[sc34wg3] TMDM / XTM 2.0 comments

[Lars Marius Garshol]

* Lars Heuer
>
> I don't want to beat a dead horse, but IMO the <itemIdentity> element
> on every topic map construct blows the XTM unnecessarily up.

It does blow it up. The question is whether this is unnecessary. :-)

Graham and I really agonized over this question as we really did not  
like the blow-up (which is real). The trouble is that if you remove  
this then only topics will have identity, and the result is that  
there are some things you will no longer be able to do.

An example probably helps here. Let's say you want to make an desktop  
topic map editor that connects to a TM server and gets the fragment  
for a topic from there (in XTM). The user edits the topic, and once  
done the editor sends the updated fragment back to the server (in XTM).

Now, if we have item identifiers on all constructs the server can  
assign an identity to each piece it sends, and transmit this in  
standard XTM. The editor can mess around with these as it pleases,  
and then send standard XTM back, and the server will now be able to  
tell what the changes are since every piece has an identifier.

All the other scenarios are really variations on this theme if I  
recall correctly.

> Is it true, that they are only kept because nobody wants to (or can)
> touch the TMDM again?

No. This was before we sent out the TMDM FDIS, so the train hadn't  
left the station then. It probably has now, though.

> Keeping item identifiers (and variants) seems to ignore the whole  
> potential of the "Atlanta decision" (<http://www.jtc1sc34.org/ 
> repository/0676.htm>), where the committe has decided to break  
> backwards compatibility (for XTM).

I agree that decision did open a window of opportunity. (Of course,  
the window was really too small to be very useful, but that's all  
water under the bridge by now.)

--
Lars Marius Garshol, Ontopian               http://www.ontopia.net
+47 98 21 55 50                             http://www.garshol.priv.no