill Gates must have looked at one of these babies and
said "that's what I want to be when I grow up". Silly Bill, sold the rights
to Xenix (and the UNIX market) to SCO so that he could buy into the Mac.
While MS-DOS users struggled to juggle 5-6 TSRs, the UNIX systems of the
day were gracefully handling 100 users on a system with about the same power
as a PC-XT (A PDP/11-77 with 1 meg ram and 4 300 meg drives).
By the time Corporate America actually began to trust PCs, UNIX was running
the telephone systems, the railroads, military systems, and hundreds of
"mission critical" systems. UNIX had 1 billion "users" who didn't even
know they were users.
By the time Windows 3.0 came out, Sun was switching from its legacy SunView
system to the X11R4 system, complete with source code compatible Xview
toolkit. The IPC lunchbox was about $8000, about $1000 more that a PC capable
of running Windows 3.0, WordPerfect, Lotus Notes, and Harvard Graphics. Only
$400 more than the cost of a PC running Win 3.0, Word, Excel, PowerPoint, and
Project Manager. Of course, the PC networking left a great deal to be
desired, PCs made great clients, but the IPCs could also be servers.
By the time Windows 3.1 came out, it was time to upgrade hardware, while
the IPC chugged along reliably through several SunOS upgrades. By the time
Windows 3.11 came out, and Office came out as a package, users were being
upgraded again. Incompatibilities with previous versions of Word often meant
"shock upgrades" for hundreds of users. Not real popular from the customer's
point of view. Meanwhile, Sun was offering the Sparc5 ond Sparc20 that ran
even faster. They charged a bit more for each machine, but the new machines
would efficiently interoperate with the old ones.
In 1992 several versions of UNIX for Intel were available. They were
expensive, upwards of $3000 each. A group of people got hold of Linus
Torvald's Linux and took up the challenge of creating an entire UNIX system
that could be sold for less than $100. In 1993, SLS linux was selling for
$98 (plus postage).
In 1993, Microsoft decided to take on AOL by creating it's own internet
service. Originally, it was going to be based on NT 3.51, NETBEUI DLC, and
Microsoft "BackOFFICE". The goal was to counter Lotus Notes, and
simultaneously take on AOL. While all of this was going on, I was working
with 3000 publishers and 2000 BBS operators, coaching them in how to create
terminal servers and web sites with Linux. They had no budgets, they
scrounged hardware from storage closets, and even set up the web servers on
personal phone lines. Dow Jones (where I worked) had a site based on a
SparcStation 20, attached to a T1. We went from 1,000 hits/day to 1000
hits/minute in less than 9 months. The publications on the
"online-newspapers" mailing list were also growing. Microsoft quickly
changed it's strategy, and caught the "last train" - it was actually a
late-comer. By the time Microsoft had come aboard, the Internet had grown to
40 million users and was growing at 20%/month.
> > Before M$ and Intel brought us the blight that was the
> > PC, it's competitors were more than capapble of delivering
> > systems (even then in 1981) of similar capacity of those today.
Microsoft wasn't a leader, it was a follower. It snatched up ideas from
the UNIX community and butchered them to fit into that 1 meg memory model
(that was Gates' idea). Even the graphics were primative compared to
those of UNIX machines connected to Tektronics graphics stations.
Microsoft's big "innovation" was a licensing policy that dictated an
"all or nothing" policy that gave OEMs deep "discounts" off of rediculously
inflated prices ("perceived value pricing"), in exchange for giving Microsoft
complete control of the system and an exclusive market.
Many upstarts initiated by offering alternatives. Dell offered SCO UNIX,
as did Micron. Gates, who holds a 25% stake in SCO, could see when these
companies were "ripe", and would "convert" them to Microsoft companies.
> > The i86's were just playing catch up to the 68k's for a
> > considerable amount of time.
i86's were playing catch up to EVERYTHING. The 68k, the Z8k, the Sparc,
the Fairchild Clipper, the AMD 19k (i'm not sure of these chip numbers).
In many cases, Intel cut deals with competitors - AMD for example, got
80x86 technology in exchange for not pushing it's chip as a PC-Priced
UNIX chip.
The Atari ST series, the Amiga, and big Macs were all delayed in FCC
certification when IBM and other PC makers "slipped in" about 500
permutations that needed to be independently certified.
> > Hardly. Microsoft has brought the microcomputer OS to
> > the point where it is more bloated than even OSes from |||
> > what was previously larger classes of machines altogether. / | \
> > This is perhaps Bill's single greatest accomplishment.
It's amazing isn't it? The system required to run NT 4.0 with
Microsoft Office is more computer power than most mainframes of
1993. A typical NT 4.0 system runs 1 meg of CACHE, 128 Meg of Ram,
4-6 gig of hard drives, 32x CD-ROM, and NO BACKUP SYSTEM! What a
wonderful configuration for a system that crashes once a week.
In 1994, a 3090/400 mainframe ran 4 25 mips processors, 16 meg of ram,
30 gig of DASD for 10,000 users, and took 6 hours to back up every night.
Just in case this was the one time per year that it might actually
require an unscheduled reboot.
In 1996, a Linux system running on a 486/100 could support a full T1
link, run 100 concurrent transactions, and could handle 300-500 simultaneous
processes. It was an "event" if it actually had to be rebooted before it's
scheduled quarterly upgrade. With a pair of servers, the users would never
know one of the machines had gone down. Some folks just liked their systems
so much that they'd upgrade them only once/year "just for grins".
> I'm not arguing your above points. My point was,for good or bad.
> Feature-itus and bloat,put us on the upgrade path.
But who was really steering the boat? The engine driving the change
was the UNIX/Internet community. Corporations didn't stop buying
UNIX machines, but they didn't have to throw them away either. In 1995,
McGraw-Hill was still using a Power 6/32 "Tahoe" machine that was
made in 1985. We finally pulled the plug because it was taking up too
much space in the computer room. It was still functional and useful
as a mail system and as a conferencing system. We also used it as
an NFS server. The box had 128k cache, 16 meg ram (as 1meg/128 bits)
and clocked at 25 mips (same as a 486/50).
> Do you think that we
> would have pentiums if it wasn't for the above? Now all a person has to
> do is run more efficient software on that pentium,viola instant cray. :)
And where was that software developed? On Vaxen? on 68K, on Z8K, on
Sparc, on Mips, on PowerPC, on PA-RISC... And finally, the 386 could
run a crippled version (no cache, no multiple pipelines...). The irony
is that the Intel chips usually cost 4 times the price of their competitors.
Even today, a K6-300 runs $100-150 compared to $400 for a P-II/400 (equivalent
performance). The MIPS R10k runs about $50 in quantities. If you focus
expandability on SCSI bus, the new serial expansion systems (Firewire,...)
You could build a full-blown Linux system running a 1 Gigaflop/s machine
(equivalent to P-II/600) for under $2500. The manufacturers cost would be
less than $700.
Rex Ballard - http://www.access.digex.net/~rballard
-----== Posted via Deja News, The Leader in Internet Discussion ==-----
http://www.dejanews.com/rg_mkgrp.xp Create Your Own Free Member Forum
robert w hall
<fewtch@dont_spam_me_serv.net> writes
>
>I know this post was tongue-in-cheek, but look at it this way: CP/M
>2.2 is pretty much featureless as far as OS's go. Does it multitask?
>Multithread? Support any networking whatsoever? Etc, etc :-)
But in another part of the forest, some of us happily ran OS9 (in my
case Level 2 on a 6809 with 256k) happily for nearly 10 years whilst the
PC world caught us up! I only gave it up when I blew the chip! It ran a
plausible wordprocessor (Stylograph) and C, and multi-tasked, and that
was all I needed to be happy.
Only the prospect of creating a cheap 'work-station' environment via
Linux/FreeBSD has moved me on.
>
>I'm not sure of the size of the MS-DOS 5.0 kernel (io.sys +
>msdos.sys)... anyone here have any idea? Around 45-50k or so? Then
>again, one might argue that MS-DOS doesn't even qualify for OS status,
>since it really doesn't control any hardware whatsoever and lets the
>BIOS handle just about everything... unless you'd consider himem.sys
>and EMM386.EXE to be "part of the kernel" ;-)
Heavens, how profligate - the guts of OS9L2 was two 8k modules!
[snip]
--
robert w hall
Richard Lyon
jedi wrote in message ...
> ...and this is just how relevant to John Doe?
>
I am John Doe. After using FBSD since 386BSD 0.1 I have yet
to link a new kernel. I just use it because it works and allows me
to concentrate on whats important. ie. Joe Doe (me) runs the following
applications:
emacs
egcs cross-compiler
octave
gnuplot
Xfig
LaTeX/TeX
This allow me to complete my work and publish it. Try generating new
fonts for TeX on a slow machine or running a complex octave script.
It sure nice also to have a high resolution display for gnuplot and xdvi.
It was a real pain trying to do these things on a 68K box.
As soon as I get my next grant I'm off to get the fastest Pentium II with
at least 128M of RAM. Also, I getting a bit fed up with my HP
Laserjet 6L. I want it faster with higher resolution and colour.
Just think, mathematical equations in colour. This would really show
those windows junkies who the boss is ..... Yeah, if I get a 25 GB
drive then I can load every BSD and Linux distribution known to
man. Doesn't this just sound like John Doe?
Rick Kelly
> Ben Sandler wrote:
>>
>> I have no experience with *BSD whatsoever (I got a BSDi account at work
>> two weeks ago, but I don't really use it). Is there binary
>> compatibility in between Linux and *BSD? If so, that certainly argues
>> your point.
> --
The Linux emulation for NetBSD is fairly good. I've had Linux Doom I
binaries running on NetBSD 1.0A, 1.2, and 1.3.x. Emulation on NetBSD
and FreeBSD (don't know about OpenBSD) use the shared libraries, etc
from Linux. It's a quite fast and useful emulation.
--
Rick Kelly r...@tencats.com r...@rmkhome.com
Stephen E. Halpin
>On 7 Jul 1998 20:54:17 GMT, Dave Kenny <d...@parka.winternet.com> wrote:
>>In comp.unix.advocacy jedi <je...@dementia.mishnet> wrote:
>>: ...
>>: Before M$ and Intel brought us the blight that was the
>>: PC, it's competitors were more than capapble of delivering
>>: systems (even then in 1981) of similar capacity of those today.
>>:
>>: The i86's were just playing catch up to the 68k's for a
>>: considerable amount of time.
>>
>>You mean Intel caught up?
>
> If you're not using M$ OSes, a 386 ~ a fast 68000.
> Although the 80x86's are still incredibly annoying
> as a cpu family go...
On what do you base this? The 386 ALU could turn a result in
>Hardly. Microsoft has brought the microcomputer OS to
>the point where it is more bloated than even OSes from |||
>what was previously larger classes of machines altogether. / | \
>This is perhaps Bill's single greatest accomplishment.
-Steve
BR
Or a hp48gx calculator.
--
************************
* Enjoy the pane-Run NT*
************************
Patrick M. Hausen
> The term "bloatware" is very subjective, especially these days. For
> example, the "average" new computer sold these days has a 7.2 gig hard
> drive, 48-64 megs RAM and a 300 MHz processor. On a machine like
> that, what would be considered "bloatware" on a Pentium 120 with a 1.2
> gig hard drive and 16 megs RAM is pretty lightweight. In my opinion,
> even products like MS Office aren't really bloated on a machine with
> 64 megs RAM and a 7 gig hard drive. It's all in how you look at it.
Despite the fact, that e.g. Clarisworks/Appleworks, a nicely integrated
office suite fits into 5 Meg of RAM an can be run on an 8 MB 68040
Macintosh when you stick to system 7.6.
> So what's wrong with
> running heavyweight software on heavyweight hardware?
Why the need for heavyweight hardware?
I've been looking for a decent laptop for years now.
It doesn't exist. What do I consider decent? Well, a 486 with
16 to 32 MB and a 1 GB harddrive. No audio, no CD but up to
10 hours of battery life at a price of < $ 2K.
It would be possible to make such a thing. It would be the
perfect FreeBSD laptop IMHO. But what do I get? Pentium II
chips, that drain your battery quicker than it takes to drive
from my home to my parents' by train. (3 hours).
*sigh*
Patrick
Jordan K. Hubbard
> ...
I find it somewhat telling that an entire analysis of the industry can
be presented in a BSD group which somehow makes it sound as if Linux
was the only player in all of this and is the only player now. Hello?
Is there anyone in there? What is it about some Linux advocates that
compells them to completely *ignore* any of the alternatives?
For something which attempts to give a historical perspective on all
this, it's doubly unforgivable since BSD has been around for a lot
LONGER than Linux and, even though it wasn't allowed to be a player in
the free software world until relatively recently (1991), it was very
serious player in the academic world for some time before that.
I am always willing to acknowledge the accomplishments of Linux, I
just wish it were something more often reciprocated. Are the afraid
of something? :-) [At which point here I fully expect Albert D. Calahan
to jump in with some bad conspiracy theory or claims that gosh, the
Linux folks just don't really even notice us. :)]
- Jordan
Alexander Viro
Jordan K. Hubbard <j...@time.cdrom.com> wrote:
>r.e.b...@usa.net writes:
>> ...
>
>I find it somewhat telling that an entire analysis of the industry can
>be presented in a BSD group which somehow makes it sound as if Linux
>was the only player in all of this and is the only player now. Hello?
>Is there anyone in there? What is it about some Linux advocates that
>compells them to completely *ignore* any of the alternatives?
Jordan, it's Rex. Please, take a *big* barf-bag and look through his stuff
on dejanews. Then read, erm, CV on his homepage. Then draw your conclusions.
The kindest description that I can give: suit par excellence, spindoctor,
ignorant, absolutely dishonest (can't tell truth from lie). His knowledge
of Linux is about the same as his knowledge of BSD or UNIX (zero). 100%
buzzword-compliant. Author of wonderful neologism: 'hacker in suit'
(absolutely seriously used it as self-description). Unusual type in the
advocacy zoo (probably not unique, but I didn't see anything similar
around Linux or FreeBSD).
BTW, I'm serious about going to DN and his webpage. It's
fascinating (albeit long) reading.
HAND,
Al
--
Luser, n.:
Human-like creature that doesn't dare to use elevator, because of
its belief that only horrible geeks can master arcane and obscure art of
using control panel.
kskei...@my-dejanews.com
j...@time.cdrom.com (Jordan K. Hubbard) wrote:
> r.e.b...@usa.net writes:
> > ...
>
> I find it somewhat telling that an entire analysis of the industry can
> be presented in a BSD group which somehow makes it sound as if Linux
> was the only player in all of this and is the only player now. Hello?
> Is there anyone in there? What is it about some Linux advocates that
> compells them to completely *ignore* any of the alternatives?
>
> this, it's doubly unforgivable since BSD has been around for a lot
> LONGER than Linux and, even though it wasn't allowed to be a player in
> the free software world until relatively recently (1991), it was very
> serious player in the academic world for some time before that.
>
> I am always willing to acknowledge the accomplishments of Linux, I
> just wish it were something more often reciprocated. Are the afraid
> of something? :-) [At which point here I fully expect Albert D. Calahan
> to jump in with some bad conspiracy theory or claims that gosh, the
> Linux folks just don't really even notice us. :)]
It's interesting that it's not uncommon to hear about Linux users who switch
to (Free)BSD, but you never ever hear about BSD users switching to Linux.
r.e.b...@usa.net
j...@time.cdrom.com (Jordan K. Hubbard) wrote:
> r.e.b...@usa.net writes:
> > ...
>
> I find it somewhat telling that an entire analysis of the industry can
> be presented in a BSD group which somehow makes it sound as if Linux
> was the only player in all of this and is the only player now. Hello?
> Is there anyone in there? What is it about some Linux advocates that
> compells them to completely *ignore* any of the alternatives?
Let's look at that. I use a number of different indicators including
web site counts, bookstore title counts, and several other metrics.
BSD is a presence. BSD was hindered by some of the AT&T code and
this kept it from being able to be distributed freely. Eventually,
FreeBSD did come out in a freely distributable form.
Linux is just in momentum right now. Lots of users are playing with
Linux because they can go to Borders or Barnes and Nobles, by a $50
book, or go to CompUSA and by a $50 package that will let them run
UNIX (Linux) on their cheap little PC.
Linux is a pretty decent version of Linux, and lot's of people have
contributed lots of code to the "Linux Distribution". Of course, most
of this could just as easily be compiled to run on FreeBSD.
Linux also has several substantial companies who are willing to
"Give away CD-ROMs and Sell Support". This would be like BSDi
or SCO saying "We'll let you have the software we used to charge
$2000/copy for, and you can have the source. If you get in trouble,
we'll charge you $200/user/year.
Caldera, Red Hat, and Workgroup Solutions are going for volume,
with the clear understanding that demand for service will increase
as more people use Linux.
The BSD/AT&T feuds have been plaguing the UNIX industry for years.
Linux shows up as an impartial 3rd party player, gives away the store,
and charges reasonable prices. Eventually, other UNICES will be
striving to be "Linux Compatible" (not a difficult task at all - since
the source is right there).
> For something which attempts to give a historical perspective on all
> this, it's doubly unforgivable since BSD has been around for a lot
> LONGER than Linux and, even though it wasn't allowed to be a player in
> the free software world until relatively recently (1991), it was very
> serious player in the academic world for some time before that.
Actually, BSD extensions existed in SunOS, SysVr4, AIX, HP-UX, Ultrix,
and several other systems. BSD was the "standard", by which other
versions of UNIX were measured. Linux has usurped that position
by "giving away the store".
> I am always willing to acknowledge the accomplishments of Linux, I
> just wish it were something more often reciprocated. Are the afraid
> of something? :-) [At which point here I fully expect Albert D. Calahan
> to jump in with some bad conspiracy theory or claims that gosh, the
> Linux folks just don't really even notice us. :)]
Actually, the Linux folks are very aware of BSD, and the distinctions
of the BSD license vs. the GPL and GPLL. Since the standardization of
UNIX depends on the agreement of hundreds of vendors, commercial customers,
programmers, support staff, and power-users, the GPL has provided a much
higher level of trust, and as a result, has made it much easier to get
that agreement.
Actually, it would be great of the BSD folks wanted to "come and play"
with us. Certainly, it's not difficult to add the applications that
would make BSD "Linux Compatible".
> - Jordan
User Rdkeys Robert D. Keys
> I know this post was tongue-in-cheek, but look at it this way: CP/M
> 2.2 is pretty much featureless as far as OS's go. Does it multitask?
> Multithread? Support any networking whatsoever? Etc, etc :-)
No, but as a single user OS, or a dumb datalogger type of application
it still works fine. It depends a lot upon the tools you need to use.
Dos still works fine in that mode. Why have the overhead of a gig of
excess windows for small discrete applications. It makes great waste
of resources and space.
> I'm not sure of the size of the MS-DOS 5.0 kernel (io.sys +
> msdos.sys)... anyone here have any idea? Around 45-50k or so? Then
> again, one might argue that MS-DOS doesn't even qualify for OS status,
> since it really doesn't control any hardware whatsoever and lets the
> BIOS handle just about everything... unless you'd consider himem.sys
> and EMM386.EXE to be "part of the kernel" ;-)
The dos kernels run to around 75K max, if memory serves me correctly.
There is nothing inherently wrong in the bios handling things. It can
simplify portability between hardware. On dos, I never use beyond
640K except as ramdisk. It makes the stilly thing fly. A 40mhz
386 running dos with unix tools on a ramdisk flys as fast as many
things can go diskbound on pentiums. If the tools satisfy the computing
needs, then who cares. (It might be fun to run a stripped pentium with
dos and a good set of unix tools, for comparison, though.) It might
even be faster than a comparable unix on the same box. But, at those
kinds of speeds, it is mostly irrelevant, except for specialized needs.
If you are into number crunching it could be important. Most everything
else crawls still.
I still run dos 5 with a set of unix tools (about 50 tools) and a
TeX and the real troff. Using a ramdisk as the I/O, it runs as fast
as the lesser Pentiums that are diskbound, and the silly thing is a
16mhz PS/2. Considering the power of the pentium... how many folks
ever actually use the horsepower they have in the chip. My expectation
is that most of it goes unused, but it sure is a good sales gimmick
with all those flashy games and webscrapers they can bring up.
It really depends upon what tools you need. If you don't need gui's
on a machine, then it can flat fly with a goodly set of unix tools,
even on dos. Dos without the unix tools is rather underpowered.
The tools I need for writing are limited by the input speed that I
can muster. Hence, for that kind of specific application, even a
CP/M box or dos box works fine. Alas, fitting a troff on CP/M is
a bust, but there are some good nroffs for it. I am still looking
for the ellusive vi for CP/M. Someday, I may port it for fun.
Considering the bloat, I can fit a stripped dos with all the unix tools,
and TeX and troff on about 7 floppies. How many floppies would
win98 take if it could fit on floppies....? Egads! That is why you
need the silly Pentium.
Good discussion....(:+}}...
RDK
Marcus Meissner
void <fl...@interport.net> wrote:
>On Tue, 07 Jul 1998 01:32:57 +0000, Ben Sandler <s...@the.sig> wrote:
>>
>>I have no experience with *BSD whatsoever (I got a BSDi account at work
>>two weeks ago, but I don't really use it). Is there binary
>>compatibility in between Linux and *BSD? If so, that certainly argues
>>your point.
>
>don't believe Linux runs BSD binaries, however.
Really? For instance, how is the 'clone' systemcall emulated?
Marcus
--
<URL:http://wwwcip.informatik.uni-erlangen.de/~msmeissn/>
jedi
>In comp.unix.advocacy Fewtch <fewtch@dont_spam_me_serv.net> wrote:
>
>> I know this post was tongue-in-cheek, but look at it this way: CP/M
>> 2.2 is pretty much featureless as far as OS's go. Does it multitask?
>> Multithread? Support any networking whatsoever? Etc, etc :-)
>
>No, but as a single user OS, or a dumb datalogger type of application
>it still works fine. It depends a lot upon the tools you need to use.
>Dos still works fine in that mode. Why have the overhead of a gig of
DOS does NOT still work fine in that mode. It's memory
management is attrocious and requires far more lower
level knowledge to deal with than even a Unix. A 386
DOS could be useful for some however...
[still holding a grudge against M$ for the state of
DOS (and thus Win16) in '94]
Jordan K. Hubbard
> Let's look at that. I use a number of different indicators including
> web site counts, bookstore title counts, and several other metrics.
> BSD is a presence. BSD was hindered by some of the AT&T code and
> this kept it from being able to be distributed freely. Eventually,
> FreeBSD did come out in a freely distributable form.
Huh? I don't see how these two points can be mated at all. Yes, *BSD
had some trouble with AT&T and friends in the beginning, but that's
all behind us now and your "web site counts and other metrics" have
probably not even attempted to gather a full picture of just how
pervasive BSD technology is in the marketplace. Ever hear of a couple
of small companies like Yahoo, Oracle or U.S. West, for example? They
all use FreeBSD technology (and, in the case of Oracle, NetBSD
technology as well) and have greatly helped to spread the message that
what we've got to offer is also pretty substantial.
Just because FreeBSD books aren't falling onto your head at the
bookstore (and that's something we're actively working on changing
now) by no means indicates that the word isn't getting out to the
front-line techs and other folks who "make things happen" at some of
the nation's largest ISPs. We may not have the sheer quantity of
users that Linux has yet, but we're well past the 1 million mark and
also have some significant *quality* users out there which we can
point to with justifable pride. I think your selection of metrics is
far too narrow.
> Linux is just in momentum right now. Lots of users are playing with
> Linux because they can go to Borders or Barnes and Nobles, by a $50
> book, or go to CompUSA and by a $50 package that will let them run
> UNIX (Linux) on their cheap little PC.
They can do the same now for FreeBSD in an increasing number of
stores. Walnut Creek CDROM is selling Greg Lehey's "The complete
FreeBSD" book + 4 CD set at Borders, Fry's, and a number of other
bookstore chains. Once I get my book for Addison Wesley finished, I
expect to see even greater penetration of the same market.
> Linux is a pretty decent version of Linux, and lot's of people have
> contributed lots of code to the "Linux Distribution". Of course, most
> of this could just as easily be compiled to run on FreeBSD.
I think it'd be pretty hard for Linux to be a version of anything else. :)
I'm really not sure just what you're trying to say in that paragraph
but there's also no "Linux distribution" that I know of, just a lot of
independent distros like RedHat, Slackware, Debian, Stampede,
Yggdrasil, OpenLinux, ummm. Did I leave anyone out? :)
> Caldera, Red Hat, and Workgroup Solutions are going for volume,
> with the clear understanding that demand for service will increase
> as more people use Linux.
Yep, us too. You'd probably be amazed to know just how many thousands
of FreeBSD CDs we sell each month, but I'm not allowed to disclose
sales figures on the net. I suspect that RH and the others aren't
allowed to either, so no big loss.
> The BSD/AT&T feuds have been plaguing the UNIX industry for years.
I think you're suffering from an attack of retro here. Those "feuds"
have been over for years and are essentially irrelevant now.
> Linux shows up as an impartial 3rd party player, gives away the store,
> and charges reasonable prices. Eventually, other UNICES will be
We do the same. Why, again, do you continue to see this as a property
which is somehow unique to Linux? I can honestly only conclude that
you haven't really made any attempt to familiarize yourself with the
*BSD world at all if you can make statements like that with a straight
face. Price, performance, you name it and we're right there in the
same ballpalk.
> > For something which attempts to give a historical perspective on all
> > this, it's doubly unforgivable since BSD has been around for a lot
> > LONGER than Linux and, even though it wasn't allowed to be a player in
> > the free software world until relatively recently (1991), it was very
> > serious player in the academic world for some time before that.
>
> Actually, BSD extensions existed in SunOS, SysVr4, AIX, HP-UX, Ultrix,
> and several other systems. BSD was the "standard", by which other
> versions of UNIX were measured. Linux has usurped that position
> by "giving away the store".
And you repeat your error again here. If you are somehow under the
impression that BSDI with their commercial product is the only
"player" in the BSD world then I can only repeat my earlier assertion
about needing to learn more about all this. It would really be nice
if you and other Linux advocates could research both "sides" of this
argument in a little more detail since I have, at least, made a more
than reasonable effort to come up to speed on what the Linux folks are
doing. I even go to their conferences occasionally.
> Actually, the Linux folks are very aware of BSD, and the distinctions
> of the BSD license vs. the GPL and GPLL. Since the standardization of
> UNIX depends on the agreement of hundreds of vendors, commercial customers,
> programmers, support staff, and power-users, the GPL has provided a much
> higher level of trust, and as a result, has made it much easier to get
> that agreement.
It has? Boy, that's a chain of logic that Socrates would have some
fun with if he were alive today. ;-)
--
- Jordan Hubbard
Co-founder/Release Manager, The FreeBSD Project
Walnut Creek CDROM
void
>r.e.b...@usa.net writes:
>
>> Let's look at that. I use a number of different indicators including
>> web site counts, bookstore title counts, and several other metrics.
>> BSD is a presence. BSD was hindered by some of the AT&T code and
>> this kept it from being able to be distributed freely. Eventually,
>> FreeBSD did come out in a freely distributable form.
>
>Huh? I don't see how these two points can be mated at all.
You don't think BSD would be more popular if it hadn't had those licensing
problems? I thought that was common wisdom, more or less. Perhaps I'm
misconstruing REB's point.
ISTR reading that Linus Torvalds said that he wouldn't have written Linux
if 386/BSD had been available at that point.
>Once I get my book for Addison Wesley finished, I
>expect to see even greater penetration of the same market.
I didn't know you had something in the works at AWL; that's very exciting,
I like that publisher a lot. What will be your focus?
>independent distros like RedHat, Slackware, Debian, Stampede,
>Yggdrasil, OpenLinux, ummm. Did I leave anyone out? :)
TurboLinux?
>Yep, us too. You'd probably be amazed to know just how many thousands
>of FreeBSD CDs we sell each month, but I'm not allowed to disclose
>sales figures on the net. I suspect that RH and the others aren't
>allowed to either, so no big loss.
Why aren't you allowed?
--
Ben
root@localhost (hi spammers!)
void
<msme...@cip.informatik.uni-erlangen.de> wrote:
>In article <slrn6q4q6j...@interport.net>,
>void <fl...@interport.net> wrote:
>>
>>FreeBSD runs Linux binaries nicely. I don't know about the other BSDs. I
>>don't believe Linux runs BSD binaries, however.
>
>Really? For instance, how is the 'clone' systemcall emulated?
Hmm. Some investigation turns up this, from sys/i386/linux/linux_dummy.c:
linux_clone(struct proc *p, void *args, int *retval)
{
printf("Linux-emul(%d): clone() not supported\n", p->p_pid);
return ENOSYS;
}
I suppose I shouldn't be surprised; clone() is awfully hard to emulate, as
syscalls go. Most of them correspond to BSD syscalls fairly closely, of
course.
Note that the lack of a clone() is not enough to keep BSD from running
Linux StarOffice, Netscape Navigator/Communicator, Quake, Doom,
RealPlayer, and Executor.
Jordan K. Hubbard
> You don't think BSD would be more popular if it hadn't had those licensing
> problems? I thought that was common wisdom, more or less. Perhaps I'm
> misconstruing REB's point.
I think it would be "more" popular, sure, but I think that the ripples
from this event have also pretty much died down given the 4 years that
have elapsed since then and it's just not as relevant in making points
about where the market is going today. To hear some folks tell it,
the licensing issues just happened yesterday and are still some sort
of major factor in our lives, and I merely wished to dispell that
particular rumor.
> I didn't know you had something in the works at AWL; that's very exciting,
> I like that publisher a lot. What will be your focus?
The ISP and "professional user" of FreeBSD.
> >independent distros like RedHat, Slackware, Debian, Stampede,
> >Yggdrasil, OpenLinux, ummm. Did I leave anyone out? :)
>
> TurboLinux?
Right! Damn, that's embarassing since Walnut Creek CDROM also
sells that one, don't they? :-)
> >Yep, us too. You'd probably be amazed to know just how many thousands
> >of FreeBSD CDs we sell each month, but I'm not allowed to disclose
> >sales figures on the net. I suspect that RH and the others aren't
> >allowed to either, so no big loss.
>
> Why aren't you allowed?
Because disclosing sales figures in general is just considered bad
juju, I guess, and I've been told by the powers that be at Walnut
Creek CDROM that they really don't want their sales figures disclosed
in public. I can only respect their wishes, since this is indeed
"internal" company data.
Roger Christie
>
> In article <yfgww9o...@time.cdrom.com>,
> >
> >I find it somewhat telling that an entire analysis of the industry can
> >be presented in a BSD group which somehow makes it sound as if Linux
> >was the only player in all of this and is the only player now. Hello?
> >Is there anyone in there? What is it about some Linux advocates that
> >compells them to completely *ignore* any of the alternatives?
>
> on dejanews. Then read, erm, CV on his homepage. Then draw your conclusions.
> The kindest description that I can give: suit par excellence, spindoctor,
> ignorant, absolutely dishonest (can't tell truth from lie). His knowledge
> of Linux is about the same as his knowledge of BSD or UNIX (zero). 100%
> buzzword-compliant. Author of wonderful neologism: 'hacker in suit'
> (absolutely seriously used it as self-description). Unusual type in the
> advocacy zoo (probably not unique, but I didn't see anything similar
> around Linux or FreeBSD).
> BTW, I'm serious about going to DN and his webpage. It's
> fascinating (albeit long) reading.
> HAND,
> Al
You're seriously contending that Rex Ballard has zero knowledge of
Unix? Based on that CV??
Er...whatever.
Rajat Datta
kskei...@my-dejanews.com <kskei...@my-dejanews.com> wrote:
>
>It's interesting that it's not uncommon to hear about Linux users who switch
>to (Free)BSD, but you never ever hear about BSD users switching to Linux.
Search dejanews for old posts to this newsgroup. There are a few,
some of whom have participated in/caused quite a few heated postings
here.
rajat
Alexander Viro
Roger Christie <rchr...@spinach.xylogics.com> wrote:
>Alexander Viro wrote:
>>
[snip]
>> Jordan, it's Rex. Please, take a *big* barf-bag and look through his stuff
>> on dejanews. Then read, erm, CV on his homepage. Then draw your conclusions.
>> The kindest description that I can give: suit par excellence, spindoctor,
>> ignorant, absolutely dishonest (can't tell truth from lie). His knowledge
>> of Linux is about the same as his knowledge of BSD or UNIX (zero). 100%
>> buzzword-compliant. Author of wonderful neologism: 'hacker in suit'
>> (absolutely seriously used it as self-description). Unusual type in the
>> advocacy zoo (probably not unique, but I didn't see anything similar
>> around Linux or FreeBSD).
>> BTW, I'm serious about going to DN and his webpage. It's
>> fascinating (albeit long) reading.
>> HAND,
>> Al
>You're seriously contending that Rex Ballard has zero knowledge of
>Unix? Based on that CV??
Based on the combination of that CV and profound ignorance demonstrated in
the newsgroups. Excuse me, _if_ the CV had anything with reality he
_would_ know better. To put it simple: his CV is outright lie. And I mean
exactly what I wrote. Look at the dejanews and you'll see yourself.
User Rdkeys Robert D. Keys
> Actually, it would be great of the BSD folks wanted to "come and play"
> with us. Certainly, it's not difficult to add the applications that
> would make BSD "Linux Compatible".
> Rex Ballard - http://www.access.digex.net/~rballard
Interesting concept..... I have been doing that since day 1, almost.
I was under the impression that almost anything in Linux or the *BSD's
could be ported back and forth between the two, almost ad nauseum.
If it is a good unix source, it should. Only if particular 'isms
are stressed would it not, or require stretching a few gotchas.
I the past 5 years or so, I have only found one of my pet sources
that compiled on Linux and not on FreeBSD. I would call that about
as compatible as unices can get. FreeBSD has some 1300? or so ports
of almost anything and everything. So does Linux. All the main stuff
is on both....
Still, swap everything you can.... that leads to a common community
sort of environment. That is only good for the ``freebie unix''
community.
RDK
Toon Moene
> Actually, a BeoWulf Linux cluster can outperform all but about 300 of the
> worlds fastest super computers. It takes a bit of doing (not for junior
> programmers), but you too can have a 7 gigaflop cluster.
It's perhaps fair to point out that the 315th entry of the worlds fastest
systems list gets 19.2 gigaflops for $152,000 - and there are years that I do
not have that kind of money lying around. A well cooled garage is a must
too.
--
Toon Moene (mailto:to...@moene.indiv.nluug.nl)
Saturnushof 14, 3738 XG Maartensdijk, The Netherlands
Phone: +31 346 214290; Fax: +31 346 214286
g77 Support: mailto:for...@gnu.org; NWP: http://www.knmi.nl/hirlam
The Monopoly Strikes Back !
Roger Christie
>
> In article <35A3AE...@spinach.xylogics.com>,
> Roger Christie <rchr...@spinach.xylogics.com> wrote:
> >Alexander Viro wrote:
> >>
> [snip]
> >> Jordan, it's Rex. Please, take a *big* barf-bag and look through his stuff
> >> on dejanews. Then read, erm, CV on his homepage. Then draw your conclusions.
> >> The kindest description that I can give: suit par excellence, spindoctor,
> >> ignorant, absolutely dishonest (can't tell truth from lie). His knowledge
> >> of Linux is about the same as his knowledge of BSD or UNIX (zero). 100%
> >> buzzword-compliant. Author of wonderful neologism: 'hacker in suit'
> >> (absolutely seriously used it as self-description). Unusual type in the
> >> advocacy zoo (probably not unique, but I didn't see anything similar
> >> around Linux or FreeBSD).
> >> BTW, I'm serious about going to DN and his webpage. It's
> >> fascinating (albeit long) reading.
> >> HAND,
> >> Al
> >You're seriously contending that Rex Ballard has zero knowledge of
> >Unix? Based on that CV??
>
> Based on the combination of that CV and profound ignorance demonstrated in
> the newsgroups. Excuse me, _if_ the CV had anything with reality he
> _would_ know better. To put it simple: his CV is outright lie. And I mean
> exactly what I wrote. Look at the dejanews and you'll see yourself.
>
Oh bullshit.
Toon Moene
> It's interesting that it's not uncommon to hear about Linux users who
switch
> to (Free)BSD, but you never ever hear about BSD users switching to Linux.
That's because there are vastly more Linux users than BSD users.
:-)
[ Sorry - couldn't resist ]
--
Toon Moene (mailto:to...@moene.indiv.nluug.nl)
Saturnushof 14, 3738 XG Maartensdijk, The Netherlands
Phone: +31 346 214290; Fax: +31 346 214286
g77 Support: mailto:for...@gnu.org; NWP: http://www.knmi.nl/hirlam
Put a Penguin in your Processor !
Chris Hedemark
wrote:
>It's perhaps fair to point out that the 315th entry of the worlds fastest
>systems list gets 19.2 gigaflops for $152,000 - and there are years that I do
>not have that kind of money lying around. A well cooled garage is a must
>too.
If you want to have a system like Avalon, yes, it is fair to point out
that $150,000 is a lot of money for the common Joe.
But I think that it is also fair to point out that $150,000 is a drop
in the bucket for that kind of power.
For the common Joe, I think it's danged cool when you think about it.
$1,000 here, $1,000 there, and before you know it for about $3,000 you
have your own Beowulf cluster with anywhere from 4 to 6 nodes. The
beauty is that you can continue to scale it as $$$ allows.
If you are handy with a soldering iron, you may be able to fit several
motherboards into one custom chassis with one big cooling fan (plus
dedicated CPU fans). But as cheap as cases are, you can get a slim
desktop case for next to nothing and stack them up 6 high without
needing a cooled garage.
You can take it one step farther in scalability by buying SMP
motherboards with only one CPU per motherboard, and as $$$ allows you
can add CPU power without having to buy a new case, motherboard,
memory, NIC, hard disk, etc.
I'd love to play around with something like this but I have no idea
what I would do with it once I had it. :-)
Christian Hedemark/Raleigh/Contr/IBM
If your response is in regards to an IBM product and you
wish to have further support above and beyond this newsgroup,
please call 1-800-CALL-AIX and open a PMR record but do not
email me directly or attempt to call me directly, please.
"From the fury of the Norsemen, oh Lord, deliver us!"
Dave Kenny
I'm stuck with a win95 box at work. I use emacs for almost everything
because its shell mode is the closest I can get to an xterm. In this
environment, emacs is proabably one of the least resource-intensive
apps I run. AND I'm able alter its behavior to fit my working style,
so I actually realize some _benefit_ in exchange for the resources
it does use.
Dave Kenny
: ...
: And now, you probably have more *cache* on your motherboard than there
: was memory and disk on the C=64.
A recurring fantasy (though not actually unrealistic) sometimes pops
up in comp.lang.forth:
"Hey, the entire _system_ would fit in the Level 1 cache!"
On the "small machines" topic, in the mid 1980's my Interactive
Development Environment was F83 forth. It had nice things like
a debugger, print spooler, multitasker (nonpre-emptive), assembler
and other goodies. It was _very_ responsive in interactive use.
This all in TINY MODEL: code, data, and stack, all in the same
64 kbyte segment of memory. Tsk tsk, of that space "only"
something like 40k was left over for user programs.
Dave Kenny
: On Mon, 6 Jul 1998, jedi wrote:
:> You've got to be kidding? Emacs is one of the bloatware KINGS.
: But it still works on an 8MB 486 - Word left that hardware behind how many
: years ago?
A few years ago my system was linux on a 20mhz 386 SL laptop with _4_ meg
ram. I didn't run X and didn't recompile the kernel on that machine, but I
could run emacs in text mode (from 4 different virtual consoles no less.)
And of course all the typical unix commands.
An even more extreme example:
The now defunct Mark Williams company had a unix called Coherent that
ran (at least in early revisions) on a 286, with 1 meg of Ram and
<gasp> 10 Meg of hard drive.
Functionality isn't usually what gobbles resources, all the cutesie nicey
nicey stuff is what costs.
Tim Hanson
>r.e.b...@usa.net writes:
>> ...
>
>I find it somewhat telling that an entire analysis of the industry can
>be presented in a BSD group which somehow makes it sound as if Linux
>was the only player in all of this and is the only player now. Hello?
>Is there anyone in there? What is it about some Linux advocates that
>compells them to completely *ignore* any of the alternatives?
>
I don't think anyone is ignoring anyone else. I capture
comp.os.linux.advocacy only, but lots of cross posts hitch along.
>For something which attempts to give a historical perspective on all
>this, it's doubly unforgivable since BSD has been around for a lot
>LONGER than Linux and, even though it wasn't allowed to be a player in
>the free software world until relatively recently (1991), it was very
>serious player in the academic world for some time before that.
>
Dunno what to say. My first guess is that Linux got its start on the
Internet, which is where we are now, after all, that's just my guess. I
can't explain the phenomenal success of Linux and not FreeBSD, but with the
user base increasing at such a rapid rate and Linux so tied to the internet
from its roots, I'm not at all surprised it dominates discussions here. As
the only OS besides Windows which is growing in user base, I am not
surprised that debate centers around the two, to the exclusion of FreeBSD.
It's unfortunate but probably not changeable in the short term.
>I am always willing to acknowledge the accomplishments of Linux, I
>just wish it were something more often reciprocated. Are the afraid
>of something? :-) [At which point here I fully expect Albert D. Calahan
>to jump in with some bad conspiracy theory or claims that gosh, the
>Linux folks just don't really even notice us. :)]
It's too bad, but they don't. Many Linux users, and certainly the vast
majority of new users, have no knowledge of FreeBSD nor of its similarities
to Linux nor of its strengths over Linux.
Tim Hanson
Tim Hanson
>In article <yfgww9o...@time.cdrom.com>,
> j...@time.cdrom.com (Jordan K. Hubbard) wrote:
>> r.e.b...@usa.net writes:
>> > ...
>>
>> I find it somewhat telling that an entire analysis of the industry can
>> be presented in a BSD group which somehow makes it sound as if Linux
>> was the only player in all of this and is the only player now. Hello?
>> Is there anyone in there? What is it about some Linux advocates that
>> compells them to completely *ignore* any of the alternatives?
>>
>> For something which attempts to give a historical perspective on all
>> this, it's doubly unforgivable since BSD has been around for a lot
>> LONGER than Linux and, even though it wasn't allowed to be a player in
>> the free software world until relatively recently (1991), it was very
>> serious player in the academic world for some time before that.
>>
>> I am always willing to acknowledge the accomplishments of Linux, I
>> just wish it were something more often reciprocated. Are the afraid
>> of something? :-) [At which point here I fully expect Albert D. Calahan
>> to jump in with some bad conspiracy theory or claims that gosh, the
>> Linux folks just don't really even notice us. :)]
>
>It's interesting that it's not uncommon to hear about Linux users who switch
>to (Free)BSD, but you never ever hear about BSD users switching to Linux.
>
>
Pointless bickering. In case you haven't noticed, a 500 pound tiger is
sitting on the sofa. No one in the Linux community wishes ill of any of
the other Unices, but MacNealy appears to have been out-connived.
Barksdale has conceded. Jobs was bought off. Microsoft is now attempting
to privatize the internet by messing with HTML without telling anyone how
to talk to it.
There is a serious predator loose. We can squabble later, when the threat
of One Microsoft Way subsides.
Tim Hanson
Lars Hecking
> Heheheh... that's my exact equipment as well. I find that when
> compiling a new kernel, "make depend" takes approximately 30 seconds,
> and "make" about 4 minutes. Under 5 minutes total to recompile the
> kernel (so much for the old documents raving about it being "one of
> the most time consuming and tedious tasks the Unix administrator has
> to face"). I'd like to try it on a PII-400 :-)
15s for make dep and 2:17m for make under RedHat 5.1, kernel 2.0.34-1,
on a PII-400 with 384MB of RAM.
> On Tue, 7 Jul 1998 19:12:51 +1000, "Richard Lyon"
> <c16...@email.mot.com> wrote:
>
> >Give me a 200MHz MMX, 64 MEG RAM and FBSD any day.
Michael C. Vergallen
>The now defunct Mark Williams company had a unix called Coherent that
>ran (at least in early revisions) on a 286, with 1 meg of Ram and
><gasp> 10 Meg of hard drive.
quite recent namely Unix has been around for 25 odd years so it used to run on
a PDP-1 with a 8 or 16 K off memory (I'm not shure off the memory).
Michael
--
Michael C. Vergallen A.k.A. Mad Mike,
Sportstraat 28 http://www.double-barrel.be/mvergall/
B 9000 Gent ftp://ftp.double-barrel.be/pub/linux/
Belgium tel : 32-9-2227764 Fax : 32-9-2224976
r.e.b...@usa.net
Toon Moene <to...@moene.indiv.nluug.nl> wrote:
> r.e.b...@usa.net wrote:
>
> > Actually, a BeoWulf Linux cluster can outperform all but about 300 of the
> > worlds fastest super computers. It takes a bit of doing (not for junior
> > programmers), but you too can have a 7 gigaflop cluster.
>
> systems list gets 19.2 gigaflops for $152,000 - and there are years that I do
> not have that kind of money lying around. A well cooled garage is a must
> too.
What is interesting is to compare this price to the price asked for a typical
Microsoft Windows NT web site. The NT site (4 proccessor compaq, software,
CALS, and other software are easily double this price.
As to the well cooled garage, you might want to just space them out a few
feet, or buy some big 19 inch racks. Minitowers are about 15 inches high
and 8 inches wide. You can put about 18 of them on steel shelves (3'x 6').
A Sun Enterprise system runs over $1/5 million. A Sparc Enterprise 10000
runs over 1.2 million when fully configured.
I've got 6 computers, each of which cost less than $300 per box. The slowest
machine is a Pentium 66, and the fastest is a Cyrix 6x6/200 on a TX
motherboard. 3 are Pentium 100s. I buy the pieces at computer fairs (near
wholesale prices) and build them from components. If you add my time (I do
this as a hobby), it's probably more like a $25,000 network :-).
The typical box:
Case/Power 20
MB/CPU 100
2 gig drive 100
ram 40
CD Rom 40
Floppy 15
Linux 30
------------------
Total $335
> --
> Toon Moene (mailto:to...@moene.indiv.nluug.nl)
> The Monopoly Strikes Back !
So do I
Rex Ballard - http://www.access.digex.net/~rballard
-----== Posted via Deja News, The Leader in Internet Discussion ==-----
r.e.b...@usa.net
> >I find it somewhat telling that an entire analysis of the industry can
> >be presented in a BSD group which somehow makes it sound as if Linux
> >was the only player in all of this and is the only player now. Hello?
> >Is there anyone in there? What is it about some Linux advocates that
> >compells them to completely *ignore* any of the alternatives?
In the "Free UNIX" market, Linux WAS the only player in the 32 bit arena.
Minux was a good 16 bit system. BSD "Lite" still had AT&T code in it up
until 1993. I remember calling for a "Free BSD" on the net, but Linux
came up instead. This was in 1992 in comp.os.unix.* when I was posting
as rbal...@softronics.com.
> Jordan, it's Rex. Please, take a *big* barf-bag and look through his stuff
> on dejanews. Then read, erm, CV on his homepage. Then draw your conclusions.
By all means, read the stuff. It's very rough. Someday I'd like to organize
it, dig out my old archives from 1984 to 1994, and put together a "Biography
in E-mail". Right now, I'm too busy creating the future to dwell on the past.
> The kindest description that I can give: suit par excellence, spindoctor,
> ignorant, absolutely dishonest (can't tell truth from lie).
I'm a promoter, I'll admit that. I've been bringing new technology to
the general public since 1974 when I started selling CB radios to
people after finding out that truckers were using them to locate speed
traps.
> His knowledge
> of Linux is about the same as his knowledge of BSD or UNIX (zero).
Of course, it's irrelevant that I've been earning a living for over 15
years primarily as a UNIX programmer, administrator, and system architect.
I may not be making as much as Bill Gates, but I make enough to pay my bills
without becoming a target for contingency lawyers.
> 100% buzzword-compliant.
Actually, when I first worked on some of these projects, those buzzwords
weren't even around. I figure it's just easier to use the "spin-doctor's"
language to communicate terms that are commonly understood today.
What we call the Internet had a very different meaning in 1987.
I've also gotten pretty "thick skinned". You don't survive asthma,
hay fever, chigella, dysintary, epilepsy, alcoholism, drug addition,
divorces, and daily beatings for being a "sissy" without getting to
the point where you don't let a few insults stop you from doing what
you think is really important.
When I went to Dow Jones in January of 1993 and said "I want to put
you on the Internet", do you think they said "What a wonderful idea!!"?
No - they looked at me like I had three heads. They were making $25/hour
on NRS and were getting over $300/person/month from some of their feeds.
I knew that it was only a matter of a few months before the Internet
grew from 1 million users to over 64 million users. I knew that if DJ
was there, at the front of the train, that they would make some big
money.
Ironically, many of my suggestions were ignored by DJ and picked up by
companies that eventually became Lycos, Yahoo, Wired, and hundreds of
other big names. Even AOL listened.
I didn't force anyone to do anything. In fact, all I did was find
courageous people who were willing to enroll leaders in their respective
companies into taking on a project that was speculative and risky, but
provided the possibility of economic abundance.
> Author of wonderful neologism: 'hacker in suit'
> (absolutely seriously used it as self-description).
Yup! There are hundreds of people who have waded through every inch
of the Linux kernel. I've waded through the AT&T V7 kernel, the BSD 4.2
kernel, the OSF/1 kernel, and the Linux kernel. In my younger days, I also
waded through CP/M, MS-DOS, The Atari 800, and the Apple II assembly code.
The technology is fun, and I love UNIX as a suporior operating system. There
are some nice features about BSD, and some nice features about Linux. What
I'm interested in doing is creating the possibility for people that Linux
can provide something for people that Microsoft can't.
If you want to make a strong case for FreeBSD, I'd love to hear it. What
would it take to generate 11 million BSD users within the next 12 months?
The fact is that I've posted nearly 1000 articles to Usenet addressing the
possibilities of Linux (reliability, stability, security, economy, ease of
use relative to other versions of UNIX, and compatibility/interoperability).
What that means, bottom line, to a user is that he will spend less time
trying to deal with a flaky machine, and spend more time gathering and
publishing information that can make a difference to the "bottom line".
BSD is also a pretty good system. It's also a very mature technology. In
fact, in many ways, it's more "robust". The Linux community took on the
challenge fo taking on Microsoft 3 years ago, and given the nature of the
organization, they are doing a pretty good job. Does the FreeBSD community
want to play this game too? You're certainly welcome, and I'm enrollable.
It's worth pointing out that the BSD community is a major contributor to the
application suite that makes up the "Linux Distribution". The typical
Linux CD-ROM is more like a Linux kernel with thousands of applications
that were originally developed for BSD under GPL.
> Unusual type in the
> advocacy zoo (probably not unique, but I didn't see anything similar
> around Linux or FreeBSD).
Actually an "Outrageous Combination" a cross between a "Computer Geek",
a "Home Electronics Salesman", and an Actor.
> BTW, I'm serious about going to DN and his webpage. It's
> fascinating (albeit long) reading.
Actually, I will have posted my 1000th message under this login.
I have also posted under the moniker "rbal...@access.digex.net",
rbal...@dowjones.com, rbal...@softronics.com, and rbal...@snp.com
If you dig up ancient history, I'm also in there as "rexb!ccivax and
"re...@ccivax.cci.com" and "re...@ccivax.rit.edu". Those go way back
to 1984 (I have reel-to-reel tape :-).
By the way, when Al, or many other posters provide corrections to substantive
information (I did a big boo-boo), I'll either clarify, or admit that I
made a mistake.
This particular posting seems to be a general personal attack on my
integrity, honesty, and intentions.
In general, I try to be as honest as possible (I don't knowingly and
intentionally lie - but I have sometimes gone on information collected
sources other than personal experience). As for integrity, I am fairly
true to my ideals, values, and standards - I haven't "sold out", even when
it might have been expedient to do so. And as for my intentions, I'm a
very happy and satisfied Linux customer who just loves to share Linux with
people because I can be pretty sure that they will be happy that they
tried it.
I've tried FreeBSD, and haven't liked it as much as the better versions
of Linux. I'm also not real crazy about Debian Linux. Both systems
are actually very impressive (either is an improvement over Microsoft),
but I just haven't been as impressed yet. I am enrollable.
> HAND,
> Al
Keep up the good work Al.
r.e.b...@usa.net
fl...@interport.net (void) wrote:
> On 08 Jul 1998 10:02:07 -0700, Jordan K. Hubbard <j...@time.cdrom.com> wrote:
> >r.e.b...@usa.net writes:
> >> Let's look at that. I use a number of different indicators including
> >> web site counts, bookstore title counts, and several other metrics.
> >> BSD is a presence. BSD was hindered by some of the AT&T code and
> >> this kept it from being able to be distributed freely. Eventually,
> >> FreeBSD did come out in a freely distributable form.
> >
> >Huh? I don't see how these two points can be mated at all.
>
> You don't think BSD would be more popular if it hadn't had those licensing
> problems? I thought that was common wisdom, more or less. Perhaps I'm
> misconstruing REB's point.
This is the point exactly. By the way, it doesn't take BSD out of the
game, but rather puts them into a "second wave" situation.
> ISTR reading that Linus Torvalds said that he wouldn't have written Linux
> if 386/BSD had been available at that point.
Linux wouldn't have been adopted either. In early 1992, as part of a project
to commercialize the internet and make it available to the general public,
I asked the comp.os.unix groups if there was any variant of UNIX that
could be sold for less than $100. The only candidates to come back
were Minix (couldn't run X11, 64kb process limit), and Linux (32 bits,
didn't even have an X11 port at the time. Even then it was pretty impressive.
In January of 1993, the field had grown to include Mark Williams' Coherant.
A variant of Minix I assume (similar limits). Novell purchased the Rights
to UNIX from AT&T and dropped the "floor" (the minimum selling price)
to $200 (for an IPX/only version). Eventually, Novell lifted the restrictions
on the BSD-Lite code that included AT&T code). This was nearly 18 months
after the Linux release.
> >Once I get my book for Addison Wesley finished, I
> >expect to see even greater penetration of the same market.
More and more "UNIX" books are coming out now. It would be nice if some of
them started including complimentary versions of BSD (with instructions for
how and where to get support contracts of course :-).
> I didn't know you had something in the works at AWL; that's very exciting,
> I like that publisher a lot. What will be your focus?
>
> >independent distros like RedHat, Slackware, Debian, Stampede,
> >Yggdrasil, OpenLinux, ummm. Did I leave anyone out? :)
>
> TurboLinux?
>
> >Yep, us too. You'd probably be amazed to know just how many thousands
> >of FreeBSD CDs we sell each month, but I'm not allowed to disclose
> >sales figures on the net. I suspect that RH and the others aren't
> >allowed to either, so no big loss.
Isn't it fun. You're watching something very interesting aren't you?
I wouldn't be surprised if both systems started growing at 20%/month
rates over the next few months (that means the population doubles every
3 months).
> Why aren't you allowed?
Can you give us a hint?
> --
>
> Ben
Alexander Viro
>In article <6nvdpp$2...@steklov.math.psu.edu>,
> vi...@steklov.math.psu.edu (Alexander Viro) wrote:
>> His knowledge
>> of Linux is about the same as his knowledge of BSD or UNIX (zero).
>
>Of course, it's irrelevant that I've been earning a living for over 15
>years primarily as a UNIX programmer, administrator, and system architect.
Sure. From our, erm, conversation in c.o.l.a:
Rex> [...]
Rex> reason all backups are performed by a process owned by root.
me> a) Nobody locks the file this weird way.
me> b) Backups may be done by anybody. System-wide backup requires
me> root, indeed. But "all backups are performed by a process owned by
me> root" is plain idiocy. Did you ever use tar? Or you have rolled suid-root
me> version?
Rex> If I use a non suid-root tar to do a full backup of all files, those
Rex> owned by users which have been chmod 600 (mail files, profiles...)
Rex> will not be backed up. At minimum, tar must have read access to the
Rex> file. Root can access all files.
That's for "administrator" bit, gentlemen. Rex, ITYM "uber-luser". Earth
to Rex: suid-root _tar_ may appear on a system in two cases: either it
is FUBARed by extremely perversive cracker, or sysadmin in question
deserves to be LARTed out of existence. Slow and painfully. 14 years seems
to be unrealistic idea, but taunting.
From the same place:
me> Um-hm. Show me a database server that fsync's on every write...
Rex> lpd, snmpd, nntpd - all of these at one time used fsyncs to post to
Rex> log files.
Enjoy the show. lpd as database server with log as database. That's for
"system architect" and "programmer".
_Any_ more questions about Rex's experience and CV? Thank you.
Matt Dillon
:>> and "make" about 4 minutes. Under 5 minutes total to recompile the
:>> kernel (so much for the old documents raving about it being "one of
:>> the most time consuming and tedious tasks the Unix administrator has
:>> to face"). I'd like to try it on a PII-400 :-)
on a duel pentium-II/300 running FreeBSD-current/SMP and softupdates,
make depend takes 22 seconds and a kernel build with make -j2
takes 1:45 (one minute, 45 seconds).
A make world (including the profiling libraries) takes one hour and
FIVE minutes, which impresses me!
-Matt
:> 15s for make dep and 2:17m for make under RedHat 5.1, kernel 2.0.34-1,
:> on a PII-400 with 384MB of RAM.
:>
>> >Give me a 200MHz MMX, 64 MEG RAM and FBSD any day.
--
Matthew Dillon Engineering, HiWay Technologies, Inc. & BEST Internet
Communications
<dil...@best.net> (Please include original email in any response)
Steinar Haug
| 15s for make dep and 2:17m for make under RedHat 5.1, kernel 2.0.34-1,
| on a PII-400 with 384MB of RAM.
Similar times with FreeBSD 3.0-980506-SNAP: 13s for make dep, 1:38m for make.
This is on a PII-400, BX chipset/100 MHz bus, 128 MB RAM.
I like the PII-400 :-)
Steinar Haug, Nethelp consulting, sth...@nethelp.no
Pierre Beyssac
>In article <yfgww9o...@time.cdrom.com>,
> j...@time.cdrom.com (Jordan K. Hubbard) wrote:
>the source is right there).
...
>Actually, it would be great of the BSD folks wanted to "come and play"
>with us. Certainly, it's not difficult to add the applications that
>would make BSD "Linux Compatible".
Funny, you don't even seem to realize how arrogant (and ignorant)
a statement this is !
First of all, most of the free *BSDs include a Linux emulator and
run most Linux applications without any problem.
But just because Linux suffers from a severe case of NIH syndrome
(I elaborate on that just below) doesn't mean the world has to
change to adapt to Linux. That the opposite isn't even considered
by most people in the Linux community is, to say the least, rather
telling. Sadly enough, this is for that kind of behaviour that
Microsoft has been having lots and lots of bad press recently. I
don't see why this should be more acceptable from Linux. Instead
of that, the non-Linux systems are accused of lack of cooperation!
Now if you want a list of things reinvented in Linux for not good
reason (this is mainly focused on the network code because that's
what I've personaly been plagued with the most), here's my list
of gratuitous changes:
- the IP packets definition includes are different from
the de-facto standard (BSD). This means that most of the
Linux code using raw sockets has to be PORTED to other
systems (and the reverse is true). Of course you can also
put BSD includes under Linux, but that doesn't seem to
be considered cosher by most Linux distributions.
- the semantics of non-blocking sockets are different. Have
a look at *BSD's Linux emulator code for connect() if you
want to have fun...
- the Linux interface for packet sniffers (SOCK_PACKET)
induces a _lot_ of overhead (packet copies between user-
and kernel-mode, followed by filtering in user mode).
The Berkeley packet filter (bpf) solved all this years
ago.
- the Linux IP code has been rewritten from scratch (because
BSD code was deemed "unclean"). Fine with me, but this
introduced yet another set of subtle incompatibilities.
Some routing flags, for example, are missing under Linux.
- Linux ifconfig has a behaviour different from any other
system I know of (you have to put a route by hand
afterwards).
- and you could certainly go on and on.
Linux has made strictly no attempt at backwards compatibility
regarding the network code, except for very basic user-level socket
behaviour, which is the least.
Now try to compile a functional Gated on a Linux system without
being a Linux network guru, and you'll quickly understand why
research work in IP currently occurs mostly on BSD-based systems:
because they don't have that much portability problems.
And the next thing we know, the BSD's will be accused of not being
Linux-compatible... :-))))
--
Pierre....@hsc.fr
Bernd Paysan
> Dunno what to say. My first guess is that Linux got its start on the
> Internet, which is where we are now, after all, that's just my guess. I
> can't explain the phenomenal success of Linux and not FreeBSD, but with the
> user base increasing at such a rapid rate and Linux so tied to the internet
> from its roots, I'm not at all surprised it dominates discussions here. As
> the only OS besides Windows which is growing in user base, I am not
> surprised that debate centers around the two, to the exclusion of FreeBSD.
> It's unfortunate but probably not changeable in the short term.
I have no difficulties to explain the success of Linux over *BSD. First,
Linux had a head start. For about a year, there was only Linux, and no
*BSD. Then, the first free BSD came out, but it still had legal problems
with AT&T. This hindered a success in a time-frame where *BSD was
technically better than Linux. And then, the free BSDs are split up into
FreeBSD, OpenBSD, and NetBSD, while Linux development is still united.
The main thing is that even within geeks, mindshare is important. The
technical quality of Linux and any of the *BSD is so small, that it
isn't important. Linux cares a bit less about "backward compatibility"
(to what? SVR4 or BSD?), *BSD more (backward compatibility to itself -
BSD). Most application compile and run on both systems without problems.
Yeah, if you want to compile a packet sniffer... *BSD is more popular
among crackers, because it can sniff packets better than Linux.
The competition is good for both systems. The NIH-syndrome is often due
to license problems. You can't just ripp of a portion of GPL'd Linux
code and include it into *BSD. It's easier to do it the other way round,
but since the enhancements won't come back, the *BSD hackers won't do
it. E.g. Linux still doesn't have a r/w FFS implementation, although it
should be not more work than take the *BSD code and put it under the
Linux VFS layer.
--
Bernd Paysan
"Late answers are wrong answers!"
http://www.jwdt.com/~paysan/
Bill Gunshannon
|>
|> On the "small machines" topic, in the mid 1980's my Interactive
|> Development Environment was F83 forth. It had nice things like
|> a debugger, print spooler, multitasker (nonpre-emptive), assembler
|> and other goodies. It was _very_ responsive in interactive use.
|> This all in TINY MODEL: code, data, and stack, all in the same
|> 64 kbyte segment of memory. Tsk tsk, of that space "only"
|> something like 40k was left over for user programs.
Did you ever run OS9?? Not only multitasking, but multiuser (I
actually was able to get 3 users simultaneously, console and 2
serial ports) and all in 64K running off of a 5.25" floppy.
Programmers today have no concept whatsoever of resource economy.
bill
--
Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves
bi...@cs.uofs.edu | and a sheep voting on what's for dinner.
University of Scranton |
Scranton, Pennsylvania | #include <std.disclaimer.h>
Donal K. Fellows
Pierre Beyssac <p...@mars.hsc.fr> wrote:
> Now if you want a list of things reinvented in Linux for not good
> reason (this is mainly focused on the network code because that's
> what I've personaly been plagued with the most), here's my list
> of gratuitous changes:
Which version of Linux are you talking about? It does make a (big)
difference, you know...
Donal.
--
Donal K. Fellows http://www.cs.man.ac.uk/~fellowsd/ fell...@cs.man.ac.uk
Department of Computer Science, University of Manchester, U.K. +44-161-275-6137
--
Never underestimate the power of the penguin...
Pierre Beyssac
>Linux had a head start. For about a year, there was only Linux, and no
>*BSD. Then, the first free BSD came out, but it still had legal problems
>with AT&T. This hindered a success in a time-frame where *BSD was
>technically better than Linux. And then, the free BSDs are split up into
>FreeBSD, OpenBSD, and NetBSD, while Linux development is still united.
Depends on how you look at things. IMHO, this is a myth. With Linux,
kernel development is united, but there are many, many different
Linux distributions (and I'm not even talking about the Debian/Red
Hat ideological war). Since what I'm using is a kernel + a set of
executables, it really looks to me like the BSD world is actually
_more_ united than the Linux world because is has mainly 3
distributions :-).
>isn't important. Linux cares a bit less about "backward compatibility"
>(to what? SVR4 or BSD?), *BSD more (backward compatibility to itself -
>BSD).
Mostly to all of the rest: SVR4, BSD, HPUX, AIX, OSF/1... All of
these have mostly BSD-based network semantics (and includes).
>Yeah, if you want to compile a packet sniffer... *BSD is more popular
>among crackers, because it can sniff packets better than Linux.
Crackers don't care about efficiency, because they steal CPU time
from others :)
The only thing I know is that most network exploits posted to the
Bugtraq security list are for Linux, and most "user-friendly" packet
sniffers are for Linux.
>to license problems. You can't just ripp of a portion of GPL'd Linux
>code and include it into *BSD. It's easier to do it the other way round,
>but since the enhancements won't come back, the *BSD hackers won't do
>it.
I don't quite understand your point here, what do you mean by "the
other way around" and "the enhancements won't come back" ?
>E.g. Linux still doesn't have a r/w FFS implementation, although it
>should be not more work than take the *BSD code and put it under the
>Linux VFS layer.
No more, but no less. The VFS layers are probably different enough
that it's easier said than done. Actually, for several reasons
(some of them technical, some of them regarding licenses), the
ext2fs code in *BSD is a compatible reimplementation, not a port.
--
Pierre....@hsc.fr
Tracy R Reed
>That's for "administrator" bit, gentlemen. Rex, ITYM "uber-luser". Earth
>to Rex: suid-root _tar_ may appear on a system in two cases: either it
I think you are nit picking and deliberately misunderstanding what Rex said.
It is perfectly clear to me that Rex is speaking of a full system backup (the
kind most often performed, in my experience) and NOT individual users backing
up their files. I am quite sure he knows that any user on the system can back
up their individual files by using tar. However, I am pretty sure that none of
the 200 users on my systems do that. That is why I have a root process dump the
filesystems to DLT every night.
>me> Um-hm. Show me a database server that fsync's on every write...
>Rex> lpd, snmpd, nntpd - all of these at one time used fsyncs to post to
>Rex> log files.
>
>Enjoy the show. lpd as database server with log as database. That's for
>"system architect" and "programmer".
It is also clear that Rex was speaking of the more traditional general puropse
SQL database products, not these specific daemons who can afford that kind of
performance hit.
>_Any_ more questions about Rex's experience and CV? Thank you.
I don't think there is any question as to Rex's experience.
--
Tracy Reed http://www.ultraviolet.org
``The lyf so short, the craft so long to lerne.'' - Chaucer
Steve Sheldon
>Pointless bickering. In case you haven't noticed, a 500 pound tiger is
>sitting on the sofa. No one in the Linux community wishes ill of any of
>the other Unices, but MacNealy appears to have been out-connived.
>Barksdale has conceded. Jobs was bought off. Microsoft is now attempting
>to privatize the internet by messing with HTML without telling anyone how
>to talk to it.
>There is a serious predator loose. We can squabble later, when the threat
>of One Microsoft Way subsides.
Talk about serious paranoia.
Steve Sheldon
>In article <6o0jo9$r2d$1...@sucker.nl.uu.net>,
> Toon Moene <to...@moene.indiv.nluug.nl> wrote:
>> It's perhaps fair to point out that the 315th entry of the worlds fastest
>> systems list gets 19.2 gigaflops for $152,000 - and there are years that I do
>> not have that kind of money lying around. A well cooled garage is a must
>> too.
>What is interesting is to compare this price to the price asked for a typical
>Microsoft Windows NT web site. The NT site (4 proccessor compaq, software,
>CALS, and other software are easily double this price.
$300,000?
Compaq Proliant 6500 $~$13,000
Processor upgrade $~1700 * 3
512MB Ram upgrade $~2100
SmartArray cont $~1500
9.1 Gig drives $~1300 * 3
Ok, you're at about $25-26k for hardware
Then add $900 or so for a Windows NT Server license. SQL Server with the
internet pack will cost around $3-4k if you use that. If you have a very
large site, then you'll probably be investing in Oracle, and that's another
story all together.
You don't need CACLS to use this as an internet server, only for intranet,
and if you are an NT shop you have NT file servers so the CACLS have been
purchased for that.
The only difference between this and a FreeBSD solution is the cost of that
software. Both of these OSs run on intel hardware so that's a nonissue.
I'm not here to argue that NT Server makes a great internet server. But
where do you pull these numbers from?
Pierre Beyssac
>Pierre Beyssac <p...@mars.hsc.fr> wrote:
>> Now if you want a list of things reinvented in Linux for not good
>> reason (this is mainly focused on the network code because that's
>
>difference, you know...
Uh, I mostly talked about kernel stuff, I believe it is the same
accross distributions. Regarding userland stuff, most Linux machines
I have access to are RedHats, but again I certainly hope things
like "ifconfig" are the same accross distributions :-)
--
Pierre....@hsc.fr
Curtis D. Levin
1g$92s$2...@uni00nw.unity.ncsu.edu>
Reply-To: kd4...@amsat.org
Followup-To:
On 8 Jul 1998 14:34:24 GMT, User Rdkeys Robert D. Keys
<rdk...@seedlab1.cropsci.ncsu.edu> wrote:
>In comp.unix.advocacy Fewtch <fewtch@dont_spam_me_serv.net> wrote:
>
>It really depends upon what tools you need. If you don't need gui's
>on a machine, then it can flat fly with a goodly set of unix tools,
>even on dos. Dos without the unix tools is rather underpowered.
>The tools I need for writing are limited by the input speed that I
>can muster. Hence, for that kind of specific application, even a
>CP/M box or dos box works fine. Alas, fitting a troff on CP/M is
>a bust, but there are some good nroffs for it. I am still looking
>for the ellusive vi for CP/M. Someday, I may port it for fun.
The nicest thing about dos is it's ability to run gui apps out of
it's native environment. I shudder to think what a netscape dos
4.0 would do on a pentium II. Probably be alot faster than win9*
will ever be, although some 32 bit upgrades might be in order.
Linux gets close with svgalibs, but they don't see as wide a usage
as they could. I don't think BSD has any graphics capabilities
for anything more intense than VGA.
The other nicest thing about Dos was the sheer availability of
applications. Dos had everything under the sun. The only reason
was/is the shareware concept. Amazing what people will do if they
think that making money is possible.
But it does suffer from lack of protection. That is the thing that
killed it. Constant crashes, and too much lost data. Other than
that, I liked it best. It's nice to sit down in front of a computer
and have it perform to expectations. Maybe that's what we've all
missed. When I first used linux, I asked if there was color available
on apps at the console, and got several replies. Things like, color
hurts my eyes, I like mono better, etc. As you can see now, even 16
colors makes a big difference.
Think about what you want in a computer. A decent OS with nothing, or
everything but a decent OS. I tell you what. If I could program like
some of those old shareware programmers used to, I'd be a millionaire
by now, writing apps for *nix. Because believe it or not, people would
buy the apps if they were good, and people were pleased. Not only that,
but we would have a much greater proliferation of both apps and users.
Maybe the demo concept works, but not as well as partially functional
shareware did. Does web TV support Java ? We'll see soon irregardless.
ttyl.
Curtis - kd4zkw
BR
>
> In article <6o0pg4$srp$2...@blackice.winternet.com>, Dave Kenny <d...@tundra.winternet.com> writes:
> |>
> |> On the "small machines" topic, in the mid 1980's my Interactive
> |> Development Environment was F83 forth. It had nice things like
> |> a debugger, print spooler, multitasker (nonpre-emptive), assembler
> |> and other goodies. It was _very_ responsive in interactive use.
> |> This all in TINY MODEL: code, data, and stack, all in the same
> |> 64 kbyte segment of memory. Tsk tsk, of that space "only"
> |> something like 40k was left over for user programs.
>
> Did you ever run OS9?? Not only multitasking, but multiuser (I
> actually was able to get 3 users simultaneously, console and 2
> serial ports) and all in 64K running off of a 5.25" floppy.
>
> Programmers today have no concept whatsoever of resource economy.
I can think of one general exception to this. Embedded systems
programmer.
> bill
>
> --
> Bill Gunshannon | de-moc-ra-cy (di mok' ra see) n. Three wolves
> bi...@cs.uofs.edu | and a sheep voting on what's for dinner.
> University of Scranton |
> Scranton, Pennsylvania | #include <std.disclaimer.h>
--
************************
* Enjoy the pane-Run NT*
************************