|
|
|
|
|
|
|
|
Prolog Standardization - What Now?Appeared in Volume 8/1, February 1995
rss@seg.npl.co.uk
Roger Scowen
1st November 1994
WG17 (i.e. ISO/IEC JTC1 SC22 WG17 - the working group responsible for standardizing Prolog) should complete its task to prepare an International Standard for the core of Prolog early in 1995. WG17 is also currently responsible for completing Part 2 (Modules).
What is the future for WG17?
ISO/IEC JTC1 Directives require the formation of an "Editing Group" associated with WG17 to consider and respond to any defects reported on the standard. This "Editing Group" may have nothing to do, but if it does, most of its work could be performed by e-mail.
Some features of Prolog, e.g. grammar rules, were left out of Part 1 in order not to delay its completion.
WG17 is currently relying on volunteer effort to coordinate and lead its work.
2. New Work Item proposals are prepared to standardize further features of Prolog. Some candidate features include: grammar rules, external interfaces, constraints.
3. WG17 regards its work as complete, and, apart from the "Editing Group", recommends to SC22 that WG17 be dissolved.
2. At least five countries are required to participate in New Work Items. Is there sufficient support for any further work?
3. Which additional features would benefit from standardization?
4. Are there adequate resources for subsequent standardization to be performed in a timely fashion?
rss@ditc.npl.co.uk
Roger Scowen
15th December 1994
I received various replies directly, and edited them all together for publication to WG17 (as annex D of ISO/IEC JTC1 SC22 WG17 N135 (Utrecht papers)).
WG17 discussed this topic of further standardization at its meeting in Utrecht (3rd - 5th December 1994) and decided not to take the lead until the standard is to be revised. Of course, if some standards body puts forward a New Work Item proposal, and five countries agree to participate, then any such work would almost certainly be given to WG17 to accomplish. Meanwhile, work on Part 2 (Modules) will continue with Jonathan Hodgson as the new project editor.
Finally, the good news, WG17 agreed how to deal with all the comments on the Draft International Standard (DIS), and publication of a standard for Part 1 (General Core) will be early next year.
Arthur Fleck wrote, 'It would be unfortunate to see the Modules effort abandoned - this would be an enhancement of significance to the application of the language to more ambitious projects'.
Bill McClay wrote, 'I think the work on modules is important and should continue'.
Jonathan Hodgson put the note on the agenda of X3J17 (the USA Prolog standardization panel) and they have produced a new proposal.
Mats Carlsson wrote, 'Completion is desirable. Unfortunately I have no good suggestions for improving progress'.
No one else had ideas for improving progress on Part 2 except Stan Szpakowicz who wrote: 'Part 2 in its present shape is, in my opinion, hopeless. The only way to get this done is to take the method adopted in (for example, but quite a good example) Quintus Prolog and make it the standard'.
Lee Naish wrote, 'It's going to be hard to get people to agree on a module system. Perhaps something simple that could be implemented on top of (some) existing module systems would be easiest to get through in a semi-timely fashion. It's a shame the standardization process has been so slow. At the start there were not many module systems around and the process would have been easier.'
However Tony Kusalik wrote, 'For modules [...] some work should start, but it is likely premature to set a standard ... yet'.
Arthur Fleck wrote, 'I would strongly encourage that standardization be extended to include grammar rules. This is an elegant aspect of the language, and it involves subtleties that would benefit from this effort'.
Mats Carlsson wrote, 'This is something I could consider working on, but it is not clear at this point whether SICS will let me do it on paid time.'
Peter Schachte wrote, ' I do believe that grammar rules should be included in the standard, and I wouldn't expect them to present a major problem. Broadening this to cover term expansion would be very desirable, but would undoubtedly present serious problems, so it should probably be avoided.'
Tony Kusalik wrote, 'there is sufficient experience with grammar rules that they could be "standardized".'
Lee Naish wrote, 'It's worth doing grammar rules. I can't see why it should take very long.'
Fergus Henderson correctly replied. 'the problem was that no-one could agree on what should happen with grammar rules containing non-logical constructs.'
Peter Schachte wrote, 'Standardizing the foreign interface would be very useful it if it isn't too divisive.'
Tony Kusalik wrote, 'For [...] interfacing to external languages (C/C++) some work should start, but it is likely premature to set a standard ... yet.'
Lee Naish wrote, 'It's worth doing foreign interfaces, though that may be more tricky.'
Bill McClay wrote, 'I definitely think that constraints should be addressed'.
Michael Jampel wrote, 'It is much too early to think about standardisation of constraints'. Compare the Eclipse syntax with CLP(R) - and these are systems which are pretty similar in implementation and philosophy. What about concurrent constraints vs concurrent logic programming....
Too many active research issues. Can't standardise until we all agree on what ought to be done and how we ought to do it. Neither is agreed yet for any aspect of constraints.'
Peter Schachte wrote, 'I believe it is a few years premature to consider standardizing constraints.'
Tony Kusalik wrote, 'It is definitely too early to try to standardize constraint features'.
Lee Naish wrote, 'I would forget about standardizing constraints at this stage.'
Fergus Henderson wrote, 'it's premature. But standardizing coroutining (as an optional extension) is worth at least considering. The systems that have it all seem to use incompatible syntax.'
An extensible way to do declarations (I'm not sure whats in the current draft standard, but declarations on predicates are becoming more important in lots of systems and it would be nice for these to be done in a way that doesn't prevent portability).'
Michael Covington wrote, 'Actually, in the category of `other things,' what we really need is a way to define your own evaluable functors (for use in arithmetic expressions).'
Arthur Fleck wrote, 'When I was invited to participate in the standards effort, I was unable to obtain funding for travel support, so my participation has been "passive". I believe that there are still highly useful things to be done, and I am willing to volunteer my effort to that end - for work that can be completed through e-mail, I am willing to help'.
Bill McClay wrote, 'I don't have any suggestions on how the work could be accelerated'.
Mats Carlsson wrote, 'Unfortunately I have no good suggestions for improving progress'. He also thought the resources for subsequent standardization to be performed in a timely fashion are probably inadequate.
|
|
|
|
|
|
|
|