The future of Plan9?

The future of Plan9?

[Digby]

Steve's comments:
>
>> [Plan 9's] main failing,
>> in my opinion, is its inability to interact with other operating
>> systems conveniently.
>
>I don't think the other operating system is a problem, because
>most users aren't concerned with it. They just use it for kicking
>off their main applications. The interchange of data is the important
>thing.
>
I should have said that this was the biggest problem for me. I don't
know how others are using it. When I kick off my main applications,
I need to be able to interact with them. Having found hp(1), this is
now just a problem for graphical apps. And it is still a problem
accessing and running plan 9 apps from an external non-plan9 system.

>
>> At least the
>> wealth of quality free software available on the net makes developing
>> alternate operating systems less of a daunting task than it used to
>> be - only the core has to be build, and a quite usable suite of
>> applications can be readily ported by a user community.
>
>Huh? If I understand you, you're saying that one can bash out a
>new kernel, and then chuck all the standard freebie tools on top
>of it. Ok, so this is handy for researchers into OS development,
>but apart from that, what does it give you that you haven't
>already got? [Bzzzt! Warning: heading off-topic]
>
That is mostly what I had in mind. I used to tinker with
operating systems when I was at University, and the thing
that always seemed to be the obstacle to making a usable system
was the thought of having to write all the compilers, editors
and utilities from scratch. I had access to Unix sources, but nothing
that could be redistributed.  Now at least it is only a porting
effort. But it only gets you the level of funtionality of plan 9,
certainly not what you need to compete with M$ in commercial
applications :-/


Regards,
DigbyT
-- 
Digby R. S. Tarvin                                              digbyt@acm.org
http://www.users.dircon.co.uk/~cthulhu/

The future of Plan9?

[Adam]

Alex you do not know me but I saw your post in the plan9 group. I have
a question maybe you can answer it. When will the Cisco 12000 be out?
Also on a 2505 when you do show version and it does not show any flash
memory is that true meaning there is no flash and no IOS or is it a
lie and there actually is simm memory and IOS hiding somewhere?

Adam Miller
adamm@adamm.net
www.adamm.net

On Tue, 29 Apr 1997 23:37:56 GMT, in comp.os.plan9 you wrote:

>As a long-time friend of Plan9 and someone who purchased the Plan9
>license back in '95, I found all this discussion from a couple of
>weeks ago very interesting.
>
>I personally have been very intrigued by Plan9 ever since I first
>heard about it and was hoping to be able to spend some time with the
>system but never really had the chance to do so.
>
>All this talk about "What's happening with Plan9?" and "How about
>Brazil?" left me wondering what the community really is hoping to get
>out of these developments. From Bell Labs' eh, I mean Lucent's
>perspective, things are pretty clear. Their primary focus right now is
>a commercially viable system that will bring in some revenue. Plan9 is
>not it and Brazil won't be it either. That's were Inferno comes in and
>anybody who hangs out at the Inferno mailing-list knows, most
>discussion there has been about commercial aspects of the system
>lately.
>
>So, the question is: Where to from here? It appears to me that the
>people who do stuff with Plan9 (mostly academic and hobbyists) can't
>really get a grip on what it is they want out of it. Is it a learning
>system? A home system? A commercial system? Whatever it is, it
>attracts people from different parts of the World with different
>interests. People who, for some reason, don't feel like tinkering
>around with *BSD, Linux or commercial OS's.
>
>OK, now, let's have some discussion here as to where you want to go
>with Plan9 (reference to Microsoft commercial unintentional ;-)
>Personally, I am hoping that there are people out there who are seeing
>a real opportunity in Plan9 as a great way for hobbyists to get
>involved again in what used to be dominated by them and is now
>completely industialized. What I am referring to are the early days of
>the microcomputer revolution where people had a real sense of communal
>belonging and it was the hobbyist who pushed the technological
>envelope.
>
>Before I start going off on some weird techno-sociological tangent, I
>think I should stop here and open the floor for discussion.
>
>Absolutely not speaking for my employer....
>
>> > I, too, am curious about the disposition of Brazil.  I actually got
>> > some funding approved for Plan 9 here at MIT (I'm part of a student
>> > ...
>> > Is there an `official' answer?  Is there an unofficial answer from
>> > someone who has put more effort than I am into trying to get an
>> > official answer?
>> =20
>> here's an unofficial answer.
>> =20
>> we've been working on other things for the past year or so,
>> so there has been little new work on brazil.  phil and rob
>> and dave say that they want to get back to working on
>> brazil, but we don't know when that will be.
>>=20
>> our plan 9 system is only used to access the old worm for archival
>> purposes.  our new file server contains only the brazil source
>> tree, so brazil is our current development system.  brazil is not
>> in a state that is releasable, so i doubt that there will be
>> a brazil release in the foreseeable future.
>>=20
>> of course, all of this could change in a wink.
>
>--
>Alex Bochannek                 Phone & Fax : +1 408 526 51 91
>Senior Network Analyst         Pager       : +1 408 485 90 92
>Engineering Computing Services Alpha Pager : (800) 225-0256 PIN 104536
>Cisco Systems, Inc.            Email       : abochannek@cisco.com
>170 West Tasman Drive, Bldg. E Pager Email : abochannek@beeper.cisco.com
>San Jose, CA 95134-1706, USA

The future of Plan9?

[Eric]

>> The central idea of Plan9 with a *single* unified structure for all
>> architectures is intensely appealing.
>Plan 9 is not the only operating system that runs on multiple
>architectures.  Solaris, Linux, NetBSD, and OpenBSD all run on
>multiple targets, and I suspect NetBSD runs on more of them than Plan
>9.

True, but one is still left with a bunch of unix systems, rather
than a single unified _structure_ that encompasses all the
architectures, not just being able to build from the same 
source tree.  I can't easily keep a single tree with objects
for all architectures available. Cross-compilation environments
in the past have been clumsy to maintain (for me at least); 
maintaining more than 1 target would be horrendous.

>The price in itself is damaging but not fatal; a lot of good OS
>development effort comes out of universities, where finding $350 to
>buy a Plan 9 distribution is easy.  But why would I want to spend my
>time working on a proprietary system when I can work on my pick of
>freely redistributable Unix systems?

Indeed.  The freenix systems are pretty stable, particularly
x86; stable enough that there are some brave souls that use
it in production.  That is, however, just another unix box
on the net somewhere... there's basically no hassle-free
integration above the single-unit level.

>I'm not happy with the state of the OS world today, but if Plan 9
>wants to be considered as a step in the right direction, it has to
>either have real commercial backing or be free.  From my opinionated
>point of view, AT&T's lawyers and management consigned Plan 9 to the
>permanent status of "interesting curio" when they set the distribution
>policy.  Maybe Inferno will go somewhere else.

I agree.. unfortunately Inferno doesn't solve the problems I need
to cope with, and nothing anywhere eases my administrative and 
software burdens without causing more pain that it's worth.  Sigh.
All that horsepower going to waste.

Sincerely,

Eric Dorman
University of California at San Diego
Department of Radiology
edorman@tanya.ucsd.edu

The future of Plan9?

[Brandon]

I can confirm that this is the correct approach for chroot-ing a network
service under Solaris.  See Sun's paper on setting up anonymous FTP under
Solaris (or was it a paper by a third party?  Its a paper by someone.. :)

They explicitly state that for anonymous ftp to work if it chroot()'s to
ftp's home directory, you must have certain files from the root filesystem
copied into the chroot() environment, including many, but not all, of what
you list below.  What you have down there, also, will only support UDP,
which I assume is fine for u9fs... you have to get /dev/tcp and the
related streams devices, etc.. to support tcp (which anon ftp needed).

Brandon


On Sun, 4 May 1997, Markus Friedl wrote:

> > file system directly" approach? And I haven't made u9fs work with chroot()
> > on the Solaris machine. To boot over the net I belive I need the chroot().
> 
> in order to make u9fs+chroot work under Solaris, I had to copy some
> files to the chroot-tree: (i did a strace and filtered the 'open' system
> calls, I'm not sure if you need all files)
> 
> 	/dev/zero
> 	/dev/ticotsord
> 	/dev/udp
> 	/etc/passwd
> 	/etc/nsswitch.conf
> 	/etc/netconfig
> 	/usr/lib/nss_compat.so.1
> 	/usr/lib/nss_nis.so.1
> 	/usr/lib/libnsl.so.1
> 	/usr/lib/libintl.so.1
> 	/usr/lib/libmp.so.1
> 	/usr/lib/libw.so.1
> 	/usr/lib/libsocket.so.1
> 	/usr/lib/straddr.so
> 
> you need chroot unless you want to mix the plan9 & the solaris
> file tree.
> 
> -markus
> 


/*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
 * Brandon Lee Black     * SysAdmin/Developer      *
 * F+E Applications Dept * WorldCom, Inc.          *
 *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
 *                Office * 281.362.6466            *
 *            Voice Mail * 800.362.1141 x6466      *
 *       Electronic Mail * photon@nol.net          *
 *                       * photon@gnu.ai.mit.edu   *
 *                       * vis_blb@unx1.shsu.edu   *
 *                       * brandon.black@wcom.com  *
 *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
 *          Numeric Page * 800.940.9797 x7012263   *
 * Alpha Page (Operator) * 800.940.9595 x7012263   *
 *   Alpha Page (E-mail) * 7012263@mobilecomm.net  *
 *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*/

The future of Plan9?

[Markus]

> file system directly" approach? And I haven't made u9fs work with chroot()
> on the Solaris machine. To boot over the net I belive I need the chroot().

in order to make u9fs+chroot work under Solaris, I had to copy some
files to the chroot-tree: (i did a strace and filtered the 'open' system
calls, I'm not sure if you need all files)

	/dev/zero
	/dev/ticotsord
	/dev/udp
	/etc/passwd
	/etc/nsswitch.conf
	/etc/netconfig
	/usr/lib/nss_compat.so.1
	/usr/lib/nss_nis.so.1
	/usr/lib/libnsl.so.1
	/usr/lib/libintl.so.1
	/usr/lib/libmp.so.1
	/usr/lib/libw.so.1
	/usr/lib/libsocket.so.1
	/usr/lib/straddr.so

you need chroot unless you want to mix the plan9 & the solaris
file tree.

-markus

The future of Plan9?

[Bengt]

Greetings,

This is just to keep you posted on my new email address. Due to various
reasons I still ahven't got my Sparc up and running. No working PC
ethernet makes installation difficult. Do you know a "CD-ROM to Solaris
file system directly" approach? And I haven't made u9fs work with chroot()
on the Solaris machine. To boot over the net I belive I need the chroot().
Correct?

-- 
Best Wishes, Bengt

Email: bengtk@damek.kth.se

The future of Plan9?

[Digby]

Greg Hudson wrote:
>
>The price in itself is damaging but not fatal; a lot of good OS
>development effort comes out of universities, where finding $350 to
>buy a Plan 9 distribution is easy.  But why would I want to spend my
>time working on a proprietary system when I can work on my pick of
>freely redistributable Unix systems?
>
That is a good point. 

I think $350 is a pretty reasonable price for a supported commercial
operating system, with source. But it is a lot for a proprietary system
which is effectivly unsupported and relies on volunteers for all future
development.

My guess is that the problem is that Bell Labs wouldn't want to make it free
for fear of it becoming competition for Inferno. And they don't want
to do more work on it because they see it as having been superceeded
by Inferno. 

Does anyone feel they can comment on how similar inferno is to plan9?

My fear would be that Bell Labs might split their potential user
base between two different operating systems, and fail to get the
critical mass needed for commercial viability on either.

If it really is just an evolution of plan9, then maybe the best option
would be offer upgrades to Inferno for Plan9 licensees, at a reduced
upgrade price (to cover Bell Labs costs). It couldn't hurt Inferno's
chances to boost the supply of technically knowledgeable poeple
working on it.

Regards,
DigbyT
-- 
Digby R. S. Tarvin                                              digbyt@acm.org
http://www.users.dircon.co.uk/~cthulhu/

The future of Plan9?

[Greg]

> I have to make changes to individual machines remotely then reboot
> them rather than making _one_ change on _one_ fileserver.

If you have to reboot the machines after making the change, then
presumably making the change on a file server wouldn't help.  But
ignoring that inconsistency, you're quite right that scalability is a
big problem with Unix systems, and it doesn't get any better with
Microsoft or Apple products.

> The central idea of Plan9 with a *single* unified structure for all
> architectures is intensely appealing.

Plan 9 is not the only operating system that runs on multiple
architectures.  Solaris, Linux, NetBSD, and OpenBSD all run on
multiple targets, and I suspect NetBSD runs on more of them than Plan
9.

> I think a real big problem with 'hobbyist involvement' is (as
> rcannings(?) points out) is that $350US is a bit of a chunk for a
> hobbyist to plunk down for the distribution.

The price in itself is damaging but not fatal; a lot of good OS
development effort comes out of universities, where finding $350 to
buy a Plan 9 distribution is easy.  But why would I want to spend my
time working on a proprietary system when I can work on my pick of
freely redistributable Unix systems?

I'm not happy with the state of the OS world today, but if Plan 9
wants to be considered as a step in the right direction, it has to
either have real commercial backing or be free.  From my opinionated
point of view, AT&T's lawyers and management consigned Plan 9 to the
permanent status of "interesting curio" when they set the distribution
policy.  Maybe Inferno will go somewhere else.

The future of Plan9?

[Eric]

On Tue, 29 Apr 97 16:28:28 PDT, Alex Bochannek <abochann@cisco.com> wrote:

>So, the question is: Where to from here? It appears to me that the
>people who do stuff with Plan9 (mostly academic and hobbyists) can't
>really get a grip on what it is they want out of it. Is it a learning
>system? A home system? A commercial system? Whatever it is, it
>attracts people from different parts of the World with different
>interests. People who, for some reason, don't feel like tinkering
>around with *BSD, Linux or commercial OS's.

  I personally have tired of fighting the old battle with
"standard" unix configurations.  Unless I clog up the network
with weird nfs hierarchies (which our netcom guys will certainly
bitch about) I have to make changes to individual machines 
remotely then reboot them rather than making _one_ change on
_one_ fileserver.  And what about multiple architectures?  I've 
got Sun-Solaris, Sun-SunOS, SGIs, HPPAs, and (god forbid) NT4.0 
boxes.  All these machines have to be coherent all the time, as 
the people who use them are basically computer illiterate.  Ghod, 
it's alot of work.  The central idea of Plan9 with a *single* 
unified structure for all architectures is intensely appealing.

  Unfortunately we cannot use Plan9 since it lacks support
for the hardware we have (weird multihead framebuffers, Sun 
Sparc5s and Ultras, etc) and we can't afford the initial 
time/cost expenditure to make it work.  We can't even get
the specs for the two-headed framebuffers we use.  Godamn
proprietary nonsense.
 
>OK, now, let's have some discussion here as to where you want to go
>with Plan9 (reference to Microsoft commercial unintentional ;-)
>Personally, I am hoping that there are people out there who are seeing
>a real opportunity in Plan9 as a great way for hobbyists to get
>involved again in what used to be dominated by them and is now
>completely industialized. What I am referring to are the early days of
>the microcomputer revolution where people had a real sense of communal
>belonging and it was the hobbyist who pushed the technological
>envelope.

  I think a real big problem with 'hobbyist involvement' is
(as rcannings(?) points out) is that $350US is a bit of a chunk
for a hobbyist to plunk down for the distribution.  They've
got to be a *serious* hobbyist :)  The real, working, *BSd
and Linux distributions are free (ignoring the old "You get what
you pay for" adage 8) ) and are much more attractive, not to
mention more familiar.

  What would be _very_ interesting would be using Plan9 in an 
academic lab in CS or EE, where Plan9 would provide the platform
upon which experiments in (may favorite) network engineering
could be done.  This exposes many more people to Plan9 that would
otherwise be stuck with a unix experience, and allow more people
to work on it without having to individually eat the $350 charge.
Plan9 is hamstrung in a single-machine mode; in a network lab 
Plan9 can really shine. 

Eric Dorman
Department of Radiology, Thornton Hospital, UCSD
edorman@ucsd.edu

The future of Plan9?

[Berry]

>>>bwong@arf.cisco.com said:
 > That is true.  When I worked at Bell Labs in 1985, they were leasing
 > lots of hp2621 terminals.  There was an hp2621 emulator for the Blit
 > and I expect the Plan9 version to be a descendant.
 
In the early 80's the hp2621 was one of the best ASCII terminals you could 
get.  No, one of the best *terminals*, there wasn't much else! The -P model 
even had a built-in thermal printer.

My company (not Xerox, then) bought about 20 of them to go with our spankin' 
new hot-stuff PDP-11/70 with PWB 1.0.

Sigh.  The Sparc-10 on my desk has 96 Mbytes and that 11/70 had just 4, yet it 
could support 30-40 users and I can drive the SParc-10 into thrashing all by 
myself.  

  --berry "and we had to chisel the ones out of stone, too, and we *liked* it..."

Berry Kercheval :: kerch@parc.xerox.com :: Xerox Palo Alto Research Cent
er

The future of Plan9?

[Tom]

On May 1,  4:54am, Digby Tarvin wrote:
> I wonder what the story is behind that emulation being chosen?
> Could it be just that that is what the developers had in their
Got it in one.

-- 
Tom Duff

The future of Plan9?

[bwong]

Digby Tarvin <digbyt@acm.org> wrote:
| Strange you should mention that - I noticed it in the docs
| for the first time only the night before last. Trouble is,
| I don't know anything about the hp2621, and had assumed it
| was something esoteric like the textronix that is often provided
| (eg in xterm) but I have never seen used...
| 
| But now that I look, there is an entry for it in my
| termcap file :-)
| 
| I wonder what the story is behind that emulation being chosen?
| Could it be just that that is what the developers had in their
| labs....??

That is true.  When I worked at Bell Labs in 1985, they were leasing
lots of hp2621 terminals.  There was an hp2621 emulator for the Blit
and I expect the Plan9 version to be a descendant.

--Bruce F. Wong		bwong@cisco.com

The future of Plan9?

[Steve_Kilbane]

digby's comments:
> I believe Unix's role has changed from experimental system
> (it used to be small, consistant, well written and reasonably portable)
> to the only surviving viable commercial alternative to the
> Microsoft juggernaught..

But then, that's always been Unix's role: the valiant, ultimately
unsuccessful prime challenger to the currently perceived market
standard. IBM, DEC, and now Microsoft have taken their turn as
the Evil Empire, but interestingly, Unix is still hanging on as
the main contender.

> [Plan 9's] main failing,
> in my opinion, is its inability to interact with other operating
> systems conveniently.

I don't think the other operating system is a problem, because
most users aren't concerned with it. They just use it for kicking
off their main applications. The interchange of data is the important
thing.

> I still live in hope that, because Microsofts offerings are SO poor,
> a sufficently good alternative may one day supplant it.

This is quite likely. The depressing thing is, the alternative will
probably also be owned by Microsoft. :-(

> At least the
> wealth of quality free software available on the net makes developing
> alternate operating systems less of a daunting task than it used to
> be - only the core has to be build, and a quite usable suite of
> applications can be readily ported by a user community.

Huh? If I understand you, you're saying that one can bash out a
new kernel, and then chuck all the standard freebie tools on top
of it. Ok, so this is handy for researchers into OS development,
but apart from that, what does it give you that you haven't
already got? [Bzzzt! Warning: heading off-topic]

The availability of a free system helps, but not as much as it
should. There's been an on-going battle in comp.databases.oracle.*
to get Oracle to support Linux. Apart from the "only an idiot would
run their business on a [spit!] free system" attitude, many say Linux
provides too many configuration combinations for Oracle to support.
Perhaps, but Oracle could just select and approve a given distribution.
They could even ship it themselves. But they don't.

Plan 9's easier than Linux, because there's a lot less flux and a lot
less variance between any two systems. It's also much, much stranger to
your average application programmer on the inside. Ok, so there's APE,
but that misses the boat a little, philosophically speaking.

Your average user will look at Plan 9, and ask, "Does it run foo",
where foo is their primary tool, and most likely produced by Microsoft.
You're unlikely to be able to answer positively in a million years,
but if you can exchange *data* with foo and an equivalent (read: better)
app on Plan 9, you've got one foot in the door. My current .sig claims,
someone naively, that if you standardise on protocols and file formats,
the applications will take care of themselves. Plan 9 is great for
interfacing at a protocol/file format level, because of the user-mode
servers, and because files need not be passive animals.

Perhaps there's a marketing trick here. I get incensed when Internet
Explorer slyly offers to make itself the default browser, bumping
other software out of the way, niggling at the subconcious. Perhaps
a Plan 9 application could do something similar: "By the way,
Microsoft's changed their file formats *again*. Do you still trust
your old files to work?"

steve

The future of Plan9?

[Digby]

>
>Digby Tarvin <digbyt@acm.org> writes:
>| But as someone else pointed out, I think the developers were
>| trying to make a point by deliberately excluding cursor
>| addressing 
>
>True, but see "man hp".
>
Strange you should mention that - I noticed it in the docs
for the first time only the night before last. Trouble is,
I don't know anything about the hp2621, and had assumed it
was something esoteric like the textronix that is often provided
(eg in xterm) but I have never seen used...

But now that I look, there is an entry for it in my
termcap file :-)

I wonder what the story is behind that emulation being chosen?
Could it be just that that is what the developers had in their
labs....??

Regards,
DigbyT
-- 
Digby R. S. Tarvin                                              digbyt@acm.org
http://www.users.dircon.co.uk/~cthulhu/

The future of Plan9?

[Scott]

Digby Tarvin <digbyt@acm.org> writes:
| But as someone else pointed out, I think the developers were
| trying to make a point by deliberately excluding cursor
| addressing 

True, but see "man hp".

The future of Plan9?

[rsc]

I want to address a couple supposéd shortcomings raised
by Digby.  He writes that it's main inconvenience is its
inability to communicate with other operating systems
effectively.  This is half correct.

I have found that Plan 9 communicates particularly well
with any operating system, and that this is a strength.
Through u9fs, I can access any Unix system through
a simple and well documented protocol.  Ftpfs, dossrv, and the 
other file servers are good ideas as well.  A couple days ago,
I `operated' on a broken DOS floppy in a Linux floppy
drive using a Plan 9 system 15 miles away via u9fs because
all the good disk editing/looking utilities I have are on 
Plan 9.

At the heart of all this is the fact that the file servers
all operate in user space and are more easily hacked at.
People have written file servers that handle Linux e2fs
file systems, BSD's FFS, and others.
File system clients are a lot easier to do under Plan 9
than under, say, Linux, where everything has to be done in 
the kernel.

I've found that Plan 9 is the best operating system for
accessing other systems.  I will concede that accessing
Plan 9 *from* other systems is harder than it
need be.  I do a lot of mail-reading and occasional programming
telnetted into Plan 9 from a 80x25 text-mode Linux box at
school, and wouldn't mind seeing a `vi' or something else small
and cursor-addressed, even if the vt100 codes were hard coded.

I also wouldn't mind seeing 9x revived.  I would do it
myself except I don't use X at all anymore.
I expect to be using X a lot in a couple of months, and
I might get sufficiently annoyed to do it then.  

Perhaps the best way to do 9x is to write one that runs
under Inferno?  Just a thought.

Russ

The future of Plan9?

[Lucio]

***
Just an essay on Plan 9 and Inferno; if you prefer exclusively 
technical information this is best skipped.  My apologies to those on 
the list who find the following irrelevant or irritating - mail me 
personally if you wish to castigate me, please do not waste any 
additional bandwidth.
***

Whereas I agree with Digby that Plan 9 failings seem to be in the realm 
of compatibility with other existing operating systems, I think this is 
merely an unintentional smokescreen and that Plan 9's future will 
strictly remain in the hands of a few hobbyist.

I am forced, quite unwillingly, to accept that Microsoft have 
successfully used the market's lack of sophistication to entrench their 
brand of mediocrity, and I have to grudgingly acknowledge that their 
superior marketing skills have sufficed to position them in a very 
solid monopolistic base.

On the other hand, the public domain Unix products (Linux first, but 
FreeBSD, NetBSD and others as well) are providing the hobbyist with 
workable alternatives and in many instances serious production use is 
being made of them.  Ironically, the traditionally Unix suppliers 
(notably Sun and AT&T, but DEC to some extent and that oddball company 
known as IBM) are showing signs that they appreciate the erosion the 
free Unixes are making in the Microsoft effort.  To expand, Inferno, 
Java, Netscape, all run on Linux and might readily be ported to the 
*BSD realms - these "application" are valid adversaries to Microsoft's 
own products and have already caused Microsoft to change from a trend 
setter (at least as seen by the unsophisticated public) to a very 
obvious "me too" organisation.

So where does Plan 9 fit in?  I think the problem is more one of "what 
applications can I run on Plan 9?".  The sad part is that the Plan 9 
license makes it clear that the answer is "none".  Were I to develop 
the ultimate alternative to MS-Word on Plan 9, I could at best hope to 
find a sympathetic ear within Lucent before considering taking it to 
the marketplace.  In such a situation, it is very difficult to dedicate 
some serious effort to such a goal.

I do believe that it was expedient for Bell Labs to lay down the rules 
as they did, and I think Plan 9 would not be what it is had commercial 
interests been more influential in its development, but I also believe 
that the innovations introduced by the Plan 9/Brazil/Inferno 
combination could go a long way to supplant MS-things in the 
marketplace.  But a "killer" application (MS-Word was for Windows what 
1-2-3 had been for the IBM PC) is essential to this end.

I think it matters little which of Plan 9, Brazil or Inferno is going 
to be the platform on which such a "killer" application can be 
developed; any of them would be superior to existing platforms, 
although my bets would go with Inferno.

On the other hand, Lotus could deliver 1-2-3 because they developed on 
a moderately inexpensive platform: I have no idea how expensive the 
Inferno licence might be, but it would be a lot better if I could walk 
into my nearest computer store and for less than a hundred dollar (US) 
I could buy a replacement for or supplement to Windows-95 that made it 
possible, for example, for me to develop a "program", say, to exchange 
music, interactively, with my friend across town.

Obviously, I have no such killer application in mind, but the right 
marketing approach, inclusive of a sensible price tag (the ability to 
download Inferno is all very well, but I think better marketing would 
be to charge a nominal fee and use the "profits" to publicise the 
product) might make Limbo the new Turbo Pascal and put Inferno firmly 
on the map of alternatives to Windows 95.  The talent for such 
development is out there and the time certainly seems ripe for exciting 
innovations, but the tools available (Java, Visual Basic, Visual C++) 
are still bound to an obsolete way to do application development.  In 
my mind, Limbo is sufficiently different to encourage original thinking.

Unfortunately, Plan 9 cannot do this, having never reached a quality 
control stage and having been orphaned before it became a commonplace, 
if rough tool.  But what I have seen of Inferno seems a very reasonable 
offspring of Plan 9, its greater market-oriented maturity, unless I'm 
much mistaken about it, bodes well for its future.  AT&T have a sad 
reputation for failing to market ideas that could have been very 
successful, let's hope that Lucent can break with this tradition :-)

Note: I believe that Plan 9 benefited greatly by not being encumbered 
with market-oriented objectives.  It is my opinion and cannot be proven 
or refuted.  Inferno now has sound foundations as a result, so it is 
left to the marketing forces to shape it further.  It would be a 
vindication of my beliefs if it could succeed on the merit of such 
foundations rather than on an aggressive and probably misguiding 
advertising campaign.  Whether marketing people can appreciate this and 
build on Inferno's strengths is left to be seen.
-- 
Lucio de Re (lucio@proxima.alt.za)
Disclaimer: I'm working at getting my opinions to agree with me.

The future of Plan9?

[Digby]

My 2c..,

I believe Unix's role has changed from experimental system
(it used to be small, consistant, well written and reasonably portable)
to the only surviving viable commercial alternative to the
Microsoft juggernaught..

As a result, Unix has bloated, and become subject to numerous standards,
so that even though sources for various versions are readily
available, it doesn't lend itself radical experimentation.
I don't begrudge it that - it was the only way those of us
wanting a commercially supported operating system could be
spared the horrors of the Microsoft monopoly. The latter having
proved how obsolete technology with a huge marketing budget
will win every time in a largely unsophisticated market.

So I see Plan 9 as a refreshing attempt to try to improve
technology, and experiment with new ideas, without worrying
about commercial viability. It probably isn't the 'operating
system for the masses', because while it can be used on a
standalone machine, it is very much a second class configuration.
And without the networking, a lot of its benefits (and some fairly
basic essentials, like security) are not realised.

It has a lot of great features, which I won't list because I
am sure I would be preaching to the converted. Its main failing,
in my opinion, is its inability to interact with other operating
systems conveniently. It is a bit like my understandingo of NT, 
which is multi-user if all the users have their own machine running
NT or MSWindows. Without X clients or support for cursor
addressing, it is very difficult to access and use from a non 
plan9 host, or to access a non-Plan9 host from Plan9. I know
this could be easily addressed by a tercap/terminfo port to
the APE environment, plus (with a bit more Plan9 experience) an
ANSI terminal emulation for 8.5. Adding to this the much mentioned,
but not available, X client libs and 8.5 server, would make it a much
less isolated system.

Computers used to be fun back in the days when operating systems
were as diverse as computer architectures.

I still live in hope that, because Microsofts offerings are SO poor,
a sufficently good alternative may one day supplant it. At least the
wealth of quality free software available on the net makes developing
alternate operating systems less of a daunting task than it used to
be - only the core has to be build, and a quite usable suite of
applications can be readily ported by a user community.

In summary, I don't see it catching on in the commercial market,
but I would like to see it develop an active user community of
technical people interested in operating systems, porting existing
applications and experimenting with new developments., 

Regards,
DigbyT
-- 
Digby R. S. Tarvin                                              digbyt@acm.org
http://www.users.dircon.co.uk/~cthulhu/

The future of Plan9?

[Borja]

> 
> 
> <just pretend I quoted the whole thing>
> 
> I can only imagine what it would feel like to bind namespaces under
> Solaris, or to use remote CPU resources to compile software without
> leaving my local environment on an SGI.... 

	I see Plan 9 as a research system, just like Amoeba,
Sun's Spring... in Plan 9 (and later, Brazil) they have tried new
concepts. Now, in Inferno, you can see those concepts applied to a commercial
OS.

	I think that if they have released Plan 9 as they did, they wouldn't
mind if other OS developers used the new facilities introduced with Plan 9.

	Are those concepts (namespace manipulation, etc) covered by any patents?


	Borja.


-- 
***********************************************************************
Borja Marcos			* Internet: borjam@we.lc.ehu.es
Alangoeta, 11 1 izq		*	    borjam@well.com
48990 - Algorta (Vizcaya)	*           borjamar@sarenet.es
SPAIN				* CompuServe: 100015,3502
***********************************************************************
NOTE TO BULK EMAIL SENDERS: When I receive an unsolicited advertisement,
I take some time to send complaints not only to your service
provider, but to the next level providers who serve your ISP. 

The future of Plan9?

[Brandon]

<just pretend I quoted the whole thing>

I have an almost clear definition in my mind of what _I_ wish would come
of Plan 9/Brazil.  Unix is entrenched in the business world, and Unix will
probably not go away for a very long time.  I am a unix person, so please
don't take my bashing wrong.....  Unix has fallen over the years... most
of the commercial vendors (Sun, SGI, IBM, DEC, ....) have concentrated on
GUI's, Systems Admin tools, and hardware speedups in recent versions.  The
commercial unix world has very miserably failed to put forth new operating
system concepts, like those introduced by Plan 9.  Now, Brazil would never
get any market share as a *nix replacement, because it is not compatible
enough or supported enough (I can see commercial buyers saying: Where's
C++?  Where's the normal *nix include files?  There's no Sybase or Oracle
server for this? etc.....).... 

However, The concepts learned from Brazil could be deployed into current
*nix systems.  Lucent could license out the Brazil source code for
incorporation into the commercial *nix's.... or could incorporate it into
Reasearch Unix... or add some plan9-ish-ness to the successor of the
SysVR4 definition..... or some combination thereof...

I can only imagine what it would feel like to bind namespaces under
Solaris, or to use remote CPU resources to compile software without
leaving my local environment on an SGI.... 

my $.02

Brandon

The future of Plan9?

[Alex]

As a long-time friend of Plan9 and someone who purchased the Plan9
license back in '95, I found all this discussion from a couple of
weeks ago very interesting.

I personally have been very intrigued by Plan9 ever since I first
heard about it and was hoping to be able to spend some time with the
system but never really had the chance to do so.

All this talk about "What's happening with Plan9?" and "How about
Brazil?" left me wondering what the community really is hoping to get
out of these developments. From Bell Labs' eh, I mean Lucent's
perspective, things are pretty clear. Their primary focus right now is
a commercially viable system that will bring in some revenue. Plan9 is
not it and Brazil won't be it either. That's were Inferno comes in and
anybody who hangs out at the Inferno mailing-list knows, most
discussion there has been about commercial aspects of the system
lately.

So, the question is: Where to from here? It appears to me that the
people who do stuff with Plan9 (mostly academic and hobbyists) can't
really get a grip on what it is they want out of it. Is it a learning
system? A home system? A commercial system? Whatever it is, it
attracts people from different parts of the World with different
interests. People who, for some reason, don't feel like tinkering
around with *BSD, Linux or commercial OS's.

OK, now, let's have some discussion here as to where you want to go
with Plan9 (reference to Microsoft commercial unintentional ;-)
Personally, I am hoping that there are people out there who are seeing
a real opportunity in Plan9 as a great way for hobbyists to get
involved again in what used to be dominated by them and is now
completely industialized. What I am referring to are the early days of
the microcomputer revolution where people had a real sense of communal
belonging and it was the hobbyist who pushed the technological
envelope.

Before I start going off on some weird techno-sociological tangent, I
think I should stop here and open the floor for discussion.

Absolutely not speaking for my employer....

> > I, too, am curious about the disposition of Brazil.  I actually got
> > some funding approved for Plan 9 here at MIT (I'm part of a student
> > ...
> > Is there an `official' answer?  Is there an unofficial answer from
> > someone who has put more effort than I am into trying to get an
> > official answer?
>  
> here's an unofficial answer.
>  
> we've been working on other things for the past year or so,
> so there has been little new work on brazil.  phil and rob
> and dave say that they want to get back to working on
> brazil, but we don't know when that will be.
> 
> our plan 9 system is only used to access the old worm for archival
> purposes.  our new file server contains only the brazil source
> tree, so brazil is our current development system.  brazil is not
> in a state that is releasable, so i doubt that there will be
> a brazil release in the foreseeable future.
> 
> of course, all of this could change in a wink.

--
Alex Bochannek                 Phone & Fax : +1 408 526 51 91
Senior Network Analyst         Pager       : +1 408 485 90 92
Engineering Computing Services Alpha Pager : (800) 225-0256 PIN 104536
Cisco Systems, Inc.            Email       : abochannek@cisco.com
170 West Tasman Drive, Bldg. E Pager Email : abochannek@beeper.cisco.com
San Jose, CA 95134-1706, USA

The future of Plan9?

[Digby]

>
>I want to address a couple supposC)d shortcomings raised
>by Digby.  He writes that it's main inconvenience is its
>inability to communicate with other operating systems
>effectively.  This is half correct.
>
>I have found that Plan 9 communicates particularly well
>...
>I've found that Plan 9 is the best operating system for
>accessing other systems.

Just to clarify, what I find difficult is running applications
on other operating systems from Plan9, not accessing the file
system. I agree that file sharing is done superbly
in Plan9, and I particularly like the way things like
ftp are integrated into the file system.

Of course there is no fundamental reason why Plan 9 can't
be made to provide better support for logging in to
non Plan9 hosts, so this is not a criticism of the operating
system. It is just an omission from the distribution which
I find limiting.

But as someone else pointed out, I think the developers were
trying to make a point by deliberately excluding cursor
addressing and X support, which in a research system I guess
they should be free to do..

>I also wouldn't mind seeing 9x revived.  I would do it
>myself except I don't use X at all anymore.
>I expect to be using X a lot in a couple of months, and
>I might get sufficiently annoyed to do it then.  
>
>Perhaps the best way to do 9x is to write one that runs
>under Inferno?  Just a thought.
>
Does anyone know what became of 9x? I had a look, but couldn't
find even a partially completed implementation to work on.

Has anyone compiled the latest X client libraries in
the APE environment? I suppose these would be a starting
point to a quick and nasty XTerm. I assume the server
would be a more formidable task, and that is really what
would be needed to solve the Plan-9 to Unix execution
requirement.

Regards,
DigbyT
-- 
Digby R. S. Tarvin                                              digbyt@acm.org
http://www.users.dircon.co.uk/~cthulhu/