PROV-ISSUE-610: prov-o query profiles and use cases not normative [PROV-O HTML]

PROV-ISSUE-610: prov-o query profiles and use cases not normative [PROV-O HTML]

http://www.w3.org/2011/prov/track/issues/610

Raised by: Timothy Lebo
On product: PROV-O HTML

http://www.w3.org/mid/67E90CC9-A4BC-4EBF-82CD-D3A172FD2B35@dydra.com



yes i have. my concern could be addressed by adding a normative appendix which specified an ontology profile for each of the three use cases that i described, below. i did read, that there are categories of terms - on one hand basic/extended/qualified and on the other subclasses of respective category, but there is
a, no indication that a compliant provenance information source may provide just a subset,
b, no coherent subset is describe for any of these use cases, and
c, it may be, that some cross section of those categories is appropriate to achieve coherence for these use cases, and, in particular, perhaps a different one for each case.

the working group may feel, that this is not within your mandate, but rather it is the responsibility of a sparql working group. to that, i can only reiterate my observation, that the two sparql/rdf implementations evident in the w3c list did not share a view of what those profiles should be, which certainly does not contribute to your group's interoperability goal.

best refards, from berlin,

Thanks
Paul

On Dec 14, 2012, at 1:38, james anderson <james@dydra.com> wrote:

good morning,

upon the publication of your latest provenance document set, i
attempted to subscribe to the indicated comments mailing list in
order to submit a comment. the w3c response then pointed me at a web
page intended to manage the group, which page permits access to
members only. please advise me once public subscription is possible.

in the meantime, if you would be able to forward a comment, that
would be helpful.

we are a rdf cloud service[1] and intend to support the provenance
standard. as such, we would very much like to respond to your
invitation for information about implementations. you will facilitate
this to a great degree, if you structure your standard such that it
expresses clearly what to implement for which purpose.

with respect to rdf storage systems, it appears that the
implementation wiki page[2] relates just two: callimachus and
openrdf. given even just these two, however, the implementations are
not particularly interoperable. it would serve your effort greatly,
if you would add to the standard clear specifications for the
ontology subset and the interpretation to be applied to the
respective terms for certain known use cases for rdf stores and
sparql services.

for example, in the literature and implementations described by the
documents referenced from the w3c pages, three clear "provenance
profiles" stand out
- statements
- named graphs
- resources
each of the three cases entails its own ontology / data model subset.
if your documents are to serve the "inter-operable interchange of
provenance information" they should specify the required vocabulary
for each of these cases and provide examples which demonstrate
generation, access to and interpretation of provenance information in
each case.

best regards, from berlin,
---
[1] : dydra.com
[2] : http://www.w3.org/2011/prov/wiki/ProvImplementations
---
james anderson

Received on Friday, 14 December 2012 15:29:40 UTC