[Biojava-l] Experimental DAS 1.900 implementation

[Thomas Down]

Hi...

I've started working on an experimental Java implementation of
the DAS 1.900 proposal, so that we've got something to play
with as the specification develops.  At the moment there's
just a Sequence service (vaguely equivalent to refererence
servers in DAS1).  I've got most of the pieces of a Features
service, and should be able to check that in in the next day
or so.

For anyone who's interested in tracking development, there's
code in the BioJava CVS repository (http://cvs.biojava.org/
for more info).  Checkout the module `xdas'.

What you get:

  - Prototype DAS sequence service

  - The SOAP toolkit it relies on.

  - An obligatory stock-quote demo ;)

To preempt the inevitable question, yes I have looked at other
SOAP toolkits (especially the IBM/Apache one).  While they're good
a lot of the time, they tend to use DOM to represent the SOAP messages
in-memory.  I know from past experience that this really isn't a good
idea where DAS is concerned -- sometimes you want to shift quite
serious amounts of data around (especially feature-tables), and I've
seen OutOfMemoryErrors (as well as poor performance) when the BioJava
DAS client was using a DOM based solution.

The new SOAP toolkit is based on S[t]AX event-based XML parsing,
so that should help scalability.


I'd be interested to hear any comments about this code...


   Thomas.