Linux-Development-Sys Digest #878, Volume #6 Thu, 24 Jun 99 22:14:24 EDT
Contents:
Re: libc and a released product. (Christopher Browne)
Re: Why we are still holding on to X Windows (Christopher Browne)
Re: TAO: the ultimate OS (Christopher Browne)
Re: Why we are still holding on to X Windows (Christopher Browne)
Re: TAO: the ultimate OS (Frank Sweetser)
Re: TAO: the ultimate OS (Stefaan A Eeckels)
Re: TAO: the ultimate OS (Christopher B. Browne)
Re: TAO: the ultimate OS (Stefaan A Eeckels)
Re: TAO: the ultimate OS (Christopher B. Browne)
Re: You can now use Winmodems in Linux!!!!!!! (Christopher Browne)
Re: TAO: the ultimate OS (Stefaan A Eeckels)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (Christopher Browne)
Crossposted-To: comp.os.linux.development.apps
Subject: Re: libc and a released product.
Reply-To: [EMAIL PROTECTED]
Date: Fri, 25 Jun 1999 00:35:08 GMT
On 23 Jun 1999 13:23:27 -0700, david parsons <[EMAIL PROTECTED]>
wrote:
>What Linux frantically needs is (a) a regression test suite that
>can be run on new libraries to validate them and (b) to have the
>major distributions refuse to ship new libraries that don't
>pass the test suite.
Agreed.
There may already be something "internal" at Cygnus, whether as part
of EGCS, GLIBC, or otherwise; if there were a public test suite, that
would be, while not the most sexy thing in the world, a terribly
valuable thing to have.
The person *vastly* more likely than anyone else to know about this
would be Ulrich Drepper.
<agenda id="hidden">
What the world needs is a nearly-automated system where one may submit
snippets of C code (possibly not much more than code inside a main())
along with expected output, that should probably amount to
"pass/fail".
The snippets, along with output, would then be compiled (as needed),
run, and output compared to expectations.
You want to report a bug? Provide a test case that is expected to
output "pass," and which currently outputs "fail," and stick it in the
database.
</agenda>
There is such a scheme for EGCS, and I think there's one for GNUStep.
TOG has worked on something vaguely like this for a draft LSB
proposal, not that it is necessarily coded up to be run automagically
just yet.
There almost certainly should be one for GLIBC.
This is one of those unsexy tasks that it would probably take some
financial sponsorship to get people to work on.
--
"It's obvious that the ``Linus personality cult'' has got to go."
-- Linus Torvalds, May 5, 1999, ABC News
[EMAIL PROTECTED] <http://www.hex.net/~cbbrowne/freeecon.html>
------------------------------
From: [EMAIL PROTECTED] (Christopher Browne)
Crossposted-To: comp.os.linux.development.apps
Subject: Re: Why we are still holding on to X Windows
Reply-To: [EMAIL PROTECTED]
Date: Fri, 25 Jun 1999 00:35:27 GMT
On 24 Jun 1999 10:57:07 -0400, "Stefan Monnier
<[EMAIL PROTECTED]>
wrote:
>>>>>> Todd Knarr <[EMAIL PROTECTED]> wrote:
>> 1. Provide a generic sound server protocol independent of X11. This has the
>> advantage of also allowing non-X11 apps to use sound.
>> 2. Provide a sound protocol as an X11 extension instead of as part of the
>> basic X11 protocol.
>
>1 and 2 are not mutually exclusive. The sound protocol should be integrated
>in X just as is OpenGL.
Some *part* of the protocol would be valuable to integrate with X.
Possibly not all of it. The thing worth having integrated with X would
be the ability to synchronize screen events with sound events.
Sound is probably an example of a "lossage" that resulted directly from
the move of control of X from the X Consortium to The Open Group; there
was a protocol effort under way at the time, NAS, that essentially fell
through the cracks. Some code is available at
<ftp://ftp.x.org/contrib/audio/nas>; it is probably dramatically
obsolete, but that presumably it could have become featureful had the
process proceeded...
>>>>>> "Daniel" == Daniel Robert Franklin <[EMAIL PROTECTED]> writes:
>> Fonts could certainly be better. GNUStep's Display PostScript is probably
>> the best answer.
>
>While I disagree that X should be scraped, I completely agree that font
>support suffers from its old design from the bitmap fonts days.
>Anti-aliasing (along with access to font information) is the most
>prominent problem and is sadly not easy to solve within the existing
>protocol. It does call for X12. Of course DPS is a strict necessity.
>How is Display GhostScript doing these days ?
Apparently progressing; see <http://www.gnustep.org>, where they depend
on it.
>> Ummm, X is a networked graphics protocol. It has nothing to do with sound.
>> If you want to "add" sound support, you probably should look at one of the
>> network audio protocols.
>
>NO! X is not a `graphics protocol'. The proof is very simple: I see
>no graphics whatsoever on my mouse or on my keyboard and yet their
>support is an intrinsic part of X. X is a protocol between an application
>running somewhere and `the outside world' which used to be limited to
>keyboard/mouse/display but where the sound-card is gaining importance.
>There are network sound protocols, and I really hope that one of them (I
>couldn't care less which one (probably because I *never* use my sound card))
>will take over and get integrated as an X extension (just like the GLX
>(or is it XGL) extension for OpenGL).
Agreed.
>> So, why don't we drop the X and innovate?
>
>Waste of time. Extend X instead. The sound extension can be easily
>accommodated within X11. The font extension will be trickier, but if
>the transition is smooth enough, it should work just fine.
Hopefully XFree86 4.0, "The Holy Grail," will come out soon, and
revitalize the ability to add useful extensions.
>> a) exists
>> b) works
>> c) does things which Windows (with no expensive add-on components) can
>> only dream about.
>
>Windows is also broken by design. Mostly because of the `builtin'
>window manager.
Litestep can replace the W95 WM, which suggests that it's not *quite* as
built-in as you're indicating...
--
"Linux and other OSS advocates are making a progressively more credible
argument that OSS software is at least as robust -- if not more -- than
commercial alternatives." - Microsoft lamenting Open Source Software in the
"Halloween Document"
[EMAIL PROTECTED] <http://www.ntlug.org/~cbbrowne/xwindows.html>
------------------------------
From: [EMAIL PROTECTED] (Christopher Browne)
Crossposted-To: alt.os.linux,comp.os.linux.advocacy,comp.os.misc,comp.unix.advocacy
Subject: Re: TAO: the ultimate OS
Reply-To: [EMAIL PROTECTED]
Date: Fri, 25 Jun 1999 00:35:35 GMT
On Thu, 24 Jun 1999 21:06:15 GMT, Vladimir Z. Nuri <[EMAIL PROTECTED]> wrote:
>Peter Samuelson ([EMAIL PROTECTED]) wrote:
>[windows registry]
>: Summary, from my POV: There are many reasons not to. One of the most
>: compelling is that you create one more wonderful single point of
>: failure for your entire system. And what a point of failure it is.
>: Ever try to recover a Windoze system whose registry had gotten hosed?
>
>if you combine a registry with a good fault-tolerant system, with
>versions of the registry & all other software.. that could be pretty
>streamlined. what you are saying to me above is that halfway implementations
>of powerful ideas often leave the user worse off. I agree!! but the
>failure of a new idea is often due to poor implentation. as on of my
>design friends once told me, ppl have a very difficult time telling
>the dif between failure of new idea and failure in a particular
>implementation of that idea..
Hiving an extra layer of pointers (the registry DB) atop a layer of
pointers (the filesystem) simply cannot be made "more fault tolerant."
People seem unable to understand this, but I suppose that this is only
to be expected when people that probably don't know how to design either
a filesystem or a DBMS start thinking that because they think they know
something about "user-friendliness" that they must be competent to
design an operating system.
>: There are many other reasons. Editing things with standard tools (not
>: just editors; all the standard Unix line-oriented tools) is a *huge*
>: win.
>
>I assure you the OS of the future is not going to have text based
>configuration files .. at least at the core. I think there is always
>a need to export any "entity" into unambiguous ASCII files however.
>that would be one of the operations I would build into the OS.
>any object whatsoever can be exported/imported as a binary or
>ASCII file..
Assure away. When Oracle decide to move their configuration information
into a binary database, that will be a good evidence that they consider
that good enough.
--
"Face it, Bill Gates is a Persian cat and a monocle away from being
a villain in a James Bond movie." --Dennis Miller.
[EMAIL PROTECTED] <http://www.hex.net/~cbbrowne/lsf.html>
------------------------------
From: [EMAIL PROTECTED] (Christopher Browne)
Crossposted-To: comp.os.linux.development.apps
Subject: Re: Why we are still holding on to X Windows
Reply-To: [EMAIL PROTECTED]
Date: Fri, 25 Jun 1999 00:35:38 GMT
On 24 Jun 1999 16:28:17 +0200, Mads Dydensborg
<[EMAIL PROTECTED]> wrote:
>I would like to add "printer support". IMHO the GDI printer
>"paradigme" is really really handy.
I tend to agree. GDI may represent a very dangerous/insidious
implementation of the idea of having an intermediate representation;
this does not forcibly establish that the idea is, *in general,* wrong.
There are probably three reasonable approaches:
a) Applications use a DPS facility (e.g. - DGS), which is displayable
and readily printable.
b) Applications use a PDF facility, which is, again, displayable and
readily rendered into printable form.
*NOTE THAT THIS DOES NOT IMPLY USING PDF AS A DOCUMENT/ARCHIVAL
FORMAT.* PDF gets generated, rendered, and *thrown away.*
The advantage that PDF has over DPS is while it keeps the Postscript
rendering model, it removes the considerable complexities of requiring a
PS interpreter.
c) The XPRINT extension, where X requests get sent to an X server that
talks to printers rather than to a screen display. (Note that this one
actually does exist already.)
--
Rules of the Evil Warlord #84. "No matter how many shorts we have in
the system, my guards will be instructed to treat every surveillance
camera malfunction as a full-scale emergency."
[EMAIL PROTECTED] <http://www.hex.net/~cbbrowne/printing.html#GDI>
------------------------------
From: Frank Sweetser <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux,comp.os.linux.advocacy,comp.os.misc,comp.unix.advocacy
Subject: Re: TAO: the ultimate OS
Date: 24 Jun 1999 21:16:30 -0400
[EMAIL PROTECTED] (Stefaan A Eeckels) writes:
> In article <[EMAIL PROTECTED]>,
> [EMAIL PROTECTED] (Vladimir Z. Nuri) writes:
> >
> > another example: the end user wants to increase the space
> > for their linux partition and decrease space on their windows
> > partition. without reformatting. how many zillions of hours
> > have been spent in this operation by end users? only because
> > programmers haven't written the code to move around partitions
> > on a hard drive? a clear cut case where programmer convienience
> > was chosen at huge sacrifice to user convenience.
> This is typical for the PC architecture. Most modern UNIX systems
> do have Logical Volume Managers, which basically allow you
> to shrink and expand a partition as required (at the cost of
> a layer of indirection between the file system and the disk).
>
> Can't have it on PCs for compatibility's sake, though.
not true. work is being done even as we speak on a generic (read: non arch
specific) LVM system for linux.
--
Frank Sweetser rasmusin at wpi.edu fsweetser at blee.net | PGP key available
paramount.ind.wpi.edu RedHat 5.2 kernel 2.2.5 i586 | at public servers
> (It's sorta like sed, but not. It's sorta like awk, but not. etc.)
Guilty as charged. Perl is happily ugly, and happily derivative.
-- Larry Wall in <[EMAIL PROTECTED]>
------------------------------
From: [EMAIL PROTECTED] (Stefaan A Eeckels)
Crossposted-To: alt.os.linux,comp.os.linux.advocacy,comp.os.misc,comp.unix.advocacy
Subject: Re: TAO: the ultimate OS
Date: 24 Jun 1999 23:02:19 GMT
In article <[EMAIL PROTECTED]>,
[EMAIL PROTECTED] (Vladimir Z. Nuri) writes:
> Stefaan A Eeckels ([EMAIL PROTECTED]) wrote:
>: As for
>: other projects, how do you know they are not designed? Is there
>: somehow an obligation to troll on Usenet before starting an open
>: source project? Or should you be asked to comment on each and every
>: free project?
>
> I vehemently detest the "troll" label. none of my posts fit into
> this category. I am being labelled that merely because I am not
> backing down from defending the article.
If you'd read what I said instead of going off like a firecracker,
you'd have noticed the "troll" referred to Terry's rather weird
idea of having requirements documents for open source projects.
If you'd give the matter some thought, you'd easily see that such
documents only make sense if a formal hierarchy / organization
exists.
--
Stefaan
--
PGP key available from PGP key servers (http://www.pgp.net/pgpnet/)
___________________________________________________________________
Perfection is reached, not when there is no longer anything to add,
but when there is no longer anything to take away. -- Saint-Exup�ry
------------------------------
From: [EMAIL PROTECTED] (Christopher B. Browne)
Crossposted-To: alt.os.linux,comp.os.linux.advocacy,comp.os.misc,comp.unix.advocacy
Subject: Re: TAO: the ultimate OS
Reply-To: [EMAIL PROTECTED]
Date: Fri, 25 Jun 1999 01:24:09 GMT
On Thu, 24 Jun 1999 20:56:37 GMT, Vladimir Z. Nuri <[EMAIL PROTECTED]> posted:
>Christopher Browne ([EMAIL PROTECTED]) wrote:
>
>: As a common "for instance," in DBMS-oriented applications, I have
>: quite regularly heard users ask for the ability to modify the primary
>: key to a DB table, which is one of those basic things that by
>: necessity *cannot* be fiddled with. I *can't* modify the application
>: to allow what they want; the system *forbids* this, and that was in
>: fact the original intent.
>
>there should be no restriction on this if the new primary key
>is unique, imho. give the user whatever they want that is not
>logically impossible.
Can you say "Foreign key"? I knew you could.
>another example: the end user wants to increase the space
>for their linux partition and decrease space on their windows
>partition. without reformatting. how many zillions of hours
>have been spent in this operation by end users? only because
>programmers haven't written the code to move around partitions
>on a hard drive? a clear cut case where programmer convienience
>was chosen at huge sacrifice to user convenience.
I certainly am having a hard time counting those hours... I can't count
that low.
--
Those who do not understand Unix are condemned to reinvent it, poorly.
-- Henry Spencer <http://www.hex.net/~cbbrowne/lsf.html>
[EMAIL PROTECTED] - "What have you contributed to free software today?..."
------------------------------
From: [EMAIL PROTECTED] (Stefaan A Eeckels)
Crossposted-To: alt.os.linux,comp.os.linux.advocacy,comp.os.misc,comp.unix.advocacy
Subject: Re: TAO: the ultimate OS
Date: 24 Jun 1999 22:56:31 GMT
In article <[EMAIL PROTECTED]>,
[EMAIL PROTECTED] (Vladimir Z. Nuri) writes:
>
> another example: the end user wants to increase the space
> for their linux partition and decrease space on their windows
> partition. without reformatting. how many zillions of hours
> have been spent in this operation by end users? only because
> programmers haven't written the code to move around partitions
> on a hard drive? a clear cut case where programmer convienience
> was chosen at huge sacrifice to user convenience.
This is typical for the PC architecture. Most modern UNIX systems
do have Logical Volume Managers, which basically allow you
to shrink and expand a partition as required (at the cost of
a layer of indirection between the file system and the disk).
Can't have it on PCs for compatibility's sake, though.
--
Stefaan
--
PGP key available from PGP key servers (http://www.pgp.net/pgpnet/)
___________________________________________________________________
Perfection is reached, not when there is no longer anything to add,
but when there is no longer anything to take away. -- Saint-Exup�ry
------------------------------
From: [EMAIL PROTECTED] (Christopher B. Browne)
Crossposted-To:
gnu.misc.discuss.comp.os.linux.advocacy,comp.os.misc,comp.unix.advocacy
Subject: Re: TAO: the ultimate OS
Reply-To: [EMAIL PROTECTED]
Date: Fri, 25 Jun 1999 01:24:08 GMT
On 24 Jun 1999 13:48:52 GMT, void <[EMAIL PROTECTED]> posted:
>On 24 Jun 1999 03:10:22 GMT, Terry Murphy <[EMAIL PROTECTED]> wrote:
>>
>>Perhaps, but I very rarely see design documents or requirements
>>documents for anything that comes out of the free software community.
>
>Then you're not looking in the right places. What about soft updates?
>What about vinum, the FreeBSD volume manager? What about Coda? I'm
>beginning to suspect that you're just stating your prejudices, without
>having done any research whatsoever.
There *are* projects that provide little in the way of design documentation,
and I would agree that this is something of a problem, particularly for
systems that are intended to be useful well on into the future.
We don't have *too* many "megaprojects" going on as used to be the case that
had a *clear* mandate to produce technical specs. Such projects have tended
to be attached to academic institutions.
Notably:
- CMU: AUIS, Coda, CMU Lisp,
- MIT: Athena, which became X
- UC:BSD: BSD
(I'd welcome examples of other such "megaprojects.")
This is a clear benefit of cooperative efforts involving universities: Both
grad students and professorial staff have clear mandates to "publish or
perish," which provides the incentive to produce scholarly documentation on
such systems.
Unfortunately, there seems to be a tendancy of late for cooperative projects
to be highly proprietary, which strikes me as being extremely hurtful. It
may provide schools with some funding today, but it diminishes the
applicability of the material that they produce.
>>Not for the GNU stuff, not for the kernel stuff, and certainly not for
>>the little three day hacks that they put on freshmeat.net.
>
>Linux and linux-oriented communities are far from the be-all and end-all
>of open source computing.
There are subsets that hold *significant* discussions, mostly on mailing
lists, as Usenet is so "flame-oriented." The archives of such lists can
contain valuable documentation, and better projects will provide coherent
summaries.
I first saw the Coda docs not too long ago, and was extremely impressed.
That does, of course, represent a situation where the community benefits
from the "publish or perish" principle...
>>Well, basically all of my ideas on software engineering came from an
>>internal book by DEC about software engineering, so there's one company
>>who does it. Their internal software products follow the metholodology
>>I am talking about. Also, VMS and its layered products were quite
>>obviously designed by committee, so there's one example of a commercial
>>product.
>
>Yes ... a dead one.
I could make a comment about a successor system, but I don't think it would
go along well with the "methodology" assertions :-).
I used to work for SHL Systemhouse, and "lived" on the same floor as their
"methodology development" group. I read some of the stuff, and concluded
that most of this sort of thing is mostly fit for lining birdcages.
If it was done really well, it might be of value, but that old story about
"Those who can, do; those who can't, teach" seems to apply pretty well to
the "methodology" doctrinal material I've seen so far.
>>I've never worked for IBM, but I know several people who have, and I
>>understand their development metholdogy is even more formal. I don't
>>think there's much doubt that their mainframe software products largely
>>follow this model.
>>
>>As for others, from what I've read about Microsoft's development
>>practices, they follow a somewhat less rigid development process, but
>>still reasonably close to what I'm talking about. I also STRONGLY
>>suspect that software by companies such as Cadence, Oracle, and SAP are
>>designed formally, purely judging from the quality of the products.
>
>Criminy, you're using Microsoft as a paradigm of software engineering?
>That's truly perverse.
R/3 is one thing I have some familiarity with; they have a system called OSS
that probably represents something relevant to the free software community.
OSS is a "support" system that collects centrally bug fixes, problem
reports, and the likes.
The free software that turns out to be viable tends to be that which reacts
successfully to problem reports. There are a lot of "Linux packages" that
don't have any such feedback mechanism, and that has a tendancy to lead to
them becoming moribund.
There are a number of "database systems" to help manage this sort of
process, such as PIKT, Bugzilla, GNAT, and probably others. My thinking is
that this could represent more useful feedback if, where possible, these
systems could collect actual executable examples of problems.
Thus, if I find a bug with some function call with GLIBC, I might submit a
bug report, along with a little program in C that exercises the bug, and
which could be used on an ongoing basis to represent a regression test.
This may still, from a design perspective, be a little more like "chipping
away the parts of the stone that don't look like an elephant" than like
"building a statue of an elephant," but remember that declarative languages
like Prolog can find answers too :-).
--
Those who do not understand Unix are condemned to reinvent it, poorly.
-- Henry Spencer <http://www.hex.net/~cbbrowne/lsf.html>
[EMAIL PROTECTED] - "What have you contributed to free software today?..."
------------------------------
From: [EMAIL PROTECTED] (Christopher Browne)
Subject: Re: You can now use Winmodems in Linux!!!!!!!
Reply-To: [EMAIL PROTECTED]
Date: Fri, 25 Jun 1999 00:34:59 GMT
On Wed, 23 June 1999 01:00:31 -0400, Allin Cottrell
<[EMAIL PROTECTED]> wrote:
>James Stevenson wrote:
>
>> i am not invloded in this but has anyone asked the manufacuters yet??
>
>Of course. Their response is basically that writing winmodem
>drivers is difficult and time-consuming, and requires a lot of
>help from the chip-makers -- and they're not willing to put
>resources into helping write Linux drivers unless they see a
>BIG market there. The more people who pester, e.g., Lucent,
>the better.
Of course, there is benefit to *continuing* to bug them, albeit doing
so in a reasonably diplomatic fashion.
Consider that the correspondence that was done previously largely
occurred a year or more ago. At which point corporate folk didn't
have much reason to believe that Linux represented a significant
potential market.
Today, with big names in hardware and software having software
releases (or possibly even actual *support* for Linux), the marketing
droids may be a little less dismissive of the value of "Linux sales."
--
The Three Laws of Thermodynamics:
1) You can't win.
2) You can't break even.
3) You can't even get out of the game.
[EMAIL PROTECTED] <http://www.hex.net/~cbbrowne/hardware.html>
------------------------------
From: [EMAIL PROTECTED] (Stefaan A Eeckels)
Crossposted-To: alt.os.linux,comp.os.linux.advocacy,comp.os.misc,comp.unix.advocacy
Subject: Re: TAO: the ultimate OS
Date: 24 Jun 1999 23:14:05 GMT
In article <[EMAIL PROTECTED]>,
[EMAIL PROTECTED] (Vladimir Z. Nuri) writes:
> Stefaan A Eeckels ([EMAIL PROTECTED]) wrote:
>: Are you perchance a PHB? Don't forget that what ESR describes
>: is the "prototype -> beta -> release" part of your
>: description of a "good development", not the "requirements ->
>: design" part, which has been done by the original author
>: *before* code is released.
>
> notice you emphasize a sort of "rugged individualist"
> approach to design that doesn't involve collaboration. i.e.
> single programmer makes all the design choices early on.
> but surely collaboration can be introduced into the requirements/
> design process, with good benefits. in fact perhaps this
> would be a significant improvement worth seriously pursuing
> on the worlds next open source OS..?
It could, but it's a fallacy to think that a design gets
better when more people get involved. In fact, there are
many examples where "design by committee" leads to very
bad specifications (cfr the OSI stack, or the EDIFACT
message format).
A committee quickly suffers political problems, such as
infighting, a desire to get one's own points integrated
into the spec at all costs (usually to justify one's
involvement in Club Med turnover to one's boss :-), etc.
Come to think of it, the OSI stack is a prime example of
what happens to designs if too much is designed before implementation
is started. The spec ends up describing a basically unworkable setup
(would you believe that it took until the mid-90ies before
the OSI committee realised they should be defining APIs),
but the investment in the design is so huge (and the benefits
of being on the committee so enticing) that it is never
critically examined.
> p.s. what is a PHB
A Pointy Haired Boss --read Dilbert, you'll enjoy it.
--
Stefaan
--
PGP key available from PGP key servers (http://www.pgp.net/pgpnet/)
___________________________________________________________________
Perfection is reached, not when there is no longer anything to add,
but when there is no longer anything to take away. -- Saint-Exup�ry
------------------------------
** 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
******************************
