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

Re: PROV-ISSUE-331 (review-dm-wd5): issue to collect feedback on prov-dm wd5 [prov-dm]

From: Satya Sahoo <satya.sahoo@case.edu>
Date: Wed, 11 Apr 2012 22:11:43 -0400
Message-ID: <CAOMwk6x24W4PC6ux5xDw-_qaB+R5P9+Cup0mWxqgWWvixTuBkw@mail.gmail.com>
To: Provenance Working Group <public-prov-wg@w3.org>
Hi DM Editors,
Overall, the document is easier to read and conceptually accesible as
compared to previous versions (though the initial sections are bit long and
it takes time to reach the actual constructs).

Please note that, I am raising issues for WD5 that may partial overlap with
previously raised issues - but as we discussed earlier while closing
previous issues I am treating WD5 as new unconnected DM version.

---------------------
My detailed comments:
Section 2.2:
1. The definition of "derivation" states "transformation", "construction,
"update" of an entity - but the example for derivation includes
"transportation of a work of art from London to New York" - how is change
of geographical location of art piece a transformation of the art piece?
2.The definition of "collection" and "account" is not clear - collection is
defined as "structure to some constituents, which are themselves entities"
and account is defined as "contains bundles of provenance descriptions".
Since provenance descriptions are also entities (from provenance of
provenance) - then effectively account is same as collection?

Section 3.1 and 3.2:
1. The examples describing generation of "tr:WD-prov-dm-20111215" is not
clear - the identifier refers to either the "second working draft" (which
requires approval by W3C) or a technical report (which can be result of
edit1, but is not a "second working draft" without W3C approval? Hence,
either two sets of ids should be used for two distinct entities or they can
only be generated by a single activity (as I believe is stated in
constraints).

Section4:
1. The statement "The fourth component consists of relations linking
entities somehow referring to a same thing." is ambiguous - suggestion is
to remove "somehow"

Section 4.1.3:
1. If the property "wasGeneratedBy" is by definition describing the
"production of new entity by an activity", then how can activity be
optional? The example "wasGeneratedBy(e,-,2001-10-26T21:32:52)" makes
little sense since we know that entity e is present hence it was generated
by some activity, but the provenance assertion with property wasGeneratedBy
is superfluous - the more informative assertion will be wasGeneratedAt?

This ties in with the issue of making activity mandatory for Usage in
Section 4.1.4.

Section 4.1.5 and 4.1.6:
1. The term "trigger" has not defined before. Is "trigger" as type of
entity or activity or agent?
As defined for "wasStartedBy" and "wasEndedBy", the "trigger" is
responsible for (change in) activity, hence it satisfied the definition of
agent - is it just a specific type of agent?

Section 4.1.8:
1. Similar to the previous point, if an activity is responsible for start
of another activity, then it satisfies the requirement for being an agent.
The example uses "computer processes" a1 and a2 which are "software agents"
as defined in Section 4.2.1?

Section 4.2.1:
1. Include "legal" along with "social" since a company is a legal
organization.

Section 4.2.3:
1. The property wasAssociatedWith is by definition describing
the assignment of responsibility to an agent for an activity", but if the
agent is optional, then what does the assertion "wasAssociatedWith(ex:a, -)
convey? If the agent is not known then there is no need for the assertion
describing an association between activity and agent - under the open world
assumption, there can always be an agent associated with the activity.
2. Is the plan can be associated with an agent or an activity - in case of
a process specification (e.g. processor in taverna) or an entity - iphone
or building (for the last two they are independent of an agent).

Section 4.3.2:
1. If the agent in wasRevisionOf optional,
then wasRevisionOf(tr:WD-prov-dm-20111215, tr:WD-prov-dm-20111018, -) is
same as wasDerivedFrom(tr:WD-prov-dm-20111215, tr:WD-prov-dm-20111018),
hence redundant. Either agent should be mandatory or it should be dropped.

Section 4.3.3:
1. Similarly to above point, if quoterAgent and originalAgent are optional,
the wasQuotedFrom(dm:bl-dagstuhl,wp:thoughts-from-the-dagstuhl-principles-of-provenance-workshop/,
- ) is same
as wasDerivedFrom(dm:bl-dagstuhl,wp:thoughts-from-the-dagstuhl-principles-of-provenance-workshop/,-),
hence redundant. Agent need to be mandatory for wasQuotedFrom.

Section 4.3.5:
1. What is the difference between
tracedTo(dm:term-original-source,go:credit-where-credit-is-due.html)
and hadOriginalSource(dm:term-original-source,go:credit-where-credit-is-due.html)?
2. What is meant by "Derivation and association are particular cases of
traceability."? Association is between agent and activity, while tracedTo
is defined to be between two entities? Further, Section 4.3 states that
"subtypes of derivations Revision, Quotation, Original Source, and
Traceability" - so what is subtype of what?

Section 4.4:
1. The statement "An entity is a specialization of another if they refer to
some common thing but the former is a more constrained entity than the
latter." is not clear - entity don't refer to (common) thing - they are the
"thing"? What does "constrained" mean?
2. "The common thing do not need to be identified." - then how is it known
that two entities are referring to the common thing if the common thing
itself is not known?
3. The example specializationOf(ex:bbcNews2012-03-23, bbc:news/) is not
clear - how is a http redirect a specialization? What common entity are
they referring to?

Section 4.5:
Similar issues as above, I will raise it against the ongoing specialization
thread. Also, the example is incorrect
"alternateOf(tr:WD-prov-dm-20111018,tr:WD-prov-dm-20111215)" how is the
second working draft an alternate of the first WD - can we replace WD2 with
WD1 (maybe in an activity) and get the same result?

Section 4.7.4.2:
1. Location is not covered by either XSD or RDF specification (from Section
4.7.5)?

Minor:
Section 4.2.4:
1. Use "superior" instead of "responsible" for actedOnBehalfOf

-----------

I believe the primary issue with WD5 are specialization and alternate
constructs and they need to be resolved before release.

Thanks.

Best,
Satya



On Thu, Mar 29, 2012 at 9:36 AM, Provenance Working Group Issue Tracker <
sysbot+tracker@w3.org> wrote:

> PROV-ISSUE-331 (review-dm-wd5): issue to collect feedback on prov-dm wd5
> [prov-dm]
>
> http://www.w3.org/2011/prov/track/issues/331
>
> Raised by: Luc Moreau
> On product: prov-dm
>
> When sending feedback, please send it under this issue or individual new
> issues.
>
>
>
>
Received on Thursday, 12 April 2012 02:12:14 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 13:07:03 GMT