W3C home > Mailing lists > Public > public-semweb-lifesci@w3.org > July 2004

Re: SemWeb BOF @ ISMB 04

From: Phillip Lord <p.lord@russet.org.uk>
Date: 27 Jul 2004 14:04:45 +0100
To: Sean Martin <sjmm@us.ibm.com>
Cc: public-semweb-lifesci@w3.org, <wangxiao@musc.edu>
Message-ID: <vfd62htxci.fsf@rpc71.cs.man.ac.uk>

>>>>> "Sean" == Sean Martin <sjmm@us.ibm.com> writes:

  >> I believe one of the main problem in life science is the same as
  >> the original problem that Eric Neumann has posted.  That is how
  >> to refer part
  Sean> of
  >> non-RDF document.

  >> In my cases, my concern is this.  For some applications, such as
  >> data
  Sean> submission
  >> and retrieval, I actually wanted the gel encoded in a "compact"
  >> form.  Personally, I don't think XML is a good solution because
  >> there are more overhead than the actual payload.  Neither will
  >> RDF a good solution for
  Sean> the
  >> same reason.  It is not that difficult to create an arbitrary
  >> format to encode the gel data in a compact text based format but
  >> then I don't know
  Sean> how
  >> the URI should be assigned to each individual spot.  The URI for
  >> the spot
  Sean> is
  >> important because if a spot is IDed or used to perform MS.  The
  >> URI can
  Sean> make
  >> it easy to associate other type of descriptions.

  Sean> If I have understood you correctly, you are describing one of
  Sean> the use cases for the Life Sciences URN identifier. A single
  Sean> LSID can describe and give remote programatic access to both
  Sean> your data (the gel data in the format you invented or some
  Sean> other standard format appropriate to that data if one exists)
  Sean> and any amount of metadata in RDF that describes that data and
  Sean> its cross-relationships with other entities.

I don't think the this is exactly the problem. My reading was that he
is worried about how to refer to part of a larger resource (in this
case a 2D gel image) when that resource itself is not represented as
RDF or XML. 

Of course, LSID's would help here; you could have one LSID for the
overall image, another image for the spots, and then the LSID metadata
facilities to link the two. And you could use the LSID metadata to
describe, perhaps, where in the image the spot is. 


Cheers

Phil
Received on Tuesday, 27 July 2004 09:08:39 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 18:00:39 GMT