Re: Review for W3C POE Deliverables - Vocabulary & Expression c'ed

Added as: https://github.com/w3c/poe/issues/185 <https://github.com/w3c/poe/issues/185>

R


> On 19 May 2017, at 08:20, Antoine Isaac <aisaac@few.vu.nl> wrote:
> 
> Renato told me it would be useful if I can continue the comments at https://lists.w3.org/Archives/Public/public-poe-comments/2017Apr/0008.html <https://lists.w3.org/Archives/Public/public-poe-comments/2017Apr/0008.html>
> I couldn't do it before today, sorry.
> This time the nature of comments is a bit random. But they are mostly editorial.
> NB: I’m still using the numbering as in the review: https://www.w3.org/TR/2017/WD-odrl-vocab-20170223/ <https://www.w3.org/TR/2017/WD-odrl-vocab-20170223/>
> 
> - a bit everywhere in the document: it is not useful to declare that a resource is a subclass of owl:thing. Every class is a subclass of owl:Thing!
> I’ve made this remark in the context of an earlier comment on the model about SKOS, but I prefer to re-iterate it separately so that it’s not forgotten.
> 
> - 4.13.1: an “in range of” could be useful to link odrl:Permission to odrl:permission, especially considering the awkwardness of their naming. It’s been done for other POE constructs in this document.
> Same for Prohibition and maybe other POE constructs.
> 
> - 4.15.1 Action is a said to be a subclass of Concept, but the link doesn’t work. Is it supposed to be a link to skos:Concept?
> 
> - 4.16. There are many problems with the wording of definitions in this section.
> First, it is dangerous to include in the definition of actions the fact that “The Assigner permits/prohibits the Assignee to”. In POE Actions are introduced in the context of permissions and prohibitions, indeed, but they could be used in other contexts, let alone for Duties (I've made comments about this elsewhere). E.g., odrl:transfer is questionable as a permission/prohibition action. I would have attached it to a Policy as a Duty. Anyway, I still find dangerous that the documentation categorizes rather generic actions according to their assumed destination in the model.
> Definitions for individual actions read also strange when one compares with the definition of odrl:Action (“An operation that can be allowed [...]”). If one applies this definition with the definition of “use” (“The Assigner permits/prohibits the Assignee to use the Asset as agreed”) it doesn’t really work. What is the ‘operation’ that can be allowed as per the definition of odrl:Action? Is it the operation whereby "the assigner permits/prohibits something”? Not really.
> Finally, many of these definitions are circular. E.g, “The Assigner permits/prohibits the Assignees to anonymize all or parts of the Asset.” doesn’t really define what the action of anonymising is.
> 
> - 4.16.17 “This action will modify an asset which is typically updated from time to time without creating a new asset like a database.”. "Like a database"? I don’t understand what this means.
> 
> - 4.16.20 I’m not sure what is meant for odlr:present. More precisely, I’m not sure I understand whether it is different from publishing. The fact that odrl:display and odrl:play are narrower concepts hints that Present is about some form of “performance”. But odrl:print is not really about performance nor presentation itseld. And how about odrl:stream? Why isn’t it here?
> 
> - 4.16.27. Transfer would be much less ambiguous if “Ownership” was added to it. Otherwise there can be confusion with electronic transfer of files. Also, the definition “The Assigner transfers/does not transfer the ownership” illustrates well the awkwardness of individual actions definitions as noted above. An “action” is defined here by the fact something happens or nothing happens!
> 
> - 4.16.28 I am tempted to suggest to add “Digital” here (i.e., naming the action as “Digital Transformation”, mentioning “a different digital format”, etc) because “transformation” can be ambiguous. Namely, it can cover things like (intellectual) adaptation of a work, which is a very precise notion in the rights domain, which POE is partly made for.
> 
> - 4.18 By now you won’t be surprised by the fact that I disapprove the pre-targetting of a set of actions for Duties :-) And why odrl:transfer (4.16.27, see above) is not here?
> 
> - 4.20.10 In English, using “Event Action” hints that what is going to be defined in this section will rather be an action. But it’s an event. Why not having used “Event”, simply?
> 
> - 4.20.13 The definition that says that the main language of the Asset is different from the original one after translation is surprising. In principle translation doesn’t change the language of an asset. It produces a new resource that is the translation of the original asset.
> 
> - 4.20 Wasn’t there any way to re-use the W3C Media Ontology and/or Web Annotation’s selectors? It is frustrating to see that POE has to reinvent so many wheels, and later on will have to pay the price of vocabulary maintenance for them.
> 
> - 5.1.1 I had comments about this section, but as it was said (https://github.com/w3c/poe/issues/184 <https://github.com/w3c/poe/issues/184>) that this section would be removed, I won’t spell them out.

Renato Iannella, Monegraph
Co-Chair, W3C Permissions & Obligations Expression (POE) Working Group