W3C home > Mailing lists > Public > public-semweb-lifesci@w3.org > March 2009

Re: blog: semantic dissonance in uniprot

From: <samwald@gmx.at>
Date: Thu, 26 Mar 2009 10:57:53 +0100
Cc: public-semweb-lifesci@w3.org, phillip.lord@newcastle.ac.uk, markw@illuminae.com, curoli@gmail.com
Message-ID: <20090326095753.116190@gmx.net>
To: Pat Hayes <phayes@ihmc.us>, bparsia@cs.manchester.ac.uk
Pat Hayes wrote:
> [...] or use rdf: seeAlso, which was put in  
> the language pretty much for this reason, to be a blank carry-all for  
> a loose association with no formal properties.

I think that rdfs:seeAlso is the way to go, as I recently suggested in a similar discussion on the Linked Data mailing list [1]. Again, people there responded that rdfs:seeAlso was too weak, and that someone, somewhere needs to come up with a kindOfLike - like property. Then the discussion ended, without approaching any kind of idea on what this property should actually imply, and how we should get everyone to agree on this property. I guess the only thing that everyone can agree on is that the property should be symmetric, but everything else is very questionable. Even if everyone in the community could be made to agree on such a property (which is quite unlikely), it would be so fuzzy that you would need to decide on a case-by-case basis on how it should be understood -- but that could also be accomplished with rdfs:seeAlso.

Besides, I still think that it is not that hard to make correct owl:sameAs and owl:equivalentClass statements. In my opinion, it is not hard or wildly impractical to distinguish something swimming in a petri dish from a database record, or to create a class of 'protein ABC' that is defined by having the amino acid sequence 'XYZ'. Using OWL does not mean that you need to define everything, in fact, OWL is so geared towards incomplete information that it hurts sometimes (i.e., unintended effects of the open-world assumption).

Using RDF/OWL to describe database records is like using e-mail to send scans of handwritten letters. You are using new technology, but you are also clinging to the old way of doing things, which makes everything quite complicated and confusing (it is hard to glue stamps on an e-mail / it is hard to make owl:sameAs statements between database records from different databases). We can type e-mails directly into our computers, and we can use RDF/OWL to describe biological phenomena without centering everything around classical database records.

[1] http://www.mail-archive.com/public-lod@w3.org/msg01694.html

Cheers,
Matthias Samwald

DERI Galway, Ireland
http://deri.ie/

Konrad Lorenz Institute for Evolution & Cognition Research, Austria
http://kli.ac.at/

-- 
Neu: GMX FreeDSL Komplettanschluss mit DSL 6.000 Flatrate + Telefonanschluss für nur 17,95 Euro/mtl.!* http://dsl.gmx.de/?ac=OM.AD.PD003K11308T4569a
Received on Thursday, 26 March 2009 09:58:35 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:20:41 UTC