- From: Deborah L. McGuinness <dlm@ksl.stanford.edu>
- Date: Sat, 19 Jan 2008 17:15:30 -0500
- To: Bijan Parsia <bparsia@cs.man.ac.uk>
- CC: conrad.bock@nist.gov, "'Web Ontology Language ((OWL)) Working Group WG'" <public-owl-wg@w3.org>
thanks for the insight into some issues related to the documents. I have some comments below. Bijan Parsia wrote: > > Conrad, > > Thanks for the feedback. I'd like to take one point out of order: > >> I don't think the burden should be >> to comment on this proposed document, but to comment on OWL 1. > > In general, WG members have an obligation to review credible > proposals, esp. on parts they care about. So, actually, there was an > appropriate burden on you, to the degree that you're interested in > these matters, to review our proposal. I put background and discussion > in: > http://www.w3.org/2007/OWL/wiki/Experiments_with_documents#Background_on_the_Proposal > I thought alan originally said of the meeting at the f2f that you would be developing the infrastructure that we might use for the folding and unfolding idea. what i thought he said was that we would potentially get the infrastructure in place and then we might work collaboratively to generate content. (alan please correct me if i am wrong). i would like to separate the evaluation of the infrastructure evaluation from the content evaluation and i suggest we consider discussing both separately. > > That section has some comments that apply even more strongly to your > comment, "it would be better to start with the OWL 1 documents and > merge them to reduce redundancy, update for new features, and make > corrections and clarifications,". This is a huge amount of work; i am in agreement with the position that it would be better to start with the owl 1 documents and update. in fact i felt so strongly about it that i made sure that the ac vote supporting the formation of the working group explicitly made this point in the support position. I would think that also as a point of order we do need to consider ac votes supporting the group with requests from the group at the outset. i also do not think it is unreasonable to believe that some number of people would have expected this from the charter that says that we will: The specification will include (at least) the following components. # *Overview:* An introduction to the new language, including a clear explanation of its differences with respect to OWL. # *Requirements:* A description of the goals and requirements that have motivated the design of OWL 1.1. # *Formal specification:* A formal specification of the language's syntax and semantics (see, e.g., OWL Semantics and Abstract Syntax <http://www.w3.org/TR/owl-semantics/>). # *Descriptive specification:* A less formal, but still comprehensive and systematic, specification of the language's syntax and semantics (see, e.g., OWL Reference <http://www.w3.org/TR/owl-ref/>). # *User guide:* Educational/outreach material aimed at (potential) users of the language (see, e.g., OWL Guide <http://www.w3.org/TR/owl-guide/>). # *Test suite:* A set of test cases aimed at facilitating software development, and at illustrating design features and correct usage (see, e.g., OWL Test Cases <http://www.w3.org/TR/owl-test/>). # *Language Fragments:* A specification of one or more subsets of the language that have been identified as having interesting or useful properties. Each component may consist of one or more documents, and may or may not extend existing OWL documents <http://www.w3.org/2001/sw/#OWLSpecs>. > more, I judge, than simply refreshing the documents "in place"; and > certainly more, to reach the same level of quality, as my proposal. I am not sure that it would be such a large amount of work if we could leverage all of the people who have expressed interest in these documents. I thought the idea that you might suggest some new technical infrastructure (ie., the folding unfolding suggestion with a solution that allows printing of the documents) that we could leverage. But, i wanted to do that with an approach that people who grew to rely on the previous documents would not feel that everything had changed. i think that could still be done by making some effort at continuity and attempt to work together. Right now, i feel that the group has splintered and i think we are best served if we can find a way to leverage all of our efforts together to create documents that we all think are useful. I am not saying that we have to maintain what people think are the bad points from the old documents and also not saying that we need to do as minimal an update as possible. i like the idea of better infrastructure. i would though like to make an effort to have some continuity and not start from scratch for the sake of starting from scratch. > > Additionally, it doesn't seem to achieve seems to be an important goal > for you: it is for me. > >> To users OWL 1.1 is a release with some >> additional expressiveness. It's not a major overhaul. The >> documentation should reflect that. > > Three points: > *I believe that most future users of OWL are new users, as I > believe the OWL community will continue to grow. So the only docs > they'll know are our docs. While i do believe that the owl community will grow, i do not think that most of the owl 1.1 users will only know the 1.1 documents. I think we will be in a critical transition period as we start to come out with owl 1.1 and i do not want to alienate current users (and for that matter people on this working group who would like to see some migration of the previous documents be useful and are offering to help). > *I think users have no problem distinguishing between "a major > overhaul" of the documentation and a major overhaul of the language. i think perceptions are important. why do what appears to be much more of a major overhaul to the documents for the sake of it. for example, why not at least reuse the example material from at least the overview where i believe we have had very few technical issues with the examples. > As I said on the discussion page: I've never heard otherwise before. > So some pointers to cases would be helpful. i provided some scenarios in the who reads our documents page earlier. > *Your proposal seems to create as much user visible upheaval as > new docs (esp. merging). > > If you think that my proposal as it is developing will lead to > significantly worse documents, I encourage you to develop your comment > into a more detail proposal. i would like to see the group try to discuss how we can develop one thread for each document instead of trying to generate what feels like competing documents to fill the same document need. i think if we can come to some agreement on the general outline of the document then we could take pieces of it and make progress collectively. > I also welcome additional feedback on my proposal. As your most used > prior document was the Reference, I suggest you examine the > experimental restructuring of the syntax spec: > http://www.w3.org/2007/OWL/wiki/Experiments_with_documents#Description.2FReference > > > [snip] > > Thanks for pointing to the parts of the reference you found most > valuable. I'll continue to study them. > >> The buttons for syntax choice are very cool. > > Thanks. > >> We need to ensure the >> document prints properly, > > They will. > >> perhaps will all the syntaxes by default, > > It's unclear to me which is better as default. I assume we'll get more > feedback. > >> because folks printing things out for travel will probably not realize >> they need to set the buttons first. >> >> Section 2.1: > [snip] > > Thanks for this feedback. Some of it is already intended (e.g., > classification). I'm not sure about centering on expressiveness as a > concept, esp. if we get into relative comparisons of expressiveness. > For example, I don't like how the current draft talks about OWL as > being "more flexible" than relational schemas. I think setting this up > oppositionally like that, esp. when we are dealing with interest > relative and context sensitive values (yes, the relational calculus > can't express disjunction in the data, but OWL can't express integrity > constraints, and most databases include turing complete langauges). > Instead, I prefer to focus on differences and affordances and > complimentary benefits. It's much better, I think, to send someone > away from the primer *understanding* what OWL can and can not do well, > even if that means they recognize that OWL's not for them, then to > make them feel put upon for their choice of technology. > > Cheers, > Bijan. >
Received on Saturday, 19 January 2008 22:15:44 UTC