On Thu, Feb 07, 2002 at 11:30:09PM -0500, Guoneng Zhong wrote: > > But I do think that it depends a lot on the programmer's style. This > leads me to ask a tangential question. Has the phrase "Web services" > entered the bio world? Since we have people using Perl and Java and > Python (wonder if there will be a PHP one), and eventually Microsoft > might want to do something when the informatics side becomes successful, > wouldn't it be time to think about interoperability using HTTP (instead > of that Corba thing that is way too complex and rigid).
Yes, there's definitely been interest in this direction. From my viewpoint, SOAP looks quite attractive for bioinformatics because it's (potentially) good for `coarse grained' communication: where each message consists of a complex (and potentially polymorphic) data structure. One existing application is the DAS protocol: http://www.biodas.org/ This is essentially a `web service' in that it uses XML messages over an HTTP transport layer. It doesn't actually use SOAP envelopes (the original protocol predates the widespread use of SOAP), but it's likely to be replaced by a SOAP-based protocol in the not-too-distant future. [If you're interested, BioJava includes a complete DAS client library. There's also a server framework which makes it reasonably easy to set up DAS servers backed by BioJava databases]. > So do words like SOAP or UDDI or just web services have a place yet in > bioinformatics? Certainly, there are quite a few people (myself included) who would love to see some framework for service discovery. Right now, UDDI looks like a very promising solution to this, but to the best of my knowledge, applications which use it are still in the planning or early prototype stages. This will certainly change, though. You might also be interested in the following projects, which are looking at a web services approach to bioinformatics infrastructure: http://www.biomoby.org/ http://omnigene.sourceforge.net/ Regards, Thomas. _______________________________________________ Biojava-l mailing list - [EMAIL PROTECTED] http://biojava.org/mailman/listinfo/biojava-l