[topicmapmail] Generating TMs out of relational Databases / How To?

[Bénédicte Le Grand]

You raised the good -and difficult- issue... visualization techniques suffer from a lack of evaluation. In the data analysis community, benchmarks exist which allow to evaluate tools (eg classification tools etc). We do not have the equivalent for visualizations.

Visualization has not been considered as a crucial question for a long time, but things are beginning to change. However, many propositions exist, but no comparison between them.

I am trying to launch a project on this subject in France. Actually, we focus on the visualization of large systems, but the goal is to define evaluation criteria and apply them to existing visualization methods.
Visualization is still not very fashionable in France, so I don't know if this project will be accepted. If you are aware of other calls in this area, please let me know, I would be glad to join other partners.

To answer your question, it seems to me a given visualization becomes gradually less efficient when the volume of data grows but it rapidly reaches a threshold above which it becomes useless. So I would rather say there is a threshold. Or many thresholds, depending, as you said, on the visualization requirements, the type of data, its distribution and so on.

Another way to deal with the problem is to adapt your data to a specific visualization technique, by using clustering techniques for example. I am currently investigating this area. 

Your idea of gathering experience and tools in this subject is very good!

Sorry for not providing more "concrete" answer...

Benedicte

>  
> I would be interested in hearing your experience in trying to
> answer a question I think is pertinent to this discussion:
> 
>   Q: In the area of Topic Map visualization, when visualizing
>      Topics and Associations between Topics, are the techniques
>      used to visualize highly-detailed, relatively small Topic
>      Maps suitable for larger scale ones? If not, is this
>      discontinuity based on a continuum, or is there roughly
>      a threshold at which a different visualization paradigm
>      is then required?
> 
> For example, one visualization technique/paradigm might be
> suitable for say, 20 Topics, another for 300 Topics, but
> neither suitable for 3000 or 30,000. Does this *gradually*
> become unsuitable, or is there a threshold (roughly) at
> which point a specific style of visualization doesn't scale
> any further?
> 
> Knowing such a threshold (even roughly) would help determine
> for a given project which visualization approach is suitable.
> 
> Of course, there are many factors in making such decisions,
> such as the required on-screen information, ability to zoom,
> filter, or localize, etc.  What I'm thinking is that it might
> be valuable to begin cataloguing the different visualization
> techniques (as you have done so well in the XML Topic Maps
> book), but perhaps on a wiki or somewhere that would allow
> new projects to be added by their authors. Because visuali-
> zation is so crucial in large scale information management
> projects, and where Topic Maps may provide important new
> abilities, it'd be great to have somewhere where we could
> keep track of existing projects and new developments.
> 
> Now, I don't mean to put any pressure on you -- I'm sure that
> we both agree that there are no simple answers, but it would
> be valuable to the community to begin to delineate the factors.
> 
> I note that easytopicmaps.org seems to have been destroyed by
> spam, so there's not a particularly good, vendor-neutral place
> to do this. I hope to have a wiki up by the end of summer, so
> perhaps there...  until such time, this list does provide
> archives, so we'll at least have a record.
> 
> All the best,
> 
> Murray
> 
> ......................................................................
> Murray Altheim                          http://www.altheim.com/murray/
> Strategic & Service Development
> The Open University Library
> Milton Keynes, Bucks, MK7 6AA, UK
> 
>           Ils ont l'orteil de Bouc, & d'un Chevreil l'oreille,
>           La corne d'un Chamois, & la face vermeille
>           Comme un rouge Croissant: & dancent toute nuict
>           Dedans un carrefour, ou pres d'une eau qui bruict.
> 
> í¢˜fj)fj)b	bh'&j_%r0~hoÜ¢oj)fjb?brf