Re: [ws-caf] How many ALSs in an activity?

[Ricardo Jimenez-Peris <rjimenez@fi.upm.es>]

Mark Little wrote:

> When going over what the Context model assumed to ALS-to-activity 
> bindings, the following questions arose:
>  
> Do multiple ALSs register for a single protocol? How does the context 
> service know how many ALSs must be registered for a particular protocol?
>  
> I hope we've been clear in the UML/text that the intention was always 
> that multiple ALSs could register with a context service (aka activity 
> service). However, what wasn't clear was the role of the 'type' field 
> in the context. This was intended to map to an implied set of ALSs and 
> recipients of the type could infer from it what was in the context. 
> Defining *what* the type is is within the specification IMO, but 
> saying *how* it maps to specific combinations of ALSs should be 
> outside the scope of the spec. What I mean is that we shouldn't say 
> that urn:foo:bar:1234 is a transactional protocol-type with only a 
> single transaction ALS registered with an activity, whereas 
> urn:foo:bar:5678 is a secure+transactional protocol-type. Other 
> specifications should do that. (It's the same as defining the formatID 
> in the OTS context). 

Thinking about ALS combinations I got the following question.
Can an ALS participate in different ALS combinations at the same time?
That is, could be the transactions ALS be used in combination with ALS1 
by some
applications and with ALS2 by others? This would be useful, since 
different applications
have different needs. I am not sure whether this is possible in current 
spec.

Ricardo

-- 
Ricardo Jimenez-Peris
Associate Professor
Universidad Politecnica de Madrid
http://lsd.ls.fi.upm.es/%7Erjimenez/