On Jan 8, 2020, at 21:50, Nick Crossley <nick_crossley@...> wrote:

Sorry, but I am unable to attend the OSLC meeting on Jan 9th. As far as I am aware currently, I will be available for OSLC meetings starting Jan 16th.

Nick.

I think there are several challenges in doing this.

For a graphical ER model to be normative, it would have to be a graphical representation of TriG so it could properly distinguish reference types - for example, inline vs rerefence vs either.

The RDF of an OSLC resource may contain vocabularies from multiple OSLC specifications. For example, a resource might have multiple RDF types, each of which has an associated resource shape in a specification.

Linked Data necessarily is focussed around the idea of links between artifacts from different application domains. For example, a test case validates a requirement. Adequately representing this in a graphical form means that such a graphical model has to span multiple OSLC specifications. This means that such a graphical model cannot live in any single specification. Constraining a graphical model to only describe a single specification would significantly reduce the utility of having such a model.

Orthogonal to resources in domain applications like requirements management or quality management, resources might be versioned or unversioned. A model might have to represent that as well.

With all the above, trying to produce an accurate graphical model is likely to result in a complex ER diagram. A concern here is whether that visual complexity reduces the utility of it to the point where the resource shape based information in the specification is easier to understand and consume.

One way of addressing this is to have one or more graphical models that spans multiple specifications but that are not normative and not formally part of any specification. It would be part of some separate non-normative guidance document. Depending on the domain(s) of interest, there might be multiple ER diagrams each representing particular use cases, and each ER diagram might be a simplification of the potential richness of data focussing on specific use case(s)..I think this is closer to what the original poster was suggesting. Of course, the danger here is that because such diagrams are not part of any specification and span multiple domains, keeping consistency with changes to specifications might be a challenge, and in all cases, there should be a caveat that the specifications are the authoritative reference are take precedence over the non-normative ER diagrams.

Best regards,
David



From:        "Jad El-Khoury" <jad@...>
To:        "OASIS OSLC Open Project (oslc-op@...)" <oslc-op@...>
Date:        07/01/2020 08:39
Subject:        [EXTERNAL] [oslc-op] Is there a Simple OSLC Metamodel Showing Entities and Relationships?
Sent by:        oslc-op@...

Hi

An opinion about introducing a graphical model of OSLC entities into the specs. I think it’s worth considering.

regards

______________________________

Jad El-khoury, PhD

KTH Royal Institute of Technology

School of Industrial Engineering and Management, Mechatronics Division

Brinellvägen 83, SE-100 44 Stockholm, Sweden

Phone: +46(0)8 790 6877 Mobile: +46(0)70 773 93 45

jad@..., www.kth.se

From: Stack Exchange <do-not-reply@...>
Sent: 28 October 2019 12:51
To: Jad El-Khoury <jad@...>
Subject: New questions in Favorite Tags filter for Oct 28 - Stack Exchange

Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number 741598.
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU

Hi

 

An opinion about introducing a graphical model of OSLC entities into the specs. I think it’s worth considering.

 

regards

______________________________

Jad El-khoury, PhD

KTH Royal Institute of Technology

School of Industrial Engineering and Management, Mechatronics Division

Brinellvägen 83, SE-100 44 Stockholm, Sweden

Phone: +46(0)8 790 6877 Mobile: +46(0)70 773 93 45

jad@..., www.kth.se

 

From: Stack Exchange <do-not-reply@...>
Sent: 28 October 2019 12:51
To: Jad El-Khoury <jad@...>
Subject: New questions in Favorite Tags filter for Oct 28 - Stack Exchange

 

The following item was added to your Stack Exchange "Favorite Tags" feed. Is there a Simple OSLC Metamodel Showing Entities and Relationships? There seems to be any amount of RDF-format for the OSLC but what I'm looking for is a simple E-R-like view of the OSLC metamodel which shows the concepts and relationships which can be used to ... tagged: metamodel, oslc Oct 28 at 11:45 Unsubscribe from this filter or change your email preferences by visiting your filter subscriptions page on stackexchange.com. Questions? Comments? Let us know on our feedback site. Stack Exchange Inc. 110 William Street, 28th floor, NY NY 10038 <3