> In these days I'm trying to understand which information integration
> systems (ontology-based) are currently available. In many papers I've
> read about TAMBIS, but since the 2000 up to now I can't understand
> what's happened: no official papers, no updated website, no mentions
> among the past projects of the Manchester University
> (http://img.cs.man.ac.uk/). This article
>
>(http://ieeexplore.ieee.org/Xplore/login.jsp?url=/iel5/6383/17062/00787647.pdf)
> says that it's available at http://img.cs.man.ac.uk/tambis, but nothing
> is there. (01)
I have recently asked Robert Stevens about TAMBIS; here are some
excerpts from our communication (cc:-ing both Carole and Robert for
further comments if needed, and for the record): (02)
[WK] I have a historical question: has the TAMBIS system ever been used
in a practical setup -- has it been accessed and used for scientific
purposes by people outside of the development team? (03)
Carole Goble wrote:
> no
> there is a commercial version and a new version as part of ComparaGrid (04)
Robert Stevens wrote:
I think there are a whole range of reasons why TAMBIS only really ever
worked at the level of research prototype: (05)
1. We could ask good questions, but the underlying resources
couldn’t answer them. The ontological model we used enabled us to ask
precise, well formed queries where only ones biological valid according
to our model were able to be constructred. The resources simply were not
able to answer them. Things are not a whole lot better now in many
respects. At least a translation in to something nasty like RDF would
enable me to use some kind of query language and this would be an
improvement.
2. Our ontological model could have been a lot better. We know more
now…. We conflated descriptions of biology and descriptions of
bioinformatics. Similar discussions are going on in the HCLS at the
moment. I’m not sure that a cleaner model would have made any practical
difference to questions, but maintenance, extension, etc. would have
been easier.
3. We modelled too close to the resources. A little more distance
from the resource would have allowed easier extension etc.
4. Starting now, we’d use pre-existing ontologies. The main part of
the TAMBIS ontology were SequenceComponent, biologcalFunction,
biologicalProcess, CellularComponent, Species and some stuff on protein
structure. Sound familiar? Today, I’d just import the GO, So, and
others. This would give me access to many resources marked up with
ontologies and thus better querying.
5. web services would have been lovely. We wrapped and scraped Web
pages. WS would have given more solid programmatic acess and again made
querying easier.
6. Choice of middleware and target for query re-writing is also
more interesting these days. One can of course think of RDF and
re-writing to an RDF query language. I’ve also thought of re-writing to
Taverna workflows. (06)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx (07)
|