[MOBY-dev] New Moby app

[Stian Soiland-Reyes]

On Tue, Jun 2, 2009 at 19:47, Dmitry Repchevsky
<dmitry.repchevski at bsc.es> wrote:

> Another issue is that Taverna yet to support schema type inheritance.
> (I don't blame them - it's a very tough task, I started to work on it, but
> have no time now, hopefully in July)

It's in our plans, but yeah, it would require quite a bit of reworking
and is hence not coming up immediately. We want to get rid of those
XML splitters and do some more kind of wizard-y 'configuration'
dialogue where you can tick the boxes for the granularity of inputs
that you actually want. This could include drop-down boxes for
abstract/subclassed types. You would then end up with a customized
WSDL service with input and output ports per your specification which
don't require any splitters in either end - comparable to how the
BioMart services works in Taverna.


As BioMoby deals with construction of data types a similar solution
might be interesting for BioMoby services in Taverna, as many Moby
workflows end up having various data object construction and parser
'processors' that are not really doing any execution - they are just
shims like the XML splitters.


> Is it possible to generate Taverna mixed workflow for moby/wsdl services?

Yes, but you would probably have to dig down to individual strings on
both sides with XML splitters and BioMoby objects/parsers.  You might
also need a few simple shims that do string joining etc. as WSDL
services 'in the wild' expect sequences, gene identifiers etc. in many
various kinds of syntactic formats.


-- 
Stian Soiland-Reyes, myGrid team
School of Computer Science
The University of Manchester