W3C home > Mailing lists > Public > public-prov-wg@w3.org > April 2012

Re: PROV-ISSUE-356 (replace-hadquoter): Replace hadQuoter by reusing wasAttributedTo [mapping prov-dm <-> prov-o]

From: Timothy Lebo <lebot@rpi.edu>
Date: Fri, 20 Apr 2012 09:06:59 -0400
Cc: Provenance Working Group <public-prov-wg@w3.org>
Message-Id: <75AD2599-B18D-4B83-B023-9E11227841D8@rpi.edu>
To: Stian Soiland-Reyes <soiland-reyes@cs.manchester.ac.uk>

I'd like to hear your thoughts on your preference for the design tradeoff:

1) replace hadQuoter and hadQuoted properties, use wasAttributedTo instead 
** increases chance of ambiguity in some cases that could be resolved with more specific entities
** reduces the number of properties in PROV-O

2) keep hadQuoter and hadQuoted


Thanks for these examples. It is definitely food for thought and I'll continue to revisit them as we move forward.

Part of my concern for your "ambiguity" argument is that you are being imprecise in your choice of quoted entity.

If one has a blogpost with a photo and other components contributed by different agents, and you take something from a paragraph, shouldn't you be identifying the _paragraph_ (and not the entire blogpost) as the quoted entity?

Then, ALL agents that attributed to that specific quoted paragraph would be the quoted agents (even Paolo, who made it italics), and we could determine this without the extra hadQuoted predicate.

I'm even more convinced that the hadQuoter predicate should be replaced by wasAttributedTo, since the quote entity is being created during the quotation and can be contextualized appropriately (if any modifications to the quote happen, then it's a new entity with new attributions).


On Apr 19, 2012, at 8:20 PM, Stian Soiland-Reyes wrote:

> On Thu, Apr 19, 2012 at 20:08, Provenance Working Group Issue Tracker
> <sysbot+tracker@w3.org> wrote:
>> hadQuoter could be replaced by reusing wasAttributedTo. This comes with some tradeoffs. Is prov-o team and wg okay with these tradeoffs?
> My argument for why they should not be replaced - see examples in
> https://dvcs.w3.org/hg/prov/file/47b52dc6a91e/ontology/components/wasQuotedFrom-hadQuoter
> https://dvcs.w3.org/hg/prov/file/47b52dc6a91e/ontology/components/wasQuotedFrom-hadQuoter/had-quoter.ttl
> shows how it is now.
> https://dvcs.w3.org/hg/prov/file/47b52dc6a91e/ontology/components/wasQuotedFrom-hadQuoter/had-quoter-was-attributed-to.ttl
> is trying to do as Tim is suggesting - using wasAttributedTo for both
> hadQuoter and hadQuoted - and why I think it does not work.
> https://dvcs.w3.org/hg/prov/file/47b52dc6a91e/ontology/components/wasQuotedFrom-hadQuoter/had-quoter-roles.ttl
> shows an attempt to solve this using roles - but this does not work
> when there are multiple quotations.
> In short, my argument is that wasQuotedFrom is more than a short hand
> for some wasAttributedTo links. As we know, wasAttributedTo just means
> that the agent wasAssociatedWith the activity that generated the
> entity. It does not say anything else about how or to what degree that
> agent was associated, it could be performing the activity, monitoring
> it. In my example, based on the PROV-DM example of the Dagstuhl
> blogpost paragraph, then there could be two agents who created the
> blog post, but only one of which was quoted. Similarly there could be
> two agents making the paragraph, but only one of them who was quoting.
> -- 
> Stian Soiland-Reyes, myGrid team
> School of Computer Science
> The University of Manchester
Received on Friday, 20 April 2012 13:07:45 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:51:11 UTC