- From: Jeff Heflin <heflin@cse.lehigh.edu>
- Date: Fri, 04 Jan 2002 16:11:00 -0500
- To: Deborah McGuinness <dlm@ksl.stanford.edu>
- CC: herman.ter.horst@philips.com, phayes@ai.uwf.edu, jos.deroo.jd@belgium.agfa.com, jjc@hplb.hpl.hp.com, ned.smith@intel.com, connolly@w3.org, hendler@cs.umd.edu, www-archive@w3.org
I am not sure I am comfortable with including the explainability requirement in this document. Although I can clearly see the usefulness of explainability, I don't think it is absolutely necessary in OWL. In fact, I think it would be better suited for the Proof layer of the Semantic Web (whenever the W3C gets around to defining it). Furthermore, I don't see how it directly impacts language design (as opposed to system design). In your possible approach section, the only thing that seems language related is the ability to tag information with its souce. What other language features would be necessary? I could be wrong, but I would think that any language which has a proof theory is inherently explainable (although systems that implement the language may or may not support explainability). Thus, I believe explainability is not a language requirement, but may instead be a requirement for specific applications. However, I'd like to get input from the rest of the group in order to make a decision on this. Second, I don't want to have our candidate requirements elevated to that same status as those that had achieved a reasonable consensus. However, I would be willing to include them in an appendix perhaps called "Other desirable features". If Explainability does not make the cut for actual requirement, then we can put it there as well. Finally, I intentionally kept the document in text format in order to make it easy for anyone to read and edit. I also wanted something that could be easily converted to whatever format the combined use cases document will be in. As I understand, many at the W3C object to exchanging documents in a proprietary format (e.g., Word). Furthermore, the HTML generated by Word is awful (just look at the source sometime, there are all sorts of junk in there). I understand your concern about reading extra long lines, but most mail readers, text editors, and word processors can be configured to word wrap when displaying such documents. Let's keep the document as ASCII text until a format for the combined document is decided on. Jeff Deborah McGuinness wrote: > > I have made another update (building on top of my previous update). > I did a few things: > > 1 - wrote up R18 - explainability > 2 - added our previous definitions for the topics that we had discussed but not come to consensus on > 3- added a section populated with only 2 topics for requirements that emerged from other groups. I took the topics that emerged from Guus' presentation yesterday and put down default information and part/whole information > 4- i put the document in word (because of the problem having to scroll left to right to read the document in its no carriage return format) and then also dumped it in html form for those who dont like word. > > I attach to two forms of the document here but I also put them both up on the web at: > http://www.ksl.Stanford.EDU/people/dlm/owl/ > > I propose that this is the new baseline. > If anyone wants to write up a case for any of the topics that just have a name and a description, please do so asap. > > thanks, > Deborah >
Received on Friday, 4 January 2002 16:11:09 UTC