[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [Elist Home]
Subject: RE: [xtm-wg] Equivalence between TM,RDF,Conceptual Graphs?
<- > I don't quite understand... you'd have a type something like <- > "INT+VARCHAR(10)" ??? <- > <- Think of a row as being like a struct in C or a record in Pascal <- - there are <- a number of slots, each slot may have a different type. The <- struct itself <- can be thought of as a type. A row instance would be like an <- instance of a <- record or struct. Got it now, thanks. <- As I re-read your original post, I see you are more thinking of <- the metadata <- rather than the data, so I may have been adding confusion in my <- response. I <- was thinking more of capturing actual row data in the TM. <- Still, metadata <- queries are likely to return row data, too, so the principles may be the <- same. Yep, the more I think about this the more self-similar it seems at different levels <- By metadata, do you mean basically the database schema? That's certainly where I'm starting. I'd at least as <- interested in the view and query definitions, since they are how <- you can get <- data into or out of a database. Good point. In a proper database application, all <- access to the data is through predefined queries. Extremely good point - I keep forgetting that. So the right metadata <- would seem to be the returned data types for each query, or the data type <- you need to feed to an update query (I'm using "type" in my <- previous sense <- as the type of a row). That makes sense. <- It seems to me that if you put such a row type definition into a TM <- association, and each member of the association refered to a topic that <- described its data type, perhaps using xml-schema data types, <- you'd be in a <- good position to define classes or whatever from the TM. Time to get the TM notes out... <- To be complete, you'd really want to figure out how to capture <- the integrity <- constraints. That may be impossible in general, since there can <- be stored <- procedures that get triggered by inserts, deletes, etc, and you probably <- can't get and understand those in an automated way. But at least the <- standard relational integrity constraints (from the foreign key <- definitions) <- would be important. Yep - the constraints strike me as being very important, though it might well be hard work (everything so far has been considerably harder than I expected!). I'm not even going to think about dealing with stored procedures & triggers... Thanks again. I've now got plenty to keep me going ;-) Cheers, Danny. 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