W3C home > Mailing lists > Public > public-prov-wg@w3.org > September 2011

PROV-ISSUE-104 (time-class): How to relate start/end time to PE, use, generation, etc [Formal Model]

From: Provenance Working Group Issue Tracker <sysbot+tracker@w3.org>
Date: Wed, 28 Sep 2011 15:02:29 +0000
To: public-prov-wg@w3.org
Message-Id: <E1R8veX-00058O-Ul@barney.w3.org>

PROV-ISSUE-104 (time-class): How to relate start/end time to PE, use, generation, etc [Formal Model]

http://www.w3.org/2011/prov/track/issues/104

Raised by: Stian Soiland-Reyes
On product: Formal Model

The conceptual model allows the optional time for:

processExecution ( identifier [ , recipeLink ] , [ time ] , [ time ] , other-attribute-values )
wasGeneratedBy ( identifier , identifier , generationQualifier [, time] )
used ( identifier , identifier , useQualifier [, time] ) 


It is defined as:
5.5.6 Time
Time instants are defined according to xsd:dateTime [XMLSCHEMA-2].
It is optional to assert time in use, generation, and process execution expressions.

I don't particularly like making time a data property like this (not allowing you to say anything about how the time was measured, uncertainty, (see neutrino experiment), relative frame of reference, etc; and does not allow non-gregorian time (seconds only, or timeframes, cpu steps, etc) - that's probably a separate issue.

In the formal model we have a class prov:Time - but no way to connect this to ProcessExecutions. Associating them to use/generation, etc is strongly related to ISSUE-103 - but for PEs it should at least be easy to do:

prov:ProcessExecution prov:startedAt [ 
    a prov:Time; 
    prov:time "2011-02-19T12:03:12Z"^xsd:datetime ] ;
  prov:endedAt [ # .. 
    a prov:Time; 
    prov:time "2011-02-19T12:05:10Z"^xsd:datetime ] ;
  ]  .

the use of a prov:Time class will open for application extensions for the concerns I am thinking of. 

If we introduce prov:follows and prov:preceeds as suggested by the conceptual model in http://dvcs.w3.org/hg/prov/raw-file/default/model/ProvenanceModel.html#event-ordering, it would be possible for asserters who only know about event ordering to use prov:Time instances which only are described using followed/preceeded with other Time instances.

(I guess these terms are in present tense in the conceptual model because we are talking about Time - the two times would always be followed/precededed by each-other - but we could change it to past tense for consistency)
Received on Wednesday, 28 September 2011 15:02:31 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 13:06:42 GMT