Constraints in Prolog

Hi Paolo (and any others interested),

I have been toying with a small implementation some of the inferences/key constraints in Prolog.  It's somewhat horrible considered as Prolog (i.e. I'm doing lots of the things we tell students not to do), but I've pushed it far enough that I think it should be possible to do the whole thing, with examples of inference, key/uniqueness constraints, ordering, and typing.  It's just sort of a pain to write each inference or constraint explicitly, since it's being done explicitly using list operations, and you have to be careful about unification and termination.

Does anyone knowledgeable about Prolog have time to work on this?  It could use:

- someone to implement the rest of the inferences and constraints, following the examples I've done already
- someone to write a prov-N definite clause grammar parser (or use a Prolog XML library) to import prov-n or prov-XML to the Prolog term representation I'm using.  (Alternatively, maybe someone could modify prov-toolbox to output Prolog terms, it shouldn't be very hard.)  

I am pretty much swamped until next Wednesday so may not be able to finish it before the 31st, but hopefully can at least report partial support for inferences/constraints.

Anyone curious can see the progress so far at:

https://dvcs.w3.org/hg/prov/raw-file/default/semantics/checker.pl

--James


On Jan 24, 2013, at 3:59 PM, Paolo Missier <Paolo.Missier@ncl.ac.uk> wrote:

> Hi,
> 
> regrets, as I am tied up in marathon meetings. For the record:
> 
> - I just sent a positive reply to Tom re: PROV-DICTIONARY, so that'sa my vote
> 
> - I have added one implementation data point  by means of a questionnaire, I will add (at least) one more (consumption of PROV-N, generation of Neo4J data) which points at code that worked fine with the previous snapshot of the ProvToolbox -- need to rebuild.
> 
> - I am also going to add the version of prov-datalog that was current in summer 2012, however it's clearly incomplete. So please let me know if that's not interesting. 
>   Lack of personal cycles and resources meant that less progress was made on that than I was hoping for. 
> 
> Regards,
>   -Paolo


-- 
The University of Edinburgh is a charitable body, registered in
Scotland, with registration number SC005336.

Received on Thursday, 24 January 2013 19:17:49 UTC