Re: [Gzz] Space in RDF structure

[Benja Fallenstein]

Asko Soukka wrote:
> Wed, 12 Mar 2003, Benja Fallenstein wrote:
>
> > > I'm trying to figure out how to create a feel of space in an RDF 
> > > structure, even though the structure can be very dense and the same nodes 
> >
> > How about simply turning off most of the properties most of the time, as 
> > in zzstructure where only 2-3 dimensions are shown at once? E.g. if you 
> > only turn on the properties making up a list/tree, then you will be able 
> > to browse that list/tree quite naturally.
>
> With swad-chart.rdf? Uh... If we think properties as dimensions, that 
> would have been difficult to visualise also from zzstructure.

I don't understand. What do you mean by, 'think properties as 
dimensions'? In zzstructure swad-chart.rdf would have used a different 
structure than it does use, just as anything in zzstructure would have 
used a different structure if designed for RDF :-)

(I.e., saying 'what if we showed RDF property X as a zzstructure 
dimension' doesn't make too much sense because an RDF structure 
designerr makes different decisions than an zzstructure designer; in 
zzstructure, you don't want to use many dimensions with few connections, 
whereas in RDF this is not a problem.)

As for how to view swad-chart, I would recommend going to the SWAD 
(root) node,

    file:/home/connolly/w3ccvs/WWW/2000/01/sw/swad-chart#SemWeb01

turn off all properties except rm:..., and choose "g:label" from 
"View/Show class/rm:subgoal" and "View/Show class/All others".

This gives you a view closer to the one seen in the actual graph.

> Probably, we 
> need some default properties corresponding with d.1, d.2 and d.3 in 
> zzstructure, which could be shown spatially in a special view.

Firstly, in which way would these be special and how would it allow them 
to be more easily shown spatially?

Secondly, I think that having 'default properties' really goes against 
the philosophy of RDF (IMHO). Well, you can use e.g. rdf:seeAlso if you 
just want to specify an associative relationship, but a 'generic' 
relationship that says nothing but, "these two nodes are related 
*somehow* but we don't tell you how (not even whether it's relevant to a 
human or not, etc.)"-- that doesn't sound useful.

-b.