[MOBY] [MOBY-dev] graphs of the services and data types

[Martin Senger]

Mark,
   Thanks for the answers.

> IMO, a service that outputs Simple articles should be considered to
> input into services that consume Collections of those types of Simples. 
> Similarly, services that output Collection articles should be considered
> competent to make inputs to services that consume Simple's.
>
   Yes, that was also my opinion. Okay, I will do that in the graphs.

> saying that services which consume **more than one type of object** in a
> single invocation should perhaps be represented differently in your
> graphics... perhaps a dotted line between the output object type and the
> consuming service to show that it isn't a *perfect* match of output to
> input, but that there is a "partial" match there.
>
   Yes, that was exactly my plan. Actually, that's the reason I have
installed a moby central locally on my laptop in order to register various
"weird" combinations of (fake) services without disturbing the "real"
central. (Now you see why I had suddenly some questions about the local
installation :-)).

> There is one connection conspicuously missing, and that is the output of
> a Blast can be input into the BlastParser service.  There may be others
> missing, but I think if you can figure out why that one is missing, the
> others might fix themselves :-)
>
   Good point. I will do that. (I already see my bug: it happened because
these inputs and outputs do not have registered any namespace. I'll fix
it.)
 
> What software is making that?
>
   graphwiz (dot, neato) + jMoby's API reads everything from the moby
central on-the-fly. I will have more comments on that software when the
rest of graphs is sorted out (it cannot be a pure java solution, but
almost; and yes, there are also perl modules to read the same graphicvs
files produced by this software). Very convenient package - we are using
it already in myGrid (the workflow graphs).
 
> Does it currently restrict connections only to those that consume
> appropriate namespaces as well as appropriate object types?
>
   It tries to match namespace/object-name (trying sequentially all
namespaces that were registered with that object). And then it does the
same for parent object, and again for its parent...
   I hope that I am using a right algorithm - but anyway I will describe
in details this algorithm in the documentation so you and others will be
able to correct me if I am mistaken.

   Okay, now I am going to code the next round on the graphs...

   Martin

-- 
Martin Senger

EMBL Outstation - Hinxton                Senger at EBI.ac.uk     
European Bioinformatics Institute        Phone: (+44) 1223 494636      
Wellcome Trust Genome Campus             (Switchboard:     494444)
Hinxton                                  Fax  : (+44) 1223 494468
Cambridge CB10 1SD
United Kingdom                           http://industry.ebi.ac.uk/~senger