Re: PROV-ISSUE-437 (prov-dm-post-f2f3-review): Final review before last call vote [prov-dm]

Hello,

Below, you can find my post-F2F3 review of this document. I don't know if
my concerns already came up in other reviews. If they did and are resolved,
they can be ignored.

Q: Reviewer question: Can the document be published as Last Call working
draft?
A: Yes, certainly, provided that two concerns are addressed (see below)
regarding the clarity of the primary source, and the OPTIONAL attribute of
bundle in the Mention construct.

Overall, I think the document reads very well. Thanks to the new structure,
and the resolutions at F2F3 of dropping some stuff (e.g. dictionaries), it
seems like a very coherent and clear document, in the final stage of
writing.

I have two concerns (non-blocking, but I would like to know the group's
opinion on them):
- 5.2.4 Primary Source:

A primary source
◊<http://dvcs.w3.org/hg/prov/raw-file/default/model/releases/ED-prov-dm-20120628/prov-dm.html#concept-primary-source>for
a topic refers to something produced by some agent with direct
> experience and knowledge about the topic, at the time of the topic's study,
> without benefit from hindsight.


In my opinion, this definition should be rephrased or clarified some more.I
find it very confusing that the word "topic" pops up here, whereas it isn't
mentioned anywhere else in the document. Couldn't this definition be
phrased using entity?
Perhaps a (rough) proposal:

> A primary source for an entity is a derivation that refers to an entity attributed
> to some agent with direct experience and knowledge about this thing,
> without benefit from hindsight.



- 5.5.3 Mention

bundle: an *optional* identifier (b) of a bundle that contains a
> description of supra and further constitutes one additional aspect
> presented by infra.


Perhaps this came up before, but I don't see why bundle would be optional.
Then why would one use this construct instead of a regular specialization?
If an example exists, I think it should be written here. If no example
exists, I suggest making the bundle attribute mandatory. If an external LC
reviewer does see (and motivate) the need for this attribute to be
optional, we can change it back.


Minor remarks and typo's:

- 2.1.1 Entity and activity

In PROV, things we want to describe the provenance of are called
*entities*and have some fixed aspect.

Typo aspect"s", and perhaps rephrase this to the actual phrasing we use in
the rest of the document. I think it's clearer

- 2.1.2 Derivation

> If an artifact was used by an activity that also generated a new artifact,
> it does not always follow that the second artifact was derived from the
> first.

Why are we talking about artifacts instead of entities here? -> confusing

-  2.2.1 Mechanisms to Define Extended Structures

> A software agent is running software.
>
This is a bit rough. (a computer also runs software)
Perhaps: A software agent is a digital agent whose actions are the result
of the execution of a piece of software.

- 2.2.1.4 Further Relations

> Finally, PROV-DM supports further relations that are not subtypes or
> expanded versions of existing relations.
>
Such as?

- 2.2.3 Collections

> Many different types of collections exist, such as *a* *sets*,
>
typo: remove the "a"

- 3. The Provenance Notation

> To further disambiguate expressions that *contains* an optional
> identifier,

typo: contains -> minus s

- 5.1.4 Usage

entity: an optional identifier (e) for the entity being used;
>
An example as in the previous section (example 19) on generation with the
optional entity left out would be great.

- 5.1.7 End
Copy-paste error: (same as start)

> ender: an optional identifier (a1) for the activity that generated the
> (possibly unspecified) entity (e);
>

- 5.3 Component 3: Agents, Responsibility, and Influence

> The second component of PROV-DM,
>
typo: should be the third component

extended structures comprise and UML association classes

typo: lose the "and"

It would maybe be nice to mention something about "scruffy"or "imprecise"
or "incomplete" provenance here, and mention that the use of influencedBy
is discouraged except for this kind of provenance, when there are no other
options.

- 5.3.2 Attribution

> agent: the identifier (ag) of the agent whom the entity is *ascribed* to,
>
Maybe use "attributed to" instead of "ascribed to" to keep consistency

- 5.5 Component 5: Alternate Entities

> The fifth component of PROV-DM is concerned with relations specialization
> and alternate between entities.
>
Why isn't "mention" mentioned?

- 5.5.3 Mention

> The following notion is a relation between two entities with regard to a
> bundle.

I would add here: "It is a special case of specialization."


Thanks for writing/editing a very nice document! I think we can be proud of
this one.

- Tom


2012/6/28 Provenance Working Group Issue Tracker <sysbot+tracker@w3.org>

> PROV-ISSUE-437 (prov-dm-post-f2f3-review): Final review before last call
> vote [prov-dm]
>
> http://www.w3.org/2011/prov/track/issues/437
>
> Raised by: Luc Moreau
> On product: prov-dm
>
>
> This is the issue to collect feedback on the prov-dm document (version
> created after F2F3)
>
> Document to review is available from:
>
>
> http://dvcs.w3.org/hg/prov/raw-file/default/model/releases/ED-prov-dm-20120628/prov-dm.html
>
> Question for reviewers: Can the document be published as Last Call working
> draft?
>
> Cheers,
> Luc
>
>
>
>