- From: David Wood <dwood@softwarememetics.com>
- Date: Wed, 2 Mar 2005 12:22:16 -0500
- To: <public-swbp-wg@w3.org> <public-swbp-wg@w3.org> <public-swbp-wg@w3.org>
Hi all, This message provides my comments on the editors' draft of the Note "Ontology Driven Architectures and Potential Uses of the Semantic Web in Software Engineering". OVERVIEW -------- I have read Mike Uschold's comments and must say that I agree with both the comments and the spirit. I am a bit troubled by the "evangelistic" nature of the document. In general, the tone of the document needs to be reworked so that it provides a balanced view of the opportunities without trying to overly sell the concepts. It is clear that the authors know the material and have good ideas, but they need to be more clearly communicated. The core of the document is section 3, Proposed Ideas. Unfortunately, the ideas are not clearly described and the descriptions are probably too brief. I can understand Mike's thought that "Ontology Driven Architectures" was not addressed, even though I presume that was the intention in section 3.1. I *do* think that the subject matter is suitable for a Note and encourage the TF to pursue it. It fits the SWBP direction, in that it provides useful ideas and even some guidance toward furthering those ideas. There are some issues with the WG Charter, especially noting that "guidelines that are not based on some form of established practice are out of scope" and "work that requires basic research is out of scope". Although we have consensus that we wish to pursue this TF anyway, we should be cautious not to go too deeply into those particular rat holes. I am confused on the provenance of this document. Did everyone on the editors' list really participate? Or was this draft primarily written by Phil and Jeff? DETAILS ------- (Status of this Document) Please note that the proper name of the host working group is the "Semantic Web Best Practices & Deployment Working Group". It is a mouthful, but the full name should be used, especially in our own documents. The "& Deployment" part of the name tends to be left out for brevity, including in this document. The public-swbp-wg@w3.org mailing list requests that public comments include "comments: " at the start of the Subject header. We should probably say so in the WG documents, if we expect anyone to do it. (Table of Contents) I personally think the document reads better if order of sections 3.1 and 3.2 are reversed. (Introduction) I think I understand (and agree with) the points made in the Introduction, but the paragraph needs some real work to make them more clearly. (2.1 Composition and Reuse) I am impressed by anyone who can define 'objects' as 'bundles' with a straight face. Well done :) (2.2 A Heritage in Model Driven Architecture) Surely other ontological approaches have been part of CASE and MDA, even if they were proprietary and non-OWL? I think this document should better address the history of what has been tried. The document says, "the more formal the modelling [sic] approach chosen, the more abstract the tools needed, often making methods difficult to implement, limiting the freedom of expression available to the engineer and proving a barrier to communication amongst practitioners with lesser experience." This seems to be a central point, and yet the document fails to address it. If the greater formalism of OWL is used, what is the solution to this problem? How do SemWeb techniques help? "MDA does not currently support automated consistency checking". Was that intended to mean that MDA implementations/tools don't? Or that it is somehow a failure of the architecture? If the former, is there a barrier to doing so? How can a SemWeb approach help? "What is needed in addition is a way to incorporate unambiguous, rich semantics into the various semi-formal notations underlying methods like the UML." Although I get that rich semantics will assist in implementing consistency checking, the document has not told me why this is so important. (3.1 Ontologies as Formal Model Specs...) "the amenable syntax of Semantic Web languages" is a value judgment and should be avoided. "the universality of the Semantic Web's XML ancestry" - Hmmmm. This seems to conflict with the W3C's official (and unofficial) positioning of the Semantic Web. I think the document could benefit from a stronger analysis of why OWL should be the ontological language of choice. "Because it is ours" probably isn't the reason. (3.2 The Semantic Web in Systems and Software Engineering) The document says that consensus is forming around two definitions. Why is this so? What are the pros and cons or tradeoffs or directionality differences between these two definitions? Similarly, the terms "Passive" and "Active" are defined, but not used. Why are these definitions valuable? (3.3 A Corpus of Reusable Content...) Who else has discussed the use of the Semantic Web as a relational database? Some references would be useful. This comment goes for other parts of the document, too. The first bullet ("Semantic Web technologies could be used to formalise associations between sub-components within a given system") seems to belong in 3.2 instead of 3.3. It would be good to reference some of the ongoing work in regard to Web Services/SemWeb integration. (4. Previous Experience) I expected more from this section than a marketing push. I do not think that the Semantic Web can rely on the history of the Web to ensure its success. (5. Issues) What is being done to address these issues? Do any of the issues appear to be a long-term problem for the uses you are suggesting? See Jennifer Golbeck's work on Trust (http://www.mindswap.org/~golbeck/). NITS ---- (2.1 Composition and Reuse) s/systems building block/system building blocks/ s/culminated/which culminated/ (3.2 The Semantic Web in Systems and Software Engineering) s/Software'/Software/ The bullets for the "Passive" and "Active" definitions should be indented once more. Regards, Dave
Received on Wednesday, 2 March 2005 22:26:20 UTC