W3C home > Mailing lists > Public > public-semweb-lifesci@w3.org > June 2011

Re: Trying to summarise: Semantic free identifiers

From: Andrea Splendiani <andrea.splendiani@bbsrc.ac.uk>
Date: Tue, 21 Jun 2011 19:41:12 +0000
Message-ID: <339406501919350@jngomktg.net>
To: James Malone <malone@ebi.ac.uk>
Cc: "andrea splendiani (RRes-Roth)" <andrea.splendiani@rothamsted.ac.uk>, Helena Deus <helenadeus@gmail.com>, Matt Vagnoni <matthew.vagnoni@uth.tmc.edu>, Michel_Dumontier <michel_dumontier@carleton.ca>, "Sivaram Arabandi, MD" <sivaram.arabandi@gmail.com>, "M. Scott Marshall" <mscottmarshall@gmail.com>, Chime Ogbuji <chimezie@gmail.com>, "MMVagnoni@mdanderson.org" <mmvagnoni@mdanderson.org>, HCLS <public-semweb-lifesci@w3.org>, Jonathan Rees <jar@creativecommons.org>
Skipping  the line of reasoning that leads to these conclusions...

I think one of the main point is the role of the 'web'.

Whether we are talking about terminologies encoded in OWL/RDF, or about a
distributed web-based information space. I bet we could make a test and see
the correlation between who prefers opaque vs transparent ids, and who
prefers OWL-apis vs Jena.

In my opinion, it is desirable to have opaque identifiers for large medical
terminologies. But transparent? identifiers work better for relations and
for tight dictionaries (e.g.: DC). It doesn't really change that much if you
use tools or not at the end (at least for me).

ciao,
Andrea



Il giorno 21/giu/2011, alle ore 20.13, James Malone ha scritto:

> So.. a long but useful discussion. That will teach me to open my big mouth
:)
> 
> Is this fair as the PRIMARY reasons for this difference in opinions:
> 
> 1. Having semantic information such as a label in a URI makes it easier
> to, at a glance, grasp some sort of meaning of a class/predicate and makes
> SPARQLing and looking at RDF easier.
> 
> 2. NOT having semantic information in a URI ensures class definitions need
> to be looked up before they can be used, hence, reducing ambiguity and
> that it potentially improves maintainability.
> 
> Can 1 be resolved by tooling? Seems to me 2 is happening already and will
> grow as practice in a lot of the bio-ontology community. If there is a
> lack of tooling surely this group should be looking at doing something
> about that - funding, lobbying..
> 
> James
> 
> 
> 
> 
> 
>> Hi,
>> 
>> I think there is some confusion going on on the subject.
>> 
>> We need to name things in an unique way. In many cases codes are just the
>> best option. No wonder we all have tax-codes and the like, it's easier
>> than
>> to try to find a unique name based on some attributes.
>> 
>> The case of terminologies is an interesting case, as we need to name
>> terms.
>> There is a temptation to use the 'face value' of the term as a name, as
>> opposed to a code. The former is clearly opening the doors for
>> ambiguities,
>> in this context.
>> 
>> Beside terminologies, there are many other cases where you name thing:
>> 
>> rdf:type
>> 
>> owl:Class
>> 
>> Is there a need for these to be semantically opaque ? I don't think so,
>> they
>> are good for mnemonics and the formal meaning is clearly defined
>> elsewhere.
>> 
>> The original thread didn't start from somebody questioning GO terms...
but
>> the need to replace 'partOf' with a code.
>> 
>> To tell a funny story... I have an (unrelated) homonymous in my home town
>> (which is a bit weird given the size of the town and the frequency of my
>> last name). Given the identifier clash... I ended up receiving funny
>> things,
>> like love letters or urgent calls from unknowns... (not sure I missed
some
>> as well...).
>> Now, when i went to register a website, which one would be better:
>> mydomain/AndreaSplendiani
>> mydomain/001
>> 
>> I cannot really see any reason for the latter, and several reasons
against
>> it.
>> Does mydomain/001 protects friend and lovers of my homonymous from
>> confusion
>> ? Most likely not.
>> 
>> ciao,
>> Andrea
>> 
>> 
>> Il giorno 21/giu/2011, alle ore 18.46, Helena Deus ha scritto:
>> 
>>> Other standards (outside of semantic web) saw the need to rely on
>>> numeric
>> identifiers, even if that created a burden for their users
>>> e.g. in SNOMED Lung = T-28000
>>> 
>>> Of course it is a pain to query SNOMED with "all the diseases that
>>> affect
>> T-28000".
>>> But the fact is that despite the inconvenience of having to fetch that
>> identifier prior to the query, SNOMED is widely used.
>>> 
>>> What is so special about semantic web identifiers that they don't need
>>> to
>> follow the same path?
>> 
>> Andrea Splendiani
>> Senior Bioinformatics Scientist
>> Centre for Mathematical and Computational Biology
>> +44(0)1582 763133 ext 2004
>> andrea.splendiani@bbsrc.ac.uk
>> 
>> 
>> 
>> 
>> 
> 
> 
> -- 
> European Bioinformatics Institute,
> Wellcome Trust Genome Campus,
> Hinxton,
> Cambridge, CB10 1SD,
> United Kingdom
> Tel: + 44 (0) 1223 494 676
> Fax: + 44 (0) 1223 492 468
> 

Andrea Splendiani
Senior Bioinformatics Scientist
Centre for Mathematical and Computational Biology
+44(0)1582 763133 ext 2004
andrea.splendiani@bbsrc.ac.uk
Received on Tuesday, 21 June 2011 19:42:02 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 14:52:47 UTC