Hi Valentine,
Thanks for the explanations. We'll have to wait for Stefanie indeed, but since we've touched upon this topic in today's call, I have to send an email now...
> I think I see what you mean by this requirement and we could need it in the context of Europeana. However I have not included it in my initial list of requirements because I thought it was to related too the way Europeana has implemented EDM (as RDF-XML) rather than being about RDF validation.
>
If you think it is relevant for your application profile, you should raise it. Even if later on we decide that it's out-of-scope [1]. This at least would allow us to clarify the different issues at hand :-)
> To me R1 is more related to the fact that every object should have their own distinct URIs and that it is therefore not possible to assign two URIs to the same object or the same URI to different objects. This requirement fits perfectly what we want in EDM.
>
I understand a part of it but not all. URI, after all, are unique by design. Do you have in mind an example of data that would be wrong according to this constraint?
Cheers,
Antoine
[1] personally I find it a debatable requirement. I would relate it to constraints on the sequence of the order of statements in a file, which is another infamous 'non-RDF' validation issue. But if someone needs to order statements, the requirement deserves to be raised andobserved...
On 6/15/15 9:26 AM, Valentine Charles wrote:
> Dear all,
>
> I have read your comments about R-1 http://lelystad.informatik.uni-mannheim.de/rdf-validation/?q=node/1).
> Your conclusion was that the requirement was about:
> -not describing of same resource in different place (is nothing for RDF: it's more about structuring of files)
>
> I think I see what you mean by this requirement and we could need it in the context of Europeana. However I have not included it in my initial list of requirements because I thought it was to related too the way Europeana has implemented EDM (as RDF-XML) rather than being about RDF validation.
>
> To me R1 is more related to the fact that every object should have their own distinct URIs and that it is therefore not possible to assign two URIs to the same object or the same URI to different objects. This requirement fits perfectly what we want in EDM.
>
> Let's wait for Stefanie's comment.
>
> Best,
> Valentine
> ________________________________________
> From: Antoine Isaac [[log in to unmask]]
> Sent: 12 June 2015 20:06
> To: DCMI Architecture Forum; Ruehle, Stefanie; Valentine Charles
> Subject: Re: [RDF AP] Fixing some RDF AP requirements
>
> Hi,
>
> Thanks a lot to Lars and Karen for the great feedback!
>
> I've done changes that reflect hopefully the discussion, both on the database [1] and the wiki list [2]
>
> The only one requirement that is still unclear is R-1 (the first will be last ;-) ) "Validate for uniqueness of URIs"
>
> Valentine, Stefanie, I'd be curious to have your opinion on this [3]!
>
> Cheers,
>
> Antoine
>
> [1] http://lelystad.informatik.uni-mannheim.de/rdf-validation/?q=requirements/dc-requirements
> [2] http://wiki.dublincore.org/index.php/RDF_Application_Profiles/Requirements
> [3]https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=DC-ARCHITECTURE;8c92de00.1506
>
> On 6/8/15 9:49 PM, Antoine Isaac wrote:
>> Dear group members,
>>
>> About the action:
>> [
>> ACTION: Antoine to look at unclear requirements in Hugo's email
>> https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=DC-ARCHITECTURE;6a1f19cb.1505
>> ]
>>
>> Hugo and I had a discussion about Hugo's mail.
>> The outcome is the list of suggestions (and questions) below.
>>
>> I'd like to ask you to review and give input. Especially, please object if you disagree with the suggestions. Otherwise I'll implement them next week.
>>
>> Cheers,
>>
>> Antoine
>>
>
|