On 05.01.2015 16:29, Stian Soiland-Reyes wrote:
This is great stuff!
Thank you, Stian. We are glad to hear that.
I tried it with my ontology PAV - and it seems it is struggling a bit
because PAV (deliberately) don't have a defined domain and range on
object properties:
http://vowl.visualdataweb.org/webvowl/#iri=http://purl.org/pav/
Hence everything goes from and to Thing in the centre, which makes it
a bit clotted - I could not use the "Gravity" setting to space out the
properties.
In our manually made diagram I show this using additional "vague"
resource boxes - perhaps each "unbound" property could just get an
empty dotted box on the outside instead of going back to Thing?
http://pav-ontology.googlecode.com/svn/trunk/images/pav-overview.png
If there is a large set of properties without domain and range axioms in
an ontology, the current WebVOWL rendering is indeed not perfect. Your
PAV ontology is a good example for this. We will think about possible
alternatives to adapt or extend the splitting rules of VOWL 2 -
http://vowl.visualdataweb.org/v2/#splittingRules .
Here's a nice view that I liked, using PROV-O
http://vowl.visualdataweb.org/webvowl/#iri=http://www.w3.org/ns/prov-o
The visualization is nice but not 100% complete yet, as the current
WebVOWL implementation does not render properties with multiple domain
or range axioms properly at the moment (e.g., the property "hadActivity"
has more than one domain). We are working on that issue and it will
likely be resolved in the next WebVOWL release.
Is it possible to turn off the "Subclass of" label and only show the line?
Several people asked for that already, so we plan to include some kind
of "expert mode" with a "reduced notation" in the next WebVOWL release.
However, we do not want to remove that label from the initial
visualization, as it clarifies the direction of the subclass relation
and is important to make VOWL understandable to casual users, as we
found out in our evaluations - see our papers on VOWL linked at
http://vowl.visualdataweb.org/v2/#references
Would it be possible to save a link to a particular view (without
having to save the SVG)? That navigation state fits much better in
the # parameters than the iri= I would belive.
This is a tricky question, as it would require to save the node
positions of the VOWL graph in an annotated JSON or the like. The
current VOWL-JSON is not optimized for that, but we may incorporate such
a feature in one of the next releases (though I cannot promise).
I know it's not a proper ontology - yet people still like it for some
reason. Here however it fails with "Ontology could not be loaded.
Conversion failed. "
http://vowl.visualdataweb.org/webvowl/#iri=http://purl.org/dc/terms/
This is because the vocabulary misses some OWL constructs that are
expected by WebVOWL (e.g., an ontology IRI). We will try to fix it and
also allow the parsing of such RDFS vocabularies in the next WebVOWL
release.
Many thanks again for your feedback - very useful,
Steffen
On 19 December 2014 at 15:49, Steffen Lohmann
<steffen.lohm...@vis.uni-stuttgart.de> wrote:
Hi all,
we are glad to announce the release of WebVOWL 0.3, which integrates our
OWL2VOWL converter now. WebVOWL works in modern web browsers without any
installation so that ontologies can be instantly visualized. Check it out
at:http://vowl.visualdataweb.org/webvowl.html
To the best of our knowledge, WebVOWL is the first comprehensive ontology
visualization completely based on open web standards (HTML, SVG, CSS,
JavaScript). It implements VOWL 2, which has been designed in a
user-oriented process and is clearly specified at
http://vowl.visualdataweb.org (incl. references to scientific papers).
Please note that:
- WebVOWL is a tool for ontology visualization, not for ontology modeling.
- VOWL considers many language constructs of OWL but not all of them yet.
- VOWL focuses on the visualization of the TBox of small to medium-size
ontologies but does not sufficiently support the visualization of very large
ontologies and detailed ABox information for the time being.
- WebVOWL 0.3 implements the VOWL 2 specification nearly completely, but the
current version of the OWL2VOWL converter does not.
These issues are subject to future work.
Have fun with it!
On behalf of the VOWL team,
Steffen
--
Dr. Steffen Lohmann . Visualization and Interactive Systems (VIS)
University of Stuttgart . Universitaetstrasse 38 . D-70569 Stuttgart
Phone: +49 711 685-88438 .http://www.vis.uni-stuttgart.de/~lohmansn
