Why not continue to support this query link for srs5?
http://<ipname>/srs/srsc?[<dbase>-id:<id>]
I plan to. It seems cleaner than the one in the
default SRS5 and has the great advantange that it is
already a standard (other's servers and software like Andy's
points to such).
Is there software dependent on this srs5 format?
http://<ipname>/srs5bin/cgi-bin/wgetz?[database-id:id]
I prefer to change this on my server to something a little
simpler, such as
http://<ipname>/srsbin/wgetz?[database-id:id]
Is the "5" in srs5 needed for any software? It would seen more
maintainable to leave the version number out, to allow for other's
links and software to have a standard path (and as I said, I think
the srs4 path should be maintained by servers who switch to srs5
to maintain for existing links).
For Andy's benefit, this syntax in SRS5 will let one query several
databanks:
wgetz?[libs={db1 db2 db3}-id:queryterm]
or for joint queries:
wgetz?[libs={db1 db2 db3}-all:query1]&[libs-org:query2]
Are there any better alternate forms of these? Esp. for the multiple
Genbank or EMBL "databanks". E.g., I have Genbank, Genbanknew,
and GBEST, because they are much easier to maintain separately then
as one indexed DB, and european sites have EMBL and EMBLNEW. Is there a
simpler syntax to query all the Genbank or EMBL at a site?
Something like "[nucleicDB-acc:accessionNumber]" would be nice
to support at all srs sites where EMBL or Genbank or DDBJ databanks reside.
One problem for my uses has been the spaces in the {db1 db2 db3} list.
I'd like to see commas as an alternate delimiter, as in {db1,db2,db3}.
To answer another of Andy's questions, I think the switchover to SRS5
will take more time. I've been spending a large chunk of time
recently getting the indexing parser (icarus) instructions to work
properly for the databanks I support (genbank, pir, some others).
Due to limited disk space, I will have to get rid of srs4 soon to
make way for srs v5 indices, but other servers may opt to wait
until srs 5 indexing details are worked out.
-- Don
--
-- d.gilbert--biocomputing--indiana u--bloomington--gilbertd at bio.indiana.edu
