Re: PROV-ISSUE-643 (TomDN): Include additional constraint hadMember implies hadDictionaryMember with unknown key [PROV-DICTIONARY]

Hi Tom,

I would phrase it as follows:

IF hadMember(d, e) and 'Dictionary' \in typeOf(d) THEN there exists a 
key k such that hadDictionaryMember(d, e, k)

- "there exists a key" ... rather than an unknown key  (if fact it can 
be known!)
- not write "k" but k : a key is a prov-dm literal.

Luc

On 03/22/2013 12:40 PM, Tom De Nies wrote:
> The editors agree with this comment.
>
> The proposed resolution is to add the constraint
> IF hadMember(d, e) and 'Dictionary' \in typeOf(d) THEN 
> hadDictionaryMember(d, e, "k") with k and unknown key
> to PROV-Dictionary.
>
> If any members of the WG have an objection to this, we ask kindly to  
> inform us by replying to this email. If no objections are received 
> before Tuesday March 26th, we will assume this resolution is accepted,
>
> - Tom
>
> 2013/3/7 Provenance Working Group Issue Tracker <sysbot+tracker@w3.org 
> <mailto:sysbot+tracker@w3.org>>
>
>     PROV-ISSUE-643 (TomDN): Include additional constraint hadMember
>     implies hadDictionaryMember with unknown key [PROV-DICTIONARY]
>
>     http://www.w3.org/2011/prov/track/issues/643
>
>     Raised by: Tom De Nies
>     On product: PROV-DICTIONARY
>
>     Originally raised by Stian in his review, but agreed to postpone
>     to next draft.
>
>     Should we add the following constraint?
>     IF hadMember(d, e) and 'Dictionary' \in typeOf(d) THEN
>     hadDictionaryMember(d, e, "k") with k and unknown key.
>
>     In Stian's original email:
>     Also I don't quite understand this.  So a prov:Dictionary kind of
>     collection can have members that don't have keys?
>
>     entity(d, [prov:type='prov:Dictionary' ])
>     // implies:
>     entity(d, [prov:type='prov:Collection ])
>
>     hadDictionaryMember(d, e1, "k1")
>     // implies:
>     hadMember(d, e1)
>
>     // But what if we also see?
>     hadMember(d, e3)
>     // are you saying this would NOT imply the below?
>     hadDictionaryMember(d, e3, ?unknownKey)
>
>     If so then I am a bit confused - a prov:Dictionary to be useful should
>     be a constrained prov:Collection in which every member is associated
>     with a key. This should be added to the Conceptual Definition of
>     Dictionary above.
>
>     If there is no such implication (of course the key is unknown until
>     stated otherwise), I am not sure in which cases such a data type could
>     be useful. It would be like describing an array type of collection,
>     but where some items are allowed to not have a position.  (which is
>     quite different from saying they have an unknown position!)
>
>
>
>

-- 
Professor Luc Moreau
Electronics and Computer Science   tel:   +44 23 8059 4487
University of Southampton          fax:   +44 23 8059 2865
Southampton SO17 1BJ               email: l.moreau@ecs.soton.ac.uk
United Kingdom                     http://www.ecs.soton.ac.uk/~lavm