[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]
Subject: Re: [xtm-wg] Equivalence between TM,RDF,Conceptual Graphs?
Danny Ayers wrote - > I'm currently putting > together some (Java) code to expose the metadata of a RDBMS. I've more or > less figured out how to do this with RDF as the target - essentially > wrapping the DB in an OO facade, then mapping it across to (probably) the > Jena RDF API. I would also like to be able to provide a TM view, and the > obvious approach I suppose here would be to aim for XTM. Unfortunately I've > virtually no knowledge or experience in this field, so any suggestions or > just pointers to materials or whatever would be greatly appreciated. > I don't really have any pointers. Maybe someone else on the list? I do have a few (fairly obvious) thoughts. A rdbms returns sets of rows. Taking C. J. Date's viewpoint, a row's type is a "relation" - a good candidate for a TM association. The type of a row is the union of each of its column types. Each actual row is an instance of that type. Each role in the TM association would be one of the column types (or labels) of the returned data. The whole association would be an instance of some association topic. The main issue to deal with,as I see it, is that the XTM dtd doesn't allow a member of an association to have just string data, it is supposed to be a hyper-reference. Specifically, it is supposed to be a simple xlink element. I think this makes it unduly clumsy when you only want to supply strings, as you might want to in your case. I doubt that you would want to create a separate topic for the value of each cell in each row, for example. With a computer generated topic map, you certainly could, but it seems like a lot of overhead for very little return. It would also make the xml serialization extremely hard for a person to comprehend, since data values that should be seen together would be spread out. The approach I have favored to get around this issue is to make the data for each cell a uri using the "data:" scheme. Of course, all processors may not understand this scheme. Maybe others on the list can provide a better approach (hint, hint!). (Parenthetically, I'd appreciate hearing from those who really know (Steve Pepper, perhaps?) why an association member can't contain PCDATA for its value. It has never made sense to me) Another thing to think about is that you could supply different scopes for different groups of results from the database. This could turn out to be a very nice feature. Finally, you might want to look at whether you can transform your rdf into a TM or vice-versa, maybe even with xslt. Given the regularity of the data from a rdbms, this is likely to be feasible. Hope this is helpful, and let's have others pitch in here, if you would. Cheers, Tom P To Post a message, send it to: xtm-wg@eGroups.com To Unsubscribe, send a blank message to: xtm-wg-unsubscribe@eGroups.com Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]
Powered by eList eXpress LLC