Re: Multidimensional vs. One-dimensional info WAS Re: [xtm-wg] A challenge on "the graph"

[Lars Marius Garshol <larsga@garshol.priv.no>]

* Bernard Vatant
|
| They say indeed : Topic Maps are so simple. Why did you complicate
| them with all that unnecessary twisted syntax ?

* Lars Marius Garshol
|
| I have to agree with Murray here: because that is where the value of
| topic maps lie. What is interesting about topic maps is the features
| they provide _beyond_ what a simple graph model does. 
 
* Sam Hunting
|
| These points of view are orthogonal.

Certainly not! If topic maps are reduced to having onyl the features
graphs have, requiring users to invent their own conventions for
representing names, occurrences etc, much of the value of topic maps
will have been lost.
 
| Only with the topic map angle brackets are somehow processed -- that
| is, only when knowledge is actually interchanged, through "the
| graph" -- do topic maps have value.

Certainly, but if that graph only has nodes and arcs valuable
information has been lost.
 
* Lars Marius Garshol
|
| Those who created the topic maps model painstakingly raised some
| features up out of the generality of the graph model (occurrences,
| names, ...). For the processing model to smudge all that back
| together would be counter-productive, I think.
 
* Sam Hunting
|
| Smudge? I don't understand this. Are you claiming that there is
| information loss when the angle bracket syntax is transformed in
| processing into the graph?

There seems to be, yes. Where did all the URIs go? Where did the URIs
of all nodes in the graph go? Where did the subject indicator URIs go?
Where did the subject address URIs go? Where did the occurrence and
variant name resource URIs go?
 
| My thinking has been that "where the URIs 'are'" is exactly the sort
| of thing that vendors should want to compete on -- 'where' they
| 'are' is a storage issue.

I doubt that there is much here to compete on, but in any case the
processing model will have to require that the URI of an occurrence
somehow be connected to the occurrence which again must be connected
to the topic. Precisely where and how the URI string is stored is up
to implementors, but the connection must be maintained. 

Please note that there is no contradiction between the data model
saying 'there must be a node representing an occurrence, which must
have as one of its properties the URI of the resource that contains
the occurrence of information about the topic'. Implementors are still
allowed lots of leeway as long as they don't lose the information that
the URI belongs to this occurrence.

| If a URI is used to establish the subject identity of a topic, for
| example, is it not sufficient to specify that it does not, without
| specifying "where" they are?

It's enough to say that the URI is connected to the topic, yes. But
that must be said, and said very clearly.

--Lars M.


------------------------ Yahoo! Groups Sponsor ---------------------~-~>
Do you have 128-bit SSL encryption server security?
Get VeriSign's FREE Guide, "Securing Your
Web Site for Business." Get it now!
http://us.click.yahoo.com/EVNB7A/c.WCAA/bT0EAA/2n6YlB/TM
---------------------------------------------------------------------_->

To Post a message, send it to:   xtm-wg@eGroups.com

To Unsubscribe, send a blank message to: xtm-wg-unsubscribe@eGroups.com 

Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/