Re: PROV-ISSUE-465 (avoid-infinite-inferences): Avoid infinite loops in inferences [prov-dm-constraints]

Hi,

To (begin to ) address the issue and hopefully also satisfy requests from reviewers for better intuition about how normalization/validity checking/equivalence checking work, I have added an outline of the validation process (complementing Tim's contributed figure).

This is work in progress, but I hope it can give people confidence that this issue is addressed (and can be explained so that it is clear to other readers that it is addressed).

I have marked this pending review, hopefully we can close it or identify further action to take to resolve it during the meeting.

--James

On Aug 9, 2012, at 11:25 AM, Jun Zhao wrote:

> James and Stian,
> 
> On 08/08/2012 11:33, James Cheney wrote:
>> OK, I will revise to make this clearer.  This can be done by adding explanation rather than by making technical changes, though, so we should focus on resolving the other issues now.
>> 
> 
> +1
> 
> -- Jun
> 
> 
>> --James
>> 
>> On Aug 8, 2012, at 11:02 AM, Stian Soiland-Reyes wrote:
>> 
>>> On Wed, Aug 8, 2012 at 10:47 AM, James Cheney <jcheney@inf.ed.ac.uk> wrote:
>>>> I understand your concern.  At a purely technical level, we avoid nontermination, but only by drawing a fine distinction between having an entity(e) statement in the instance, and "knowing e is an entity" (represented by entity' \in typeOf(e)).
>>>> 
>>>> We (me, Luc, Paolo, Tom) discussed three ways of avoiding this problem, before finalizing the review copy:
>>>> 
>>>> 1.  Drop the entity-generation-invalidation and activity-start-end inferences altogether.
>>>> 2.  Add some limitation to inference (such as your suggestion of not triggering inferences on generated existential variables, or applying the ) that recovers finiteness.
>>>> 3.  [what we have done] demote the type inferences to only infer constraints like 'entity' in typeof(id) , not add new statements like entity(e) to the PROV instance.
>>>> 
>>>> Of the three, the one with strongest consensus was (3).  Some of us strongly felt that the e-g-i and a-s-e inferences are needed.  Others, including me, strongly felt that (2) would be a bad idea, as it breaks the connection to logic (i.e., the e-g-i and a-s-e  and may have more radical unforeseen consequences.
>>> 
>>> OK, I see now that you have thought about this. It would be useful if
>>> some of those considerations shone through to the document. I can
>>> agree on argument 3 if we formulate it well, explicitly.
>>> 
>>> I agree that the inferences are needed, or at least useful.
>>> 
>>>> We believed that (3) was an acceptable compromise (if a bit hacky), but, I'm not sure how the group as a whole would feel.  That is why I'm laying out the options we considered.
>>>> So, my proposal would be to make this distinction clearer (and explain why) so that it does not surprise or bite people...
>>> 
>>> I would think this is the best approach, rather than dropping the
>>> inferences all together, as the rest of the constraints rely on them.
>>> 
>>> --
>>> Stian Soiland-Reyes, myGrid team
>>> School of Computer Science
>>> The University of Manchester
>>> 
>>> 
>> 
>> 
> 
> -- 
> Jun Zhao, PhD
> EPSRC Postdoctoral Fellow
> Department of Zoology
> University of Oxford
> Tinbergen Building, South Parks Road
> Oxford, OX1 3PS, UK
> 
> 


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

Received on Thursday, 9 August 2012 14:04:24 UTC