Re: Agenda for Methodology Task Group Meeting in Cachan

[Giuseppe Scollo <scollo@turing.dipmat.unict.it>]

Bonjour Michel (and cofi-methodology friends),

  Here's a first bit of reaction to Michel's questions; details are deferred
to the discussion at the meeting.

> b) The expected level (style) of explanation should be carefully designed
> too. E.g. do we write a User's Manual for people having already a good
> experience in algebraic specification, or do we rather merge both algebraic
> specification and CASL user's manual ?
> c) Suggestion of relevant user's manual that we could look at (and perhaps
> choose as example/model) is welcome. Concrete ideas are needed !

  With programming languages, I found it most convenient to have three things:

 1. Tutorial
 2. Reference Manual
 3. Quick Reference

Two examples of 1+2: CAML-Light, Perl. (Perl also has 3, in the form of a very
handy _depliant_, useful to "real" users who need rapid memory-refresh about
syntax matters and the like.) Then, reader's good experience with algebraic
specification could be assumed for the Reference Manual, whilst it should not
for the Tutorial.

> Please consider seriously the above issues. If you intend to assist to the
> meeting, then I would appreciate if you could come prepared with a few
> concrete suggestions. If you do not intend to attend the meeting, please
> send me directly your suggestions, which are wellcome as well.

  My sending of this message does not imply that I do not intend to attend
the meeting, of course I do!

> We should manage to have a draft version of the CASL User's Manual ready
> for the next ETAPS conference.

  Great! Then it would be ready for my next course in Verona too ;-)

> Please send me asap any suggestion for the agenda of this Cachan meeting.

  Here's one (maybe not so urgent, but pieces of art often take long to
get into final form):

   * www shapes of tutorial

(besides an obvious HyperLaTeX text, wouldn't it be nice to have something
more _animated_?)

> Proposals for presentations are wellcome, provided they help us to make
> concrete progress.
> 
> Note: The idea of producing notes on some specific topics is not abandoned
> (see the minutes of previous meetings in the archive, and please read again
> the existing notes), but since the production of the user's manual is now
> felt as being more urgent, these notes will be either postponed or
> integrated in the user's manual. Any reaction on this issue is wellcome as
> well.

  The relationship between subsorting and inheritance issues seems very
relevant to the usability of CASL for OO software development. Technical
discussion on this topic is relevant to the _contents_ of a good tutorial
rather than to its _structure_, and I understand the latter takes priority
at the forthcoming meeting. Yet, once a structure will be chosen, it will
be necessary to reach agreement about its contents, which seems not at all
obvious for the topic in question. Maybe at Cachan subtasks should be
defined for `critical methodology topics' such as this one (and any other
which would be identified as such, of course)? Then, here's a proposal for
yet another item on the agenda:

   * identification of critical methodology topics

and (short) presentations would be in order here, if aimed at explaining
why a certain topic is felt as 'critical' with respect to the user manual,
e.g. because of expected problems with agreeing on its contents.

  Avec mes meilleurs amities,
                              Pippo
_____________
Dr. G. Scollo
 _______________________________ __________________________________
| Caalma B.V.       Consultancy |              Universiteit Twente | 
| use of abstract algebraic and |         Fac. Informatica, SE_TKI | 
| logical methods in automation |     e-mail: scollo@cs.utwente.nl |
|_______________________________| http://www.cs.utwente.nl/~scollo |
 niet veel is goed, goed is veel __________________________________|