[CoFI] CASL finalization - STATUS

["Peter D. Mosses" <pdm@daimi.aau.dk>]

Bernd wrote:

> The fact hat there has been little work on finalization (regarding
> almost no reactions) is a little disturbing. I guess everybody has been
> very busy (as I have), but we must get on with it!

I hope to be able to release a DRAFT CASL v1.0 Summary in
mid-September, including a list of all proposed changes 
(relative to v0.99).  There is to be a 14-day period for 
comments and objections to the proposed changes.

> I do know that the semantics group is very active in updating the
> semantics and thus checking the feasability of recent changes, as
> requested by Appendix F (Finalization). This is obviously the most
> important feedback.

That feedback will be incorporated in the above document.  Those
working on the formal semantics are encouraged to send messages
explaining any proposals for significant language design adjustments
to this mailing-list AS SOON AS POSSIBLE!

I gather that some further restrictions on datatype declarations may
be proposed, to ensure that the semantics of free datatype
declarations is the same as that of the corresponding free extension.
An attempt is currently being made to find a simple way to formulate
these restrictions.

> ALL OF YOU: please read the latest version of the CASL Summary carefully
> and react until the 
> 	19th June
> with any further comments you may have. This does NOT mean that I want
> to encourage you to propose any changes (the time for this is long
> gone).

Apart from Bernd's own message proposing a different form of
conditional term, and the discussion of views (and imports), there has
been agreement on an adjustment of the relationship between compound
identifiers and subsort inclusions.  It appears that everyone is
perfectly satisfied that the rest of the CASL design is now optimal...

> However, (cf. Appendix F), some "remaining points need to be
> settled", and details to be decided upon as they are tentative:
> 	views

As I proposed in a message that I submitted yesterday to this list, 
I think that we should make a final effort to find a solution more
compatible with the OBJ3 treatment of views than the current proposal
(which was outlined in my message of 23 July).  But time is short...

> 	architectural specifications

The updated Methodology Note M-4 now provides examples.  Till has
provided extensive feedback in a message on cofi-methodology on 23
July; perhaps the authors of M-4 would like to send a message to
cofi-language, indicating whether or not any adjustment of the
language design is required here.

> 	specification libraries

Till has drafted a study note on the issue of version control, which
should be finalized and released for general comment a.s.a.p.

> 	annotations

The main unresolved question here concerns parsing (precedence, etc.)
annotations and their relationship to attributes of operations.

> Moreover, feedback from
> 	semantics

see above

> 	parsing

As far as I know, the prototype parsers (bottom-up, top-down, and
non-deterministic) have now been extended to the full v0.99 concrete
syntax.  Would those responsible please advise me of any remaining
problems a.s.a.p.?

> 	formatting

I am halfway through polishing the CASL package for LaTeX and
Hyperlatex.  This should be ready by mid-September.

> is needed; when can we expected a reaction??
> 
> I know that I still have to produce the minutes of the last meeting;
> coming up soon [in fact, its conclusions have already be incorporated in
> the last version of the summary by Peter].

Real soon now?  :-)

> NB: the chances for a CoFI WG, funded by the EC, look good; I hope this
> inspires new enthusiasm in all of you!!

As previously announced on cofi-list, we have indeed been successful
with our application, and CoFI WG is planned to start at the beginning
of October, 1998, for 2 years.  It will be nice if CASL v1.0 can be
finished by that time !

> Best regards
> Bernd

Peter

PS  I'll be in Pont-a-Mousson, France, most of the first week of
September, attending the Workshop on Rewriting Logic and Applications.
I hope to read e-mail from there, at least once a day.