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

[Axel Belinfante <Axel.Belinfante@cs.utwente.nl>]

> i've posted that before, but...
> 
> here's my p9 presentation for an undergraduate project i did last year,
> html-ed off of powerpoint:
> 
> http://homepage.usask.ca/aam396/P9BL.html

I tried to grab this using wget (for off-line browsing) but did not succeed.
Would it be possible to package the whole thing in a gzip-ed tar file or so?

On the topic of spreading the word:
I will be trying to convince my colleages about some of the nice
features of plan9/inferno - in particular the advantage (abstraction)
that you may get by using a file system as interface.
The group I work in works on 'formal methods' both on theory and tools;
the tools usualy consist of components that somehow interface which
each other, and using file-system like interfaces seems very attractive.
As proof-of-concept I made my first limbo program to 'unfold' (construct)
the labelled transition system for a given automaton (finite state machine)
step by step by repeatedly reading 'successor' files, and the result is
that 'on top of that' we would be able to write a 'simulator'
in a shell script. This turned out to be pretty convincing to the few
coleagues I showed this so far. 

Now (one of) the questions that probably will come up
(rephrased: that I hope to come up :-)
is how we can integrate (profit from) the inferno/plan9 niceties
(ideas, implementation environment) in our current development
environment (sun/solaris, pc/linux), without comitting us completely
to plan9 (I have running it on left-over hardware, I would be delighted
but pretty surprised if I could convince people here to run plan9 on
the desktop - most of our desktops are now sun; drawterm works)
Foreseeable constraints are that:
on the one hand we need still to be able to connect to legacy tools
that live in the unix world and cannot be ported (e.g. because they
depend on libraries for which we don't have source)
and offer e.g. a C api on which we want to build,
and on the other hand we want to still be able to distribute the
tools we develope in such a way/shape that others can use them -
and they likely run unix, not plan9.
My first guess would be that developing in inferno/limbo could do
the trick w.r.t. distribution, and that we could deal with legacy
tools by building file-system interface wrappers around them.

Any ideas, comments?
Axel.