[sc34wg3] Problem with wildcards

[Robert Barta]

On Wed, Mar 04, 2009 at 03:23:22PM +0100, Lars Heuer wrote:
> > No, I meant "map A 'imports' map B".
> 
> CTM provides the mergemap-directive and the include-directive. The
> former works like the XTM 2.0 equivalent; the templates are invisible
> for the master file.

Ok, got that now.

> > And I would say "that is exactly how it should be". So if I use
> > 'include' I get the ?capital to be locally active.
> 
> Yep, but the item identifier <geo.ctm#$__1.capital> screams for
> trouble. If we have a CTM source which is used by many topic maps,
> like the normative TMCL constraints source, all generated topics will
> carry the same item identifier in all topic maps which include the
> that source.

Which simply tells me that that TMCL constraints source (or whatever)
should NOT be _included_, but _mergemapped_.

> ............. And if you merge topic maps which have included that
> common source you'll get ugly side effects, like the one from Lars
> Marius' example where "Oslo" and "Stockholm" merge even if it was not
> the intention of the author.

Well, the author is the author is the author. Nothing to prevent
stupid things to happen.

Or, are we crosstalking and the issue is that with _mergemap_ you do
not get access to the templates within? And you want that? And with
_include_ you get more (sideeffects) than you wanted?

\rho