[topicmapmail] Interpretive Difference between Occurrences and SI or SL

[Tobias Redmann]

Thank you so far, but let me explain my current problem:

I want to write an "topic resolver" this means, I want to get the
subject of a website from a URL. I want to use this as an extension to
retrieve the subject for every site. The input for my function is the
URL of the site, for example http://www.meshed.de/topicmaps. Now in my
Topic Map I can model the following:

Topic: Topic Maps identified by "http://www.meshed.de/topicmaps"

Topic: Semantic Technology with the occurrence
"http://www.meshed.de/topicmaps"

Now I can resolve the topics "Topic Maps" and "Semantic Technology". The
first by the subject identifier and the second by occurrence. The only
problem is, that Topic Maps obviously also has an occurrence namely
"http://www.meshed.de/topicmaps".

Perhaps the example Topic Maps isn't good, but there are subjects
defined by the customer, there cannot be identified by an "external" URL.

So, I guess that the occurrence definetly HAS a meaning for the identity
- probably not by definition (standard) but by interpretation.

Regards,

Tobias


Lars Marius Garshol schrieb:
>
> * Tobias Redmann
>>
>> What is the interpretive difference between occurrences and subject
>> indicators or locators. Both uses URIs to indicate relevant resources.
>> You can use both to get the related topic.
>>
>> The subject indicator describes the subject, but there are a lot of
>> cases where the indicator is the one and only occurrence of the topic.
>> Every occurrence is in some way relevant to the subject. So adding a
>> type "description" or "indicator" to an occurrence mean the same.
>
> Not really. It goes like this:
>
>  - subject locator: the resource *is* the subject of the topic
>  - subject indicator: the resource describes the subject of the
>                       topic *unambiguously*; most likely the
>                       resource is not interesting to an end-user
>  - occurrence: the resource has information about the subject of
>                the topic in some way, and is suitable for an
>                end-user
>
> So in the first two cases the connection between the subject and the
> resource is very close, whereas in the third case it need not be.
>
> A description is not the same as a subject indicator, although
> admittedly they share some features. A description would be very poor
> if it did not tell you what the subject was, but it doesn't
> necessarily do this unambiguously, and it's meant for an end-user more
> than for someone developing Topic Maps solutions.
>
> In general, the whole situation is different, because when you use a
> subject indicator what you want is really an *identifier* for your
> subject, and that's generally more important than the content of the
> indicator. When you add a description you're only interested in the
> *content*.
>
>> This also leads to the question: What to do with topics where the si
>> equals the occurrence and how to model them? What means this?
>
> If you find a case where this happens you have a dubious topic map.
> Why does the subject indicator contain information that's relevant to
> a different topic? It should just be a short and sweet unambiguous
> definition of the subject, not some long discursive piece that rambles
> over different subjects.
>
> I hope that helped.
>
> --Lars M.
> _______________________________________________
> topicmapmail mailing list
> topicmapmail at infoloom.com
> http://www.infoloom.com/mailman/listinfo/topicmapmail
>


-- 
Dipl.-Wi-Inform (FH) Tobias Redmann
Erfurter Str. 31a
98693 Ilmenau

Mobil: 0178 - 8338 795
Mail:  tobias at meshed.de
Web:   http://www.meshed.de