Re: OWL-S question: multiple atomic processes

On Nov 28, 2006, at 9:14 AM, Jo Vermeulen wrote:

> Dear Brian,

It's "Bijan", not Brian :)

> Thanks for your response! I was indeed thinking of defining my own  
> "Service" OWL class, which could contain multiple OWL-S processes,  
> and thereby would still do what I want. However, this is a custom  
> extension to OWL-S.
>
> Is defining the three distinct services as a composite choice  
> process a good alternative, or is it just a hack, abusing the OWL-S  
> spec? Can you still choose yourself which of the three services  
> will be executed?

Why not? :)

> From the spec:
>
> <quote>
> : Choice calls for the execution of a single control construct from  
> a given bag of control constructs (given by the components  
> property). Any of the given control constructs may be chosen for  
> execution.
> </quote>
>
> The sentence "any of the given control constructs may be chosen for  
> execution" makes me think it it not possible.

Well, that's true either way, yes? It doesn't say who is in control,  
but you could choose to invoke any of the services.

If the inputs are distinguishable, then I think it's fair to expect  
the system to pick the service that *can* process it. In the case of  
compatibility (either operation could handle the input) there is,  
indeed, no further constraint on the system.

> I am going to add more metadata to OWL-S services  as a custom  
> extension, so it is actually no disaster if I would have to define  
> a "service" of my own, on top of the three OWL-S services.

I would call it something else like "service collection" or "service  
set" but other than that, I would go for it. Why not :) Done right,  
it could get into the spec, I imagine.

What you won't get is toolkit support out of the box for it. But if  
you are just collecting services into a set, plus a little metadata  
on top, I don't see that this is a problem. as long as the collected  
services are described "normally", any owl-s toolkit should be able  
to extract and use the descriptions just fine. So you'll have  
graceful degradation.

Cheers,
Bijan "Not Brian" Parsia.

Received on Tuesday, 28 November 2006 09:40:25 UTC