Apps Messaging - Semantics of a Message

[John Taylor]

On 25 Apr 2007, at 15:58, Paul Harrison wrote:

>
> On 16.04.2007, at 11:50, Paul Harrison wrote:
>
>> This extended "application messaging" use case seems to overlap  
>> hugely with the "Parameterized Application" use case of the  
>> Universal Worker Service (UWS-PA) that the Grid WG has developed.  
>> The UWS was primarily designed as a client-server protocol, so  
>> already has a SOAP profile, and will have a pure REST profile by  
>> Bejiing interop. However,  it would be conceivable to create an  
>> intra-desktop profile to meet the "applications messaging" use  
>> case, all using essentially the same registry-storable message  
>> description metalanguage.
>
> Although this message was conspicuously ignored, being stubborn I  
> will try again.

Hi Paul,
I think you have a good case here.  This may well be a better fit to  
invoking application-specific "task-like" functionality than PLASTIC  
is, even on a single machine.   The self-describing features of CEA  
make it a good solution.
It's conceivable that you could come up with a UWS-PA/CEA profile  
that meshed quite well with PLASTIC/SAMP, allowing people to reuse  
libraries (and brain power).   We should discuss it at Beijing.

John


>
> I note that CEA (UWC-PA) already has many of the features proposed  
> for a PLASTIC 1.0, namely
>
> http://www.ivoa.net/twiki/bin/view/IVOA/PlasticOnePointOh#AddSecurity
> http://www.ivoa.net/twiki/bin/view/IVOA/PlasticOnePointOh#MessageIds
> http://www.ivoa.net/twiki/bin/view/IVOA/PlasticOnePointOh#AsynchOnly
> http://www.ivoa.net/twiki/bin/view/IVOA/PlasticOnePointOh#ParamTypes
> http://www.ivoa.net/twiki/bin/view/IVOA/PlasticOnePointOh#ParamsByName
> http://www.ivoa.net/twiki/bin/view/IVOA/PlasticOnePointOh#UrlParams
>
>
> However, even if you want to press on along separate tracks to the  
> Grid WG, I would be interested in collaborating over the mtypes  
> idea (though I am not convinced that it is sufficiently different  
> from a UCD or a UType to have a separate name), as CEA does lack a  
> machine decidable description of
>
> * what an application actually does
> * what a parameter means
>
>
> Paul Harrison
> ESO Garching
> www.eso.org
>