[sc34wg3] Question on TNC / Montreal minutes

Lars Marius Garshol sc34wg3@isotopicmaps.org
18 Sep 2002 22:36:18 +0200

* Marc de Graauw
| Of course in general it would make sense to look at where a Topic
| Map comes from before you merge it with your own carefully crafted
| masterpiece.  Just merging with the TNC on, certainly with obvious
| names for scoping topics ('city', 'language') is a sure-fire way to
| disaster.

I know. That's why I want the thing made optional or moved into
application space.
* Lars Marius Garshol
| [...] I don't have any use cases for allowing us to turn this
| capability on anywhere.
* Marc de Graauw
| Liar :-)
| You told me once you use the TNC for some of your customers.

Either I misspoke, or you misheard. We do merging by name several
places (which is probably what I've told you), but it is *not* TNC
merging. In all of those cases it is merging by base name within a
particular topic type. As far as I am aware, none of our customers use
the TNC.
| I agree. The only reason to do this [merging] only for basenames is
| we already have it.  But that's a good reason.

Well, it's sufficient reason for the idea to merit discussion. I don't
think letting past mistakes warp the shape of the standards family is
a good idea. Look at what a mess XML has turned into because of this.
For us it is not yet too late to fix our mistakes.
| I would just keep the TNC (made optional), that's the easy way.=20

It seemed easy in Montr=E9al, but people have been sufficiently
disgusted with the solution we spent a whole day coming up with, so I
am no longer convinced that solution is very easy.

| But the only thing I think would be wrong is drop the TNC from the
| SAM without simultaneously having the possibility to reapply it
| (through TMCL).

Well, we might provide other means of reapplying it, without putting
it in the SAM. The SAM will be published long before TMCL, I am fairly
certain, so we'd have to do something like that.

| But sure, TMCL should support more than just basename+scope merging.

This, at least, appears to be uncontroversial. (That's a little
victory in itself. :-)
* Lars Marius Garshol
| If you have a tool that can read in a topic map and export a version
| of the topic map where TNC mergings have been applied, the problem
| goes away. Just preprocess your topic map before you use it.
* Marc de Graauw
| Smart thinking. (It could become a selling point for TM vendors who
| have the preprocessor built into their engine :-))

I very much doubt that it would, unless you could show me a real topic
map that relies on the TNC for internal integrity. I have yet to see

Lars Marius Garshol, Ontopian         <URL: http://www.ontopia.net >
ISO SC34/WG3, OASIS GeoLang TC        <URL: http://www.garshol.priv.no >