Is Operating Systems Research Dead?
Shinji Ikari
My conjecture is this "Operating Systems research is dead" The towel
has been thrown in. There is no excitement, no passion or drive left
in Universities, research departments, or companies to push the
cutting edge of OS development.
What has happened with OS's in recent years? Most of the OS
excitement centers around 1970's vintage operating systems such as
linux and Mac OS-X. BEOS is basically down for the count after being
bought out by Palm. Plan 9 never went anywhere. The Inferno
virtualized OS is an interesting commercial OS, but it seems to be
heavily biased towards the Unix philosophy. In this environment, its
sad to say that the most innovative OS work is being done by
Microsoft.
So what has happened? Why, except for a small minority of OS
"eccentrics", are people no longer motivated to experiment, to
explore, or to push the envelope of operating systems research?
Why have we given up?
Edward A. Feustel
My conjecture:
First, operating systems encompass to much stuff -- they are almost too
large for
a small group of individuals to do and writing drivers for all the new
devices
is incredibly challenging. I'd bet this is one reason that IBM gave up
on a
consumer version of OS/2. Instead many seem inclined to take on a
manageable section
of an existing, open source OS, e.g., Linux SE and its add-in security
features.
Second and more daunting: Few companies are writing applications. Those
that do
concentrate on a few operating systems that can produce the most
customers, viz. Windows vs. Apple
and Linux. They don't want to maintain additional versions for small
customer bases.
Ed Feustel
--
Edward A. Feustel Research Associate
efeu...@ists.dartmouth.edu Institute for Security Technology Studies
phone - (603)646-0671 Cybersecurity Research Group
fax - (603)646-0660 Dartmouth College, Hinman Box 6226
http://www.ists.dartmouth.edu Hanover, NH 03755
Yoann Padioleau
shi...@swirve.com (Shinji Ikari) writes:
> My conjecture is this "Operating Systems research is dead" The towel
> has been thrown in. There is no excitement, no passion or drive left
> in Universities, research departments, or companies to push the
> cutting edge of OS development.
>
> What has happened with OS's in recent years? Most of the OS
> excitement centers around 1970's vintage operating systems such as
> linux and Mac OS-X. BEOS is basically down for the count after being
> bought out by Palm. Plan 9 never went anywhere. The Inferno
> virtualized OS is an interesting commercial OS, but it seems to be
> heavily biased towards the Unix philosophy. In this environment, its
> sad to say that the most innovative OS work is being done by
> Microsoft.
What microsoft innovation are you talking about ?
>
> So what has happened? Why, except for a small minority of OS
> "eccentrics", are people no longer motivated to experiment, to
> explore, or to push the envelope of operating systems research?
There is a paper from rob pike which is called
operation system research is irrelevant: http://www.cs.bell-labs.com/cm/cs/who/rob/utah2000.ps
you should read it.
>
> Why have we given up?
>
>
--
Yoann Padioleau, INSA de Rennes, France, http://www.irisa.fr/prive/padiolea
Opinions expressed here are only mine. Je n'écris qu'à titre personnel.
**____ Get Free. Be Smart. Simply use Linux and Free Software. ____**
Shinji Ikari
Return-Path: <ne...@google.com>
Received: from rutgers.rutgers.edu (rutgers.edu [165.230.4.76])
by ftp.cse.ucsc.edu (8.9.3/8.9.1) with ESMTP id BAA28299
for <o...@ftp.cse.ucsc.edu>; Fri, 24 Aug 2001 01:11:50 -0700 (PDT)
Received: from smtp3.google.com (smtp3.google.com [216.239.33.28])
by rutgers.rutgers.edu (8.8.8/8.8.8) with ESMTP id EAA10380
for <comp-os-...@moderators.isc.org>; Fri, 24 Aug 2001 04:11:49 -0400 (EDT)
Received: from exins2.exo.google.com (exins2.exo.google.com [10.4.0.8])
by smtp3.google.com with ESMTP id BAA11663
for <comp-os-...@moderators.isc.org>; Fri, 24 Aug 2001 01:11:15 -0700
Received: from sjmm36.sj.google.com (sjmm36.sj.google.com [10.6.39.36])
by exins2.exo.google.com (8.9.3/8.9.3) with ESMTP id BAA30264
for <comp-os-...@moderators.isc.org>; Fri, 24 Aug 2001 01:11:15 -0700
Received: (from news@localhost)
by sjmm36.sj.google.com (8.9.3/8.9.3) id BAA29013
for comp-os-...@moderators.isc.org; Fri, 24 Aug 2001 01:11:15 -0700
To: comp-os-...@moderators.isc.org
Path: not-for-mail
From: shi...@swirve.com (Shinji Ikari)
Newsgroups: comp.os.research
Subject: Re: Is Operating Systems Research Dead?
Date: 24 Aug 2001 01:11:15 -0700
Organization: http://groups.google.com/
Lines: 28
Message-ID: <8cd70836.0108...@posting.google.com>
References: <3b8446bf$1...@news.ucsc.edu> <3b85770a$1...@news.ucsc.edu>
NNTP-Posting-Host: 208.168.16.141
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 8bit
X-Trace: posting.google.com 998640675 29012 127.0.0.1 (24 Aug 2001 08:11:15 GMT)
X-Complaints-To: groups...@google.com
NNTP-Posting-Date: 24 Aug 2001 08:11:15 GMT
Status: R
Content-Length: 1144
"Edward A. Feustel" <efeu...@ists.dartmouth.edu> wrote in message news:<3b85770a$1...@news.ucsc.edu>...
> My conjecture:
>
> First, operating systems encompass to much stuff -- they are almost too
> large for a small group of individuals to do and writing drivers for
> all the new devices is incredibly challenging.
....
Yes, it is challenging. Yes, 3rd party software vendors may not write
software for it. Yes, these are certainly reasons why few groups, if
any, are embarking on developing full-fledged commercial operating
systems.
But these are only some *reasons* why operating systems resarch may be
dead; it is implying that if it is difficult and there will be little
commercial benefit then its not worth doing.
But surely there are reasons beyond the purely commercial for pushing
for OS research to be revived.
Or have we really come to the end of the road. Is it like building
chip fabs; its just to big, too complex, and just too expensive for
anyone except huge corporations to even consider experimenting with
novel operating system concepts?
So, my conjecture still holds, is operating systems research really
dead?
Shinji Ikari
Yoann Padioleau <Yoann.P...@irisa.fr> wrote in message news:<3b857710$1...@news.ucsc.edu>...
> There is a paper from rob pike which is called
> operation system research is irrelevant: http://www.cs.bell-labs.com/cm/cs/who/rob/utah2000.ps
> you should read it.
Interesting paper; it is in total agreement with my conjecture. Rob
Pike's conclusions are worth repeating here:
"It has reached the point where I doubt that a brilliant systems
project would even be funded, and if funded, wouldn't find the bodies
to do the work. The odds of success were always low; now they're
essentially zero."
"The community, universities, students, industry, funding bodies must
change its priorities."
"The community must accept and explore unorthodox ideas."
"The community must seperate research from market capitalization."
Any comments on Rob's conclusions?
John Jensen
Yoann Padioleau <Yoann.P...@irisa.fr> writes:
: shi...@swirve.com (Shinji Ikari) writes:
: > So what has happened? Why, except for a small minority of OS
: > "eccentrics", are people no longer motivated to experiment, to
: > explore, or to push the envelope of operating systems research?
: There is a paper from rob pike which is called
: operation system research is irrelevant:
: http://www.cs.bell-labs.com/cm/cs/who/rob/utah2000.ps
: you should read it.
Thanks for the link. FWIW, I've started to put together a little page
as I figure out what is going on (or not going on) with operating
systems research:
http://members.home.net/jjens/FooOS/weblog1.html
So far it is a two day old weblog ;-), but if all goes well I'd like to
link to interesting papers and articles as they appear on the web.
I am a working programmer (for twenty years now), but with a degree in
chemistry and some computer science courses to fill in the gaps. It is
interesting for me now to go back and read textbooks I didn't see then
... to see how well theory matches the practices I've seen.
I'd appreciate any links you can throw me as time goes on.
John
--
33° 47' 36N 117° 54' 51W
Edd Doutre
Shinji Ikari wrote:
>My conjecture is this "Operating Systems research is dead" The towel
>has been thrown in. There is no excitement, no passion or drive left
>in Universities, research departments, or companies to push the
>cutting edge of OS development.
>
>What has happened with OS's in recent years? Most of the OS
>excitement centers around 1970's vintage operating systems such as
>linux and Mac OS-X. BEOS is basically down for the count after being
>bought out by Palm. Plan 9 never went anywhere. The Inferno
>virtualized OS is an interesting commercial OS, but it seems to be
>heavily biased towards the Unix philosophy. In this environment, its
>sad to say that the most innovative OS work is being done by
>Microsoft.
>
Actually, some of the 'research' being done at Microsoft appears to be
documenting
minor improvements on fairly well known technology and applying for
patents. In some
research for a patent a year or two ago, I noticed several patents in
the area of filesystems
that seemed to involve FAT and NTFS and various well known operations
such as formatting,
disk partitioning and related. The claims were intriguing... I felt
then and still do that it would
be difficult at best, if not nearly impossible, to do research in those
areas without infringing on
some of the claims.
Now, before I am accused of Microsoft-bashing, several other companies
have also filed
for lists of patents in certain areas that 'lock-up' research and
innovation in those areas. It
reminds me of a game of Go, where a winning strategy is to form a 'wall'
across one's
opponent's line of expansion.
Do you suppose that the recent years of IP (Intellectual Property)
conscious corporate
research/development has stifled OS research in some areas? Everyone I
know in
corporate employ seems to be quite aware of the bonuses offered by their
employers
for patentable IP. Some companies' records are impressive in this regard.
Frank Miller
As evidence that OS Research is truly dead, I offer the 2001 SOSP
program, http://www.cs.ucsd.edu/sosp01. This is arguably the
preeminent conference in OS design. I count 15 papers from some of
the biggest names in OS research. I'm still looking for something
interesting. Well, I guess there is the study of errors, oh, and
theres the one on power management. I just cant wait until this
program becomes available. I want to read everything, cover to
cover...
FM
--
Frank W. Miller
Cornfed Systems Inc
www.cornfed.com
Warren W. Gay VE3WWG
Shinji Ikari wrote:
I think he made some good points (apart from the M$ bias). However,
when I looked at:
http://www.cs.arizona.edu/people/bridges/oses.html
this clearly demonstrates that there has been a lot of effort
expended at various times by various people. I think some of the
most interesting work is currently the exokernel related stuff,
combined possibly with the Opal O/S work.
So while it might be true that there "is not enough" of this going
on, there does seem to be a fair amount of research activity going
on. Each of these projects represent considerable effort.
I suspect that our overal impression about the demise of O/S
research, is actually an impatience with the ongoing efforts--
we all want it to happen faster. I will agree that there is
considerable duplication of effort in some of the projects
listed there.
Thats my $0.02 worth, anyway. ;-)
--
Warren W. Gay VE3WWG
http://members.home.net/ve3wwg
.
Markus Peuhkuri
"Edward A. Feustel" <efeu...@ists.dartmouth.edu> writes:
> too large for a small group of individuals to do and writing drivers
...
> Second and more daunting: Few companies are writing applications.
I think that is very much the point. OS is between applications and
hardware: it must be able to communicate in both directions.
>From Rob Pike's presentation 2000-02-21:
New operating systems today tend to be just ways of
reimplementing Unix. If they have a novel architecture (and
some do) the first thing to build is the Unix emulation layer.
How can operating systems research be relevant when the
resulting operating systems are all indistinguishable?
Before your OS is for any use, it must have drivers and applications
-- and making/adpating them will take much longer than innovating new
kernel -- if you don't emulate.
--
Markus Peuhkuri ! http://www.iki.fi/puhuri/
Alex
What about creating small operating systems for educational purposes?
I believe this is an excellent platform for students to grasp the
concepts of operating systems in a practical manner. There are plenty
of struggling OS programmers out there sweating over boot sectors and
tiny kernels, and some results are promising. Is this group not
interested in discussing the progress in that area?
I find digging into the machine very interesting. I know you can also
study the existing Linux kernel, for example, but I believe it would
be wise to always broaden your view beyond the horizon of this one
operating system, which also began from scratch (as a better Minix).
Developments like SkyOS, ShagOS, RedPants (and many, many more) are
worth monitoring in my opinion.
Kind regards,
Alex van Oostenrijk
vanoos...@hotmail.com