[MOBY-dev] data by reference - a request for comments

[Martin Senger]

> It can be claimed from the RDF of the service as a predicate, including the
> understood reference protocols (which should be described elsewhere in an
> RDF ontology).


I do not like using the RDF for mandatory features. The RDF, at least in the
current moby, is optional and should not carry anything that is significant
for the service and its behaviour. The data by reference should be part of
the main BioMoby API (or whatever we call it).


> I suggest using serviceNotes child element


I do not have a strong opinion where in the message this could be - I have
just used the tag name and location that was suggested in our February
meeting. The ServiceNotes, however, seems a bit a strange place when used by
the clients.


> and the service answers using the acceptRefs explaining what it have done.


Why? Why is not enough the reference itself? It is a URL and the URL has a
wel-defined protocol. And what if a service uses a different protocol for
different parts of the message (remember the purpose B)? No, the
ServiceNotes is not a suitable place for it.


> The service must answer a subset of the understood reference protocols to
> the client in order to maintain the contract.


And what about if the service does not want to return references, at all?
That should be an error, as well? No, I don't think so. A client just
suggests that she is able to accept references but there is no guarantee
that she can get it (many services will not have any provision for
references).


> A passive, static way is publishing the service capabilities about
> references in the RDF.


It can be there, why not, - but not as the main way.


> An active, dynamic way is creating a new service port (like we did with
> asynchronous services)


Do you mean a new method name? But for what? Not sure I understand how this
new method may help here (and not breaking the purpose B, for example).

Just my opinions.
Martin

-- 
Martin Senger
email: martin.senger at gmail.com,m.senger at cgiar.org
skype: martinsenger