Re: [9fans] I've got 4 student intern (undergraduates) here.

Re: [9fans] I've got 4 student intern (undergraduates) here.

[okamoto]

[-- Attachment #1: Type: text/plain, Size: 688 bytes --]

When I tried to show it to our students, yes, I also felt it's not easy to do.
However, as I know Plan 9 is very easy to maintain system and users, and
to write programs, I started to do it by reading papers such as "Plan 9 from
Bell Labs" etc. with them.

Today, I asked Yoshitatsu what was his first impression of Plan 9 when he
touched, and his answer is that it was "then, so what?". :-)  I understand it,
because Plan 9 doesn't have "killer application" except of acme. which is
only for programmers or system maintainers.  However, he is writing Plan 9
application now.  However, this is not accepted by many of Japanese _SOFTWARE_
companies, if any... (sigh)

Kenji


[-- Attachment #2: Type: message/rfc822, Size: 2031 bytes --]

From: Laura Creighton <lac@cd.chalmers.se>
To: 9fans@cse.psu.edu
Cc: lac@cd.chalmers.se
Subject: [9fans] I've got 4 student intern (undergraduates) here.
Date: Fri, 8 Jun 2001 11:30:27 +0200 (MET DST)
Message-ID: <200106080930.LAA04591@boris.cd.chalmers.se>

I want to make a `this is plan 9' presentation to them.  Can I get notes
and stuff from you?  Also, what works?  What will impress them?  And
yes i _have_ the papers from 1127, the stuff I want to steal from you
is what you did with that.

Thanks very much,
Laura Creighton

Laura's theme (Long) (Was: [9fans] I've got 4 student intern (undergraduates) here.)

[Lucio De Re]

Apologies for taking Laura Creighton's name in vain :-)

On Mon, Jun 11, 2001 at 02:13:56PM +0900, okamoto@granite.cias.osakafu-u.ac.jp wrote:
> 
> When I tried to show it to our students, yes, I also felt it's not easy to do.
> However, as I know Plan 9 is very easy to maintain system and users, and
> to write programs, I started to do it by reading papers such as "Plan 9 from
> Bell Labs" etc. with them.
> 
My enthusiasm was certainly fired by the papers, it was a long time
after I discovered the first set of Plan 9 documents that I was given
a chance to look at it.

> Today, I asked Yoshitatsu what was his first impression of Plan 9 when he
> touched, and his answer is that it was "then, so what?". :-)  I understand it,
> because Plan 9 doesn't have "killer application" except of acme. which is
> only for programmers or system maintainers.  However, he is writing Plan 9
> application now.  However, this is not accepted by many of Japanese _SOFTWARE_
> companies, if any... (sigh)
> 
Based on my somewhat narrow view of the computer field in the last few
years, there have been very few killer apps: Lotus 1-2-3 (a better
Visicalc, I should imagine), the web browser, MS Office (in an odd
sense); underrated, but susprisingly significant (in my opinion, of
course) are Quake and Microsoft Outlook.

Quake is to Linux what 1-2-3 was to the IBM Personal Computer.
Without it, I believe, Linux would still be as much of a curiosity as
Plan 9 is today.

Outlook is important because it is practically transparent.  Most
users are unlikely to know whether it is the Office version or
Express as delivered with Internet Explorer.  All Outlook users seem
to believe that everyone else is also using Outlook, a bit like all
automobile drivers assume that all other automobile driver knows what
a steering wheel is.  This insidious penetration is Microsoft's
strongest marketing ploy.

Plan 9 has too many of these aces up its sleeve.  Like Linux, it
should concentrate all its efforts on a single one, but of course,
like with Linux, it is its community that defines the direction
development takes.

Without excluding myself from the criticism, the Plan 9 culture is
inherently elitist, rather than evangelical.  I'm about to embark on a
cross-host debugging session (Inferno, rather than Plan 9, I'm not
sure how successful I'm likely to be) and only when confronted by this
need, did I realise that Laura's question about impressing students
should have been answered by a cross-host task, the more convoluted,
the better.  No one else has proposed such an idea, and yet the main
strength of the fileserver approach lies clearly in that quarter.

"We hold these truths to be self-evident: ..."  I really love that
phrase.  Plan 9's strengths are so self-evident, even we can't see
them anymore :-)

++L

PS:  The jist of it all, is that ACME is Plan 9's killer app.  If we
could turn a Plan 9 system into a cross-platform (hardware _and_
Operating System platform, that is) development tool, it would
doubtlessly eclipse the current IDEs.  Porting Limbo (Alef) to the
popular platforms (thank you, Bell-Labs and Vitanuova for Inferno,
it's a pity we have to ask for more), with Acme, RC, the Plumber and
MK, would be a great start.  But these are not market oriented dreams,
they address a more ideal than realistic world.

Re: [9fans] Laura's theme (Long) (Was: [9fans] I've got 4 student intern (undergraduates) here.)

[Theo Honohan]

On Monday 11 June, Lucio De Re wrote ("Laura's theme (Long) (Was: [9fans] I've got 4 student intern (undergraduates) here.)"):
> 
> Without excluding myself from the criticism, the Plan 9 culture is
> inherently elitist, rather than evangelical.  I'm about to embark on a
> cross-host debugging session (Inferno, rather than Plan 9, I'm not
> sure how successful I'm likely to be) and only when confronted by this
> need, did I realise that Laura's question about impressing students
> should have been answered by a cross-host task, the more convoluted,
> the better.  No one else has proposed such an idea, and yet the main
> strength of the fileserver approach lies clearly in that quarter.

Absolutely!  Well said.  I spent several futile hours at the weekend
trying to articulate exactly that point.

If the students have encountered the pseudo-network-transparency of X11
or NFS, they'll immediately appreciate the advantages of the Plan 9
approach.  It's all about just presenting the three main design
principles listed in the overview paper.  Then, you show that the Plan 9
approach solves all of the same problems as X, ssh X-forwarding, NFS and
VNC, in a much cleaner way.  That should appeal to any CS student.  Then
you show how it goes way beyond that, allowing fine-grained cross-host
sharing of resources.  Russ's anecdote about Windows and ssh is a good
example: 
http://lists.cse.psu.edu/archives/9fans/2000-October/007937.html

Not everyone will immediately recognize the need for that kind of
contortion, but people who have developed software across multiple
machines, or have run their own small network of Linux boxes, will be
sent into transports of delight.

I'm less convinced that Acme is a good selling point; people are
unlikely to see a new editor or IDE and think "Oh, I want to change
editors!"  Habits are hard to break, and editors are a matter of
personal taste (as are mouse chords...)

Theo

Re: [9fans] Laura's theme (Long) (Was: [9fans] I've got 4 student intern (undergraduates) here.)

[Boyd Roberts]

> If the students have encountered the pseudo-network-transparency of X11
> or NFS, they'll immediately appreciate the advantages of the Plan 9
> approach.

as i recall, the returning NFS guru's from Sun (have they served an
TM injunction on the owner of that star yet?), when i was working for
_the instruction set_, set about to teach us about it, in an internal
course.

i must have been 23 at the time and as soon as i heard UDP mentioned
i asked questions along the lines of 'are they mad'?  i was told not
to worry -- hah, hah.  real worry set in when they asked me to port
the kernel in 7 weeks (16 weeks being the current VAX -> 68020 port
speed of the 2.0 release and the '20 was far from stable -- all
sorts of glue to get it to take its faults/lumps correctly).

well it didn't take too long, once i realised that it was just a
in-line diff-ing of similar sources driven by someone who knew both
kernels, the odd machine dependent glop and byte order problem,
followed by a good thrashing to death against a Sun to see if it
worked -- it didn't.  all the world's a vax so some of XDR was just
hosed.  fix that, thrash it, play chess -- we're _ahead_ of schedule.

i think this is the 'storm trooper' school of coding.

maybe 'starship trooper' school if you're a yes fan:

    shining, flying, purple wolfhound show me where you are

much like those XDR bugs.

but the line is from 'yours is no disgrace'.

Re: [9fans] I've got 4 student intern (undergraduates) here.

[Borja Marcos]

On Monday 11 June 2001 07:13, you wrote:
> When I tried to show it to our students, yes, I also felt it's not easy to
> do. However, as I know Plan 9 is very easy to maintain system and users,
> and to write programs, I started to do it by reading papers such as "Plan 9
> from Bell Labs" etc. with them.

	Plan 9 is a distributed system. It looks especially designed for 
heterogeneous networks, the kind of real-world applications you can find. 
When I read the first paper about the design of Plan 9 and its way to 
accomodate heterogeneous networks, with both high-speed local networks and 
slow-speed dialup accesses, it was quite obvious that a similar design (and 
it was the Inferno design goal) would be excellent for a service provider 
such as a cable modem network (or xDSL) who deploys "network computers" 
instead of simple set-top boxes.

	Other distributed OSs (Amoeba, for example) are designed with local area 
networks in mind, using the network as a bus for a processor pool if I 
oversimplify it. Plan 9, on the other hand, allows the developer and network 
architect distribute the applications taking into account the available 
bandwidth in each of the networks involved.

	In a real world network, the bandwidths can be as low as 33600 bps (a pstn 
modem) and as high as a gigabit. Apart from the elegant design, I think it 
can be the most interesting feature for many people.



	Borja.

GL again (was Re: [9fans] I've got 4 student intern (undergraduates) here.)

[Theo Honohan]

On Monday 11 June, Borja Marcos wrote:
>
> 	In a real world network, the bandwidths can be as low as 33600
> bps (a pstn modem) and as high as a gigabit. Apart from the elegant
> design, I think it can be the most interesting feature for many
> people.

Speaking of gigabit, and the idea of supporting distributed OpenGL, I
just noticed some recent work from the Stanford "Mural" people, which
describes a new OpenGL stream protocol, called WireGL.  It's intended to
be a portable and scalable replacement for GLX, with a view to cluster
rendering.  So, that might be interesting to look at, given that they've
already done all the hard work...

http://graphics.stanford.edu/software/wiregl/research.html

Re: [9fans] I've got 4 student intern (undergraduates) here.

[pac]

Hi,
I am a computer self -made-man, and will, probably, be pretty green in CS forever (I am a researcher in biology).
However, I am _so_ impressed by Plan 9, and I'll tell you why:

i)   it shifts the paradigm  from kernel to a universal protocol
ii)  everything is a file
iii)  it is _clean_ as it  consistently uses i and ii,
iv)  it was designed from scratch with an _idea_ in mind
v)   the creators were not afraid (unlike many others) to throw away anything from the old stuff that seemed garbage
vi)  it is _elegant_, AFAI can judge, (cf., e.g., how elegantly it got rid of the X11, sic!)
vii)  it is minimalistic, I mean that everything is implemented carefully with respect to possible overhead, instead of
       "just be satisfied, boy, that the app works, and don't bother how (and why :-)
I have some previous experience as a user/programmer under the following "OSes" (strating in mid-eighties): TI programmable calculator :-)
-> CP/M -> DOS -> Windoze -> NEXTSTEP -> OpenStep -> GNU-Linux . I repeat I am a _greenhorn_ in CS,
but take this as a "look from outside".
Hope Plan 9 will remain clean, elegant, and open in the future (to remain also free would be nice, at least for us `Easties').
Thanks for it.

Cheers,
Peter.

--
Peter A. Cejchan
Dept. Paleobiology, Inst. Geology Acad. Sci.,
Rozvojova 135, Prague 6
CZ-16502 Czech Republic
<cej@cejchan.gli.cas.cz>

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
A plea:
Please, consider your support to the Public Library of Science initiative at
http://www.publiclibraryofscience.org

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Re: [9fans] I've got 4 student intern (undergraduates) here.

[pac]

ix)  it is modular
x)   believe me or not, I like it's UI (actually, originally I was attracted to Plan 9 by 9wm of Linux...but in 
those days Plan 9 was not free and open source) - it is so different and minimalistic, and I was impressed


--
Peter A. Cejchan
Dept. Paleobiology, Inst. Geology Acad. Sci.,
Rozvojova 135, Prague 6
CZ-16502 Czech Republic
<cej@cejchan.gli.cas.cz>

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
A plea:
Please, consider your support to the Public Library of Science initiative at
http://www.publiclibraryofscience.org

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Re: [9fans] I've got 4 student intern (undergraduates) here.

[Axel Belinfante]

> x)   believe me or not, I like it's UI (actually, originally I was attracted 
> to Plan 9 by 9wm of Linux...but in 
> those days Plan 9 was not free and open source) - it is so different and mini
> malistic, and I was impressed

Same here:
 sam -> reading about help, and being impressed -> wily -> acme,
 and of course 9term  :-)
(I still remember the day when I in 'automatic non-thinking editing mode'
 I used an editing chord in my mail program (exmh) - did not work, sigh ;-))
Axel.

Re: [9fans] I've got 4 student intern (undergraduates) here.

[pac]

xi)  Unicode
xii)  system-wide cut and paste


--
Peter A. Cejchan
Dept. Paleobiology, Inst. Geology Acad. Sci.,
Rozvojova 135, Prague 6
CZ-16502 Czech Republic
<cej@cejchan.gli.cas.cz>

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
A plea:
Please, consider your support to the Public Library of Science initiative at
http://www.publiclibraryofscience.org

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~