[MOBY-l] Incoporating existing services

[Ken Steube]

This is just Martin's message CC'd to Doug Brown and with a new subject.
Please include Doug in future e-mails on this topic

Doug you might go to biomoby.org and sign up for moby-l at biomoby.org.

Ken


On Mon, 20 Oct 2003, Martin Senger wrote:

> > Let's bring the discussion onto the moby-l list and get everyone talking
> > about it.  Martin, Doug, and Catherine all have the same barrier with
> > wrapping MOBY around their pre-existing services - I'm sure we can find
> > a way to deal with properly if we, together, put our minds to it.
> >
>    Okay. In order to remove barrier for me I need:
>
>    1) To be able to register my service by sending its WSDL or a URL to
> the WSDL (I do not mind which one), to be able to say "this service is not
> MOBY-native". The API says now "currently only ;moby; protocol is
> supported.
>    I do not like much the name 'soap' for it because it looks like the
> Moby-native services use something else (which is not true, they use soap
> as well). I would prefer probably name 'external', or 'external-wsdl'. But
> this is not critical.
>
>    2) To be able to register input and output data types in the same way
> as you do with the Moby-native services.
>
>    3) To be able to create (and use) a Query service object with a
> different Protocol (not only with value 'moby' as the API dictates now).
>
>    4) To add definition what is returned back by the retrieveService
> method (the current API says "yet to be defined".
>
>    5) To consider whether The Service List Response Object should not
> contain also wsdl (or a URL of a wsdl) of the found service (if it is
> 'soap' ('external') service.
>
>    6) Later, to consider to allow to register 'soap' ('external') service
> without registering data types and extract their automatically from the
> WSDL provided (I think that it will not be possible in all cases, but it
> may help tremendously for those who have simpe WSDLs).
>
>    From the discussions in Vancouver, I understand that this can be
> accomplished quite quicky. Having that, I can try to register something
> and to see other barriers (if any) I encounter.
>
>    Martin
>
>

-- 
-------------------------------------
Ken Steube            steube at sdsc.edu
San Diego Supercomputer Center @ UCSD
San Diego, California             USA