David J. States
Thu, 31 Dec 1998 12:09:44 -0600
Ewan Birney wrote:
I think XML has a big part to play in serving up biological
data in a more compatible form than hard-to-digest ASCII formats. XML
fits very well with perl's 'just-do-it' philosohpy...
What I might suggest is the following things:
a) that you start a web site...
b) encourage other people to comment...
c) Come up with some conrete proposals
For both a) and c) I can strongly encourage you to do this inside the
'bioperl' web site and code base.
I put together a web page, http://www.ibc.wustl.edu/bioXML, with some links
and references relevant to the use of XML in biocomputing. Would be happy
to move this into the bioperl site at some point. The bioperl developers
mailing list does seem like an appropriate discussion forum (if the traffic
becomes excessive, we can always split it off).
NCBI is very interested in developing an XML interface to their databases,
but they have not finalized their approach. One alternative would be to
use the existing ASN.1 spec as the basis for a lexically correct XML data
island, but there are some constructs in ASN.1 that do not map precisely
into XML. For example, "CHOICE" in ASN.1 specifies one and only one of a
set of alternatives. These could be coded as "OPTIONAL" in XML, but then
the user might expect 0 or more than 1 occurence when in fact precisely 1
is appropriate. Developing an independent XML spec would also give them a
chance to rectify some issues locked into the existing ASN.1 spec and
perhaps to simplify the definitions somewhat.
I have not heard what EBI is planning. Perhaps something based on the
PDB is moving to San Diego, and the people involved have been quite devoted
to CIF, but again, I have not heard anything specific. There is a real
need for a biological entities (as opposed to crystallographic unit cells)
view of PDB, and XML would be a great place to start.
=========== Bioperl Project Mailing List Message Footer =======
Project URL: http://bio.perl.org/
For info about how to (un)subscribe, where messages are archived, etc: