Re: Visualization of domain and range

All,

Thanks for an interesting discussion. It is interesting how many visual
styles different people can come up with. Is there a place that gathers
these up for comparison? I made a twine at
http://www.twine.com/twine/12df6wvd7-1f0/owl-the-web-ontology-language-visualization.
It includes the examples from this thread; feel free to add more.

I have developed my own visual styles over the years and would like to see
if I got it wrong :-)

Is there a good source for a description of the commonly-accepted
conventions?

I also love the fact that Dan started to enumerate the OWL tasks that the
visual should support. We can't evaluate a visualization unless we know the
tasks that it must support.

With the new OWL 2 constructs, I'll have to go back to the drawing board for
things like property chains and keys. I was excited to find out that the
disjointUnionOf construct came through, because I already made the visual
abbreviation for the verbose OWL 1 expressions.

Regards,
Tim Lebo

ps - svn co
http://svn.foaf-project.org/foaf/trunk/xmlns.com/htdocs/foaf/spec/images/only
gives a jpg, svg, and graffle. Can you include a URL for the foaf.owl
spec that inspired these works of art? I want to make sure I work on the
same input.




>
> In http://www.flickr.com/photos/danbri/1856478164/ ([4]) I try to do too
> many things at once:
>  * show the classes that each property is used with
>  * show sub-property relationships
>  * show sub-class relationships
>  * show some typical properties
>  * show attachment points for "friends of FOAF" namespaces (DOAP, SIOC, DC,
> Geo etc), with classes and with sample properties
>
> This is a lot of information.
>
> I did try to make a "gradual reveal" slideshow version, building up from
> something simple. It wasn't great. The layout was done by hand to minimise
> crossovers, and looking at it, I think the whole structure could be
> twisted/stretched to be more evenly presented. It was fiddly to do though.
>
> A sample of instance-data would probably convey most of the same
> information about domain/range, and would allow subclasses reasonably too.
> Sub-property would remain hard...
>
> If anyone wants to mess around with the FOAF example, source data in
> OmniGraffle format is here and also in SVG: just do "svn co
> http://svn.foaf-project.org/foaf/trunk/xmlns.com/htdocs/foaf/spec/images/"