Re: [wsbpel] implicite links of the runtime engine (was: Implicit<sequence> macro)

[Assaf Arkin <arkin@intalio.com>]

David RR Webber - XML ebusiness wrote:

>Assaf,
>
>There's a very big difference between a model approach,
>and the sort of interface specification I'm talking about.
>
>An abstraction layer contains all the programming support
>you need - but based around your functional set - not the
>necessarily the functional set of the target(s).
>  
>
Then I guess I missed the point.

To write my functional set as far as the process is concerned is 
something I can do using a WSDL interface. To write someone else's 
function set as far as my process is concerned is something I can also 
do using WSDL. Or not, but then I won't know what service they offer 
than I can use ;-)

Mobility, of the ability to execute in different environments, is given 
since the definition is based on the interface. It is not based on a 
particular protocol or end-point. So a process that uses service X with 
protocol Y in one environment would just as well use service A with 
protocol B in another environment. As long as it understands what the 
communication pattern is.

arkin

>This gives you better interoperability - and ability to support
>lots of different environments thru one common interface,
>and publish bindings which bring you in a wider audience.
>
>Adopting one flavour wholesale is not only a 'cheap fix',
>but limits you drastically long term.
>
>Thanks, DW.
>=================================================
>  
>