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

Re: Identity of URIs. WAS: Re: Is there an NCBI taxonomy in OWL ?

From: Xiaoshu Wang <wangxiao@musc.edu>
Date: Mon, 02 Mar 2009 22:43:18 +0000
Message-ID: <49AC6106.4010001@musc.edu>
To: Andrea Splendiani <andrea.splendiani@bbsrc.ac.uk>
CC: Kei Cheung <kei.cheung@yale.edu>, Alan Ruttenberg <alanruttenberg@gmail.com>, Chris Mungall <cjm@berkeleybop.org>, public-semweb-lifesci hcls <public-semweb-lifesci@w3.org>, Maryann Martone <maryann@ncmir.ucsd.edu>


Andrea Splendiani wrote:
> One thing that I think would be very useful, though it poses some  
> semantic problem... is the possibility to assert equivalence in rdf.
> At the moment equivalence can be asserted only in owl (and this  
> implies a distinction between individuals, properties, classes...).
>
> But that is a lower level statement we can make about URI.
>
> For instance, I can say that:
>
>   
>> http://dbpedia.org/resource/Dopamine_receptor
>>     
> is the same uri as:
>   
>> http://dbpedia.org/resource/DopamineReceptor
>>     
>
>
> And I want to say this independently from the interpretation  
> associated to this URI, as my knowledge about the identity is a-priori:
> if this leads to inconsistency, then this is because what is expressed  
> is inconsistent, not because of this equivalence.
>
> Is there a way to express this in RDF ? Don't think so...
>   
I am not sure exactly what you intend.  Do you mean that you want to 
assert the equivalence of the URI as the symbols or you want to assert 
that the referent of the two URIs are the same?  If your intension is 
the latter, I don't think RDF offers this kind of vocabulary.  But you 
can always mint your own terms just as with all other concepts that is 
not in RDF, right?

If you want to assert the equivalence of two URI, but not the resource 
that they references.  There isn't a straight-forward answer but there 
can be work-around.

For instance, you can design a class, say URI, which has a string 
property that confirms to URI spec.  Of course, if you want, you can 
also make it several properties according to the URI spec.  Then, you 
can create instance of this URI class, which URI can be either a b-node 
or you can designate another URI to it.   This would allow you to make 
assertions on URI rather than its referent..

Such an inconvenience of describing URI is due to the incomplete syntax 
of URI.  Among the three essential things of the Web, URI, 
Representation, and Resource, only Resource is "conveniently* in URI's 
referential realm.  I have suggested a solution in [1] to offer some 
syntactic sugar.  In that proposal,

"http://dbpedia.org/resource/Dopamine_receptor?" would by definition 
denote the URI.  This would greatly simplify the task.

Xiaoshu 

1. http://dfdf.inesc-id.pt/tr/uri-issues
> ciao,
> Andrea
>
> Il giorno 27/feb/09, alle ore 03:03, Kei Cheung ha scritto:
>
>   
>> I gave the following neuroscience URI examples in my biordf talk at  
>> C-SHALS yesterday.
>>
>> http://dbpedia.org/resource/Dopamine_receptor
>> http://purl.org/ycmi/senselab/ 
>> neuron_ontology.owl#Dopaminergic_Receptor
>> http://purl.org/nif/ontology/NIF-Molecule.owl#nifext_5832
>>
>> I pointed out that the last one might be a possible solution. There  
>> might be hope. :-)
>>
>> -Kei
>>
>>
>> Alan Ruttenberg wrote:
>>
>>     
>>> So I count three different sets of URIs for NCBI taxonomy so far. :(
>>> -Alan
>>>
>>> On Fri, Feb 27, 2009 at 11:37 AM, Chris Mungall  
>>> <cjm@berkeleybop.org> wrote:
>>>
>>>       
>>>> also..
>>>>
>>>> part of the NCBI taxonomy is in NIF Organism:
>>>>
>>>> http://ontology.neuinfo.org/NIF/BiomaterialEntities/NIF-Organism.owl
>>>>
>>>> See also:
>>>>
>>>> https://wiki.neuinfo.org/xwiki/bin/view/Main/NIFSTDoverview
>>>> http://neuinfo.org
>>>>
>>>>
>>>> On Feb 25, 2009, at 4:17 PM, andrea splendiani (RRes-Roth) wrote:
>>>>
>>>>
>>>>         
>>>>> Thanks!
>>>>>
>>>>> It'2 240M, but compressed is only 9.
>>>>> I wonder whether there is some architecture to transparently  
>>>>> transfer
>>>>> compressed ontologies...
>>>>>
>>>>> Ciao,
>>>>> Andrea
>>>>>
>>>>> -----Original Message-----
>>>>> From: Chris Mungall [mailto:cjm@berkeleybop.org]
>>>>> Sent: 25 February 2009 20:53
>>>>> To: andrea splendiani (RRes-Roth)
>>>>> Cc: public-semweb-lifesci hcls
>>>>> Subject: Re: Is there an NCBI taxonomy in OWL ?
>>>>>
>>>>>
>>>>> On Feb 25, 2009, at 11:58 AM, Andrea Splendiani wrote:
>>>>>
>>>>>
>>>>>           
>>>>>> Hi,
>>>>>>
>>>>>> I was looking for an NCBI Taxnomoy in OWL, but I didn't find it  
>>>>>> (or
>>>>>> better, could find fragment from other projects...)
>>>>>>
>>>>>> What is strange though, is that on the obo foundry website
>>>>>> (berkeleybop.org/ontologies) there are notes on the ncbi taxonomy
>>>>>> representation in owl... but not the representation itself.
>>>>>>
>>>>>>             
>>>>> Temporarily dropped from the summary page but still available at  
>>>>> the
>>>>> usual URL
>>>>> http://purl.org/obo/owl/NCBITaxon
>>>>>
>>>>> (warning: large..)
>>>>>
>>>>>
>>>>>           
>>>>>> Does anybody have some hint about where I can fin an OWL version ?
>>>>>> Or even an RDF version ? Even better would a sparql endpoint
>>>>>> containing it...
>>>>>>
>>>>>> best,
>>>>>> Andrea Splendiani
>>>>>>
>>>>>>
>>>>>>
>>>>>>
>>>>>>             
>>>>         
>>>       
>
>
>   
Received on Monday, 2 March 2009 22:44:07 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:20:40 UTC