[Biojava-l] To Do Question

[David Block]

> -----Original Message-----
> From: Dickson, Mike [mailto:mdickson@netgenics.com]
> Sent: Friday, May 31, 2002 10:29 AM
> Subject: RE: [Biojava-l] To Do Question
> 
> 
> I avoided responding earlier because I just wasn't sure how to respond
> constructively.  I'm still a bit mystified that new software 
> specific to the
> life sciences is being developed to connect to a web services 
> directory.
> Why re-solve in possibly incompatible ways problems that platform and
> infrastructure vendors are already addressing?
> 
I agree with you.  I see BioMoby as being a web service registry and
a place for people to publicize their bio-specific clients.  These
clients can and should use public infrastructure (JAXR and SOAP::Lite),
but bioinformatics will always require tweaking (otherwise none of us
would have jobs :)

> JAXR already exists to provide a simple access mechanism in 
> Java to talk to
> web services registries.  The calls are easy to use.  I don't 
> know what soap
> layer is being targeted for Perl projects but I do know at 
> least SOAP:Lite
> includes a simple mechanism for access to UDDI.  There are a couple of
> open-source implementations of UDDI directories available now 
> to support
> private directory services.  Why create new mechanisms and 
> semantics for
> this?
> 

BioMoby is not.

> This is clearly me on a soapbox so take it for what its 
> worth.  One of the
> things I think has slowed progress and adoption of standards in life
> sciences is the constant tendency to reinvent platform 
> technology. There's
> been some really cool stuff done (like DAS for instance) that solves
> important scientific problems.  It seems to me that's a much 
> better place
> for us to focus and simply use the platform stuff as is (even 
> if it seems
> "yucky" or we think we can do it slightly better).
> 

You know, we started thinking about BioMoby before we knew about UDDI,
but Moby is now all about embracing and extending what's out there.

> Mike 

Dave