[topicmapmail] An Idea about TopicMap intelligent views

[Carlo Moneti]

Thomas B. Passin wrote:

> 
> As for a common subject ontology, that is always good for an argument! 
> There are a number of reports of studies that have found that there is 
> not much overlap in terms (and maybe concepts, I'm not sure) between 
> several people, even in areas they know well (e.g., Tony Buzan cites work 
> where bankers were asked to give associations about "money").
> 
> I think that to use a common ontology in an easy, everyday way, a person 
> would have to get really familiar with it.  This is not likely to happen 
> on any large scale in an organization.  Heck, I cannot always figure out 
> where to file a given bookmark, even though I myself created the 
> classifying folders in the bookmark collection!
> 
> But then, how to manage without a common ontology?  Hmm, a real conundrum!
> 

I understand. But I would like to suggest a (novice) thought: That a common 
ontology needs to be learned doesn't justify giving up on one. All words 
are learned. All labels for concepts are learned. Any given organization is 
in a good position to establish and enforce an ontology for its domain. I 
good ontology doesn't have to be completely and instantly self-evident to 
everyone; it is good and useful if it's easily understood and learned. The 
LOC subject headings is certainly useful (despite the various professional 
criticism about the current implementation).

Doesn't a common ontology in XTM equate with a common set of 
subjectIndicatorRef? These are developing, and I assume will develop far 
more over time. I also figure archetypal association patterns for processes 
and contexts can be developed, and that TM engines will include the 
processing instructions to render them.

Thanks,
Carlo Moneti
http://arsteca.net