Linux-Development-Sys Digest #785, Volume #6 Sat, 5 Jun 99 08:14:06 EDT
Contents:
Re: Configuration Manager for Linux ([EMAIL PROTECTED])
Re: Configuration Manager for Linux ([EMAIL PROTECTED])
Re: TAO: the ultimate OS (Vladimir Z. Nuri)
Re: TAO: the ultimate OS (Vladimir Z. Nuri)
Re: TAO: the ultimate OS (Alexander Viro)
Using PIII with Linux (Tim Burgess)
Re: TAO: the ultimate OS (Vladimir Z. Nuri)
Re: 2.2.9 kernel too big? (IanP)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Configuration Manager for Linux
Date: Sat, 05 Jun 1999 07:33:26 GMT
I think the bottom line is this:
A commercial, non-open sourced piece of software (bits and pieces of OSS
do not count) really wouldn't fly in the Linux community. If you made a
really nice product, yes you'd probably gain a number of users and make a
nice bit of money, but it would never fully catch on in the Linux
community. Distributors such as Red Hat, SuSE, Caldera, etc. couldn't
simply toss in a copy. Perhaps they could include a demo version (such as
Red Hat has done with a number of packages), but many users would opt not
to use it.
If the underlying architecture is commercial, most Linux/OSS developers
are going to ignore it. Their module might be OSS, but it would never
reach full potential without a freely available+distributable+changeable
base.
Supposing you do make the "perfect" admin tool for Linux,
commercialization will simply push Linux developers to make a new,
free-software system, this pushing you largely out of the picture.
This message isn't supposed to be a flame against your idea here, but more
of a warning. A commercial product of this nature might be relatively
successful (you'll make money), however it will never acheive widespread
adoption in the Linux community.
--
David Ludwig | "The Linux philosophy is laugh in the face of
davidl<at>wpi.edu | danger. Oops. Wrong One. 'Do it yourself.'
http://www.wpi.edu/~davidl | That's it." - Linus Torvalds
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Configuration Manager for Linux
Date: Sat, 05 Jun 1999 07:52:35 GMT
Vassilis Virvilis <[EMAIL PROTECTED]> wrote:
> Both opinions have their strong and weak points and I wouldn't like you to think
> that I am advocating either of two.
Why not use both?
First, you create a system which has different modules that interact with
the various types of configuration files. Caldera's COAS system [1] had a
name for these: "mappers".
Then, create a libConfig and try to persuade developers to use it.
Implementing it into the config system should be as simple as adding in a
new mapper/module/whatever.
[1] COAS probably has the best architecture for any config system to date,
however it leaves a lot to be desired. IMHO however, there are enough
architectural differences with what I perceive to be the "Right Way" of
doing this such that a new project/architecture is warranted.
--
David Ludwig | "The Linux philosophy is laugh in the face of
davidl<at>wpi.edu | danger. Oops. Wrong One. 'Do it yourself.'
http://www.wpi.edu/~davidl | That's it." - Linus Torvalds
------------------------------
Crossposted-To: alt.os.linux,comp.os.linux.advocacy,comp.os.misc,comp.unix.advocacy
From: [EMAIL PROTECTED] (Vladimir Z. Nuri)
Subject: Re: TAO: the ultimate OS
Date: Sat, 5 Jun 1999 09:03:34 GMT
G. Sumner Hayes ([EMAIL PROTECTED]) wrote:
: Human-computer interaction.
"ergonomics".. a term I was already using
: HCI should be the primary focus of your research.
its a good/ fruitful area. basically I am considering
my own use of OSes (all the major ones) as my own
research program. the HCI you cite is not specifically
focused on the goal of creating an OS, no? in fact
perhaps only a minority percent of it is? what in the
essay do you feel directly contradicts known aspects
of HCI?
But you show
: a complete lack of understanding of the field.
not fair. I show a lack of citation for your pet area
of speciality.
: 3. Move to a different forum that might be more receptive.
name it and I will crosspost there in the next time.
does "more receptive" mean "no comments"? ahhaha I am
comfortable with the reaction. the flames are less with
every post.
: That is assuming that you're actually interested in coming up
: with a better system. If you just like flamewars then your
: approach is doing fine.
I enjoy both flamewars & reasoned discussions, and recognize
that it usually isn't my choice.. it's a crapshoot.
: In your document you say things that show that you really don't
: understand common object models -- for instance, you conflate
: "C++-style object" with "interface", which shows that you either
: don't understand the C++ object paradigm or you don't know what
: interface means in the OO community.
there are a lot of technical terms that have different meanings
within different communities. C++ objects have an interface
in the sense of encapsulation. I am not interested in debating
trivial semantic differences.
: But you haven't clearly defined what an object is. You've
: listed a few properties of objects, but not enough to give the
: impression that you've seriously considered the options and
: made informed choices.
what is an object? that sounds like a philosophical debate
to me, dude. I listed the basic properties of the objects
I am considering in my essay, and pointed out that is part
of the design goal to nail down all the crucial ones.
all your design questions are to be answered in
specific implementations. its a low level detail
relative to the level of detail of the essay.
I agree it needs to be fleshed out.
: All of these issues have huge amounts of literature behind them, and
: yet you make no effort to discuss them.
ok, next time I will just scan every book on objects and upload
them as an addendum to your mailbox. the post is already
over 20 pages or so (or whatever) printed.
Throwing a word like "type"
: or "interface" out there and then not discussing the basics of it
: is unprofessional at the least.
hooooooo boy!! someone call the object police and arrest me!!
: but you've demonstrated a complete
: lack of understanding of the past 20 years of research in both
: programming models (e.g. object orientation, functional programming,
: structural programming, information-flow programming) and HCI (and
: the two are sometimes closely entwined).
correction, I've demonstrated a lack of interest in quoting
authorities (perhaps such as yourself)
& focusing on minutia that is of no relevance
to the basic vision outlined in the document, which btw
you have shown a "complete lack of understanding of" by
your own posts ...
--
~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^
"in theory, there's no difference [EMAIL PROTECTED]
between theory and practice, mad genius research lab
but in practice there is!" http://www8.pair.com/mnajtiv/
------------------------------
Crossposted-To: alt.os.linux,comp.os.linux.advocacy,comp.os.misc,comp.unix.advocacy
From: [EMAIL PROTECTED] (Vladimir Z. Nuri)
Subject: Re: TAO: the ultimate OS
Date: Sat, 5 Jun 1999 09:08:25 GMT
G. Sumner Hayes ([EMAIL PROTECTED]) wrote:
GSH, I am getting tired of your obvious
nitpicking, mr alpha male. what
is your definition of object? use your favorite definition
and plug it into the essay and see if it contradicts
a single design goal of the document.
the document does need someone such as yourself to
help flesh out exact defn of "object" for use by the
system. I admit that. so what?
the essay is basically suggesting a c++ like object
is what is preferred. its intentionally vague so
that advances in object design can be considered to
be part of the goal. do you go and harangue everyone
in object newsgroups about the defn of "object"
every time they use the term, saying they are not
citing the last 20 years of literature, or are
obviously ignorant of it because they are not
citing it? I've written small amts of c++, what more
familiarity with objects do you want?
--
~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^
"in theory, there's no difference [EMAIL PROTECTED]
between theory and practice, mad genius research lab
but in practice there is!" http://www8.pair.com/mnajtiv/
------------------------------
From: [EMAIL PROTECTED] (Alexander Viro)
Crossposted-To: alt.os.linux,comp.os.linux.advocacy,comp.os.misc,comp.unix.advocacy
Subject: Re: TAO: the ultimate OS
Date: 5 Jun 1999 05:19:53 -0400
In article <[EMAIL PROTECTED]>,
Vladimir Z. Nuri <[EMAIL PROTECTED]> wrote:
>I've written small amts of c++, what more
>familiarity with objects do you want?
ROFLMAO
Man, that's it. Exactly. Thanks for .sig, BTW.
--
I've written small amts of c++, what more familiarity with objects do you want?
Vladimir Nuri on c.o.l.d.s, after lenghty drivel re developing OO-OS.
------------------------------
From: Tim Burgess <[EMAIL PROTECTED]>
Crossposted-To: comp.lang.asm.x86
Subject: Using PIII with Linux
Date: 5 Jun 1999 03:26:31 GMT
I've got an app thats spends about 95% of it's total
time in a subroutine that does *many* floating point
calculations, so I'm keen to use the KNI features.
I have applied Ingo Molnar's 2.2.5 patch cleanly to
my 2.2.9 kernel (nice to see PentiumIII appear at boot).
I'm using ecgs 2.91.66-2 and binutils 2.9.1.0.24. Yet
the little test app that I found at:
http://csl.anu.edu.au/~daa/katmai.html
fails. It correctly detects KNI but then seg faults at
the first MOVAPS_MR instruction. Any suggestions as to
what is wrong?
void sane(void) {
float A[] = {1.0, 2.0, 3.0, 4.0};
float B[] = {0.5, 1.5, 2.5, 3.5}; /* 0.5, 3.0, 7.5, 14.0 */
float *a;
float *b;
printf("Doing sanity test...Should get {0.5, 3.0, 7.5, 14.0}\n");
/* Allocate two floating point arrays aligned on 128 bit boundary */
a = (float *)malloc(sizeof(float)*(4 + PAD));
a = ALIGN(a);
if (a == NULL) printf("aligned malloc of a failed\n");
else printf("Called malloc of a\n");
memcpy(a, A, sizeof(float)*4); /* Copy non-aligned into aligned */
b = (float *)malloc(sizeof(float)*(4 + PAD));
b = ALIGN(b);
if (b == NULL) printf("aligned malloc of b failed\n");
else printf("Called malloc of b\n");
memcpy(b, B, sizeof(float)*4);
MOV_MR(a, eax);
MOVAPS_MR(eax, 0); <-spits the dummy here
MOV_MR(b, eax);
MOVAPS_MR(eax, 1);
MULPS(0, 1);
MOVAPS_RM(1, eax);
printf("Result: {%g, %g, %g, %g}\n", b[0], b[1], b[2], b[3]);
}
The header file I'm using is:
#define PAD 4
#define ALIGN(a) (float *)( ( ((int)(a+PAD)) >>4) <<4 )
#define CPUID {__asm__ __volatile__ ("cpuid");}
#define MOV_MR(mem,reg) /* Move 32-bit data (mem to reg) */ \
{__asm__ __volatile__ ("mov %0,%%" #reg : : "m" (mem));}
#define MOV_RM(reg,mem) /* Move 32-bit data (reg to mem) */ \
{__asm__ __volatile__ ("mov %%" #reg ",%0" : "=m" (mem) : );}
#define MOVAPS_MR(regl,reg) /* Move aligned 128-bit data (mem to reg) */
\
{__asm__ __volatile__ ("movaps (%" #regl "),%xmm" #reg);}
#define MOVAPS_RM(reg,regl) /* Move aligned 128-bit data (reg to mem) */
\
{__asm__ __volatile__ ("movaps %xmm" #reg ",(%" #regl ")");}
#define MOVUPS_MR(regl,reg) /* Move 128-bit data (mem to reg) */ \
{__asm__ __volatile__ ("movups (%" #regl "),%xmm" #reg);}
#define MOVUPS_RM(reg,regl) /* Move 128-bit data (reg to mem) */ \
{__asm__ __volatile__ ("movups %xmm" #reg ",(%" #regl ")");}
#define MULPS(reg1,reg2) /* Multiply parallel scalars */ \
{__asm__ __volatile__ ("mulps %xmm" #reg1 ",%xmm" #reg2);}
_Tim Burgess_____________________Ba.App.Sc.(Hons) RMIT__
| InterCollegiate Network Administrator |
| Ormond College, University of Melbourne |
| Parkville, VIC 3052 Australia |
|_Mail:[EMAIL PROTECTED]: 03-9344-1212___|
------------------------------
Crossposted-To: alt.os.linux,comp.os.linux.advocacy,comp.os.misc,comp.unix.advocacy
From: [EMAIL PROTECTED] (Vladimir Z. Nuri)
Subject: Re: TAO: the ultimate OS
Date: Sat, 5 Jun 1999 09:37:56 GMT
Jimen Ching ([EMAIL PROTECTED]) wrote:
[esr cathedral vs. bazaar]
: Finally, someone said it. ;-) The fame that this document received is
: an indication of the type of audience you need to expect. Any well trained
: software engineer would see through that document after the first few
: paragraphs.
I like the general theme, and think he nicely pulled out the
basic dichotomy.. its basically "microsoft vs. linux" with
the former being the cathedral, the latter being the bazaar..
and much of the hostility to my essay is perhaps because
people are perceiving it as "cathedral like" .. I respect
ESR's publicity/proseletyzing ability and his writing is
somewhat decent.
: My suggestion to you is to grow thicker skin. There will always be immature
: people who say "show me the code". A good software engineer will see the
: flaw in this line of thinking and ignore it. Anyone who has tried writing
: complex software know that a concept must be developed first, before the
: code can be developed. You can't write something if you don't have the
: ideas in your head. And if you don't write it down, you'll forget it.
hahah I have very thick skin. very flameproof underwear.
but thanks for understanding, for one. nice to see a single
person who has a similar point of view. ah yes, now
you go and flame me over the essay. hahhaha
: 1. The first mistake you made was to umbrella the OS concept. An operating
: system is more than just what the user sees. The idea that an OS should be
: designed from the ground up to cater to the end-user is looking at only half
: the picture.
its the whole picture, man. that's the point. the end user is the
point. linux developers currently refuse to recognize this. they
think the OS designers are part of the equation. in the end
the OS creators (hackers/programmers) must be irrelevant in the face of the
end user.
: 2. As someone else already pointed out, your essay may be ahead of its
: time.
I take that as a compliment<g> I am not saying it will be implemented
tomorrow. in fact the longer it takes to pull of what I have
envisioned, the more proud I would be of the essay. we'll see
who was on target in 2-3-5 years. in the heat of ppl with
nanosecond attn span, of course it all melts into incomprehensibility.
: 3. The third problem I saw can be summed up with, "theory without data".
: This is the same problem ESR had with all his papers. There are no data
: to support any of the ideas and concepts being discussed. Examples are
: not data. For instance, do you have proof that the object file system
: design is superior to everything else?
superior how? the point is to appeal to the end user/ergonomics.
I say in the essay, some of the key design features are gonna make
it slower. just as garbage collection in java makes it slower, but
they put it in after very careful consideration and made it a key
elemet of the language..
What are the alternatives? Did
: you do a study of file system and concluded that an object design is
: better?
I contrast it to the existing system in which everything is
basically disconnected and organized only at the flimsiest
level-- free form files and directories.
: You have some good ideas. But the world is full of good ideas...
the world is full of everything imaginable. the future world
will be full of many things unimaginable.
: Good luck.
good luck 2u2
--
~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^~^
"in theory, there's no difference [EMAIL PROTECTED]
between theory and practice, mad genius research lab
but in practice there is!" http://www8.pair.com/mnajtiv/
------------------------------
From: IanP <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.setup
Subject: Re: 2.2.9 kernel too big?
Date: Sat, 5 Jun 1999 21:07:58 +1000
On 4 Jun 1999, Mohd H Misnan wrote:
> On Fri, 4 Jun 1999 13:58:22 +1000, IanP wrote:
> >
> >LILO was complaining that my newly compiled 2.2.2 kernel was too big even
> >with "make bzimage". After searching for some help I found "make bzlilo",
> >problem fixed. I don't know if its the same problem but I hope it helps.
>
> There is nothing wrong of putting some of the stuff into loadable modules and I
> don't think it'll affect the performance of your newly built kernel. Putting to
> much into the kernel itself will give the above mentioned problem.
>
Well I will if I can't fit all I desire or need into a monolithic kernel
but I can and it saves a little time.
IanP
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list (and comp.os.linux.development.system) via:
Internet: [EMAIL PROTECTED]
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Development-System Digest
******************************
