[topicmapmail] Survey of Actual Scope Use in Topic Maps

[Lars Marius Garshol]

* Lars Marius Garshol
|
| If you are interested in extending the coverage, here are some more
| topic maps you may look at: <skipped>

* Marc de Graauw
| 
| One the one hand it would be interesting to extend the TM base, but
| on the other there is a methodological advantage in my approach:
| I've taken a sample compiled by someone else for different purposes,
| so probably this sample is random with respect to scope use.

I se what you mean, but the three additions listed are topic maps that
ought to have been in that registry to begin with, and at least one of
them now is...
 
| (Jan did have a link to your i18n.ltm, but it was broken...)

Now corrected. :)
 
* Lars Marius Garshol
|
| What does the "eml" category actually mean? I saw the description,
| but it didn't tell me anything, I must admit.
 
* Marc de Graauw
|
| The first table lists scope use with:
| - long name
| - short name
| - description
| 
| The second table uses the short names as column headers, so "eml"
| means "email addresses". I should have described this better.

That much I actually managed to figure out. :) But the description is
"Email addresses are unique" and I'm having trouble finding out what
sort of scope use that is.
 
* Lars Marius Garshol
|
| In any case some conclusions seem clear:
| 
|  - very few topic maps use scope to avoid the TNC, and so probably
|    they would break if the TNC were applied,
 
* Marc de Graauw
|
| Either that or many of those Topic Maps have no ambigious names. By
| looking at them, I found quite a few are based on structured data
| from elsewhere, so maybe the names were unique to start with. But I
| must admit I did not study all the Topic Maps in-depth, I just
| looked at <scope> and scoping topics.

Very few structured data sources have a restriction that names must be
unique, though of course in practice they very often are simply
because the data set is limited in size. Of course, as you imply, that
may not mean that the author has not cared about the TNC but simply
that the author has not had the need to apply scoping to avoid trouble.

I was thinking that one might look for elaborate name scoping designed
to avoid trouble in the case of merging with other topic maps, but I'm
not sure what one could conclude from the absence of such scoping. I
think it would imply that the original intent of the TNC has failed,
but to conclude that the author ignored the TNC would probably not be
safe.

* Lars Marius Garshol
|
| In addition it appears that the idea of using scope for controlled
| vocabularies is relatively popular.
| 
| Not sure what that means, though.
 
* Marc de Graauw
|
| I think it is a quite natural use, given the TNC which is current
| standard.  Controlled vocabularies have unique names, so using scope
| is a way to state they are unique within this scope.

I agree that given that the standard has this feature and the
community emphasizes this way of using it it may not be so surprising
that people actually do use it. But what we can conclude from that I
am not sure. Perhaps that it is also intuitively appealing.
 
| It also seems the Ontopia invention of using scope for name
| direction in associations has caught on.

I think you can generally consider it a pretty reliable sign that the
author has used the Omnigator to look at the topic map while creating
it. :-)

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