> 'big' components that provide some service relevant to the business
> domain, use coarse grained interfaces). AFAIK there is not even the
> idea of making these interfaces as similar as possible to facilitate
> component integration.

This whole "big components with coarse grained interfaces" is a
canard. To me that is not a valid design constraint by choice. Rather
it is a design constraint based on the implementation approach
(e.g. CORBA and WS-* / XML).

Depending on the setting (say a large IT data center) and the
implementation technology (say Erlang rather than Java or C#) then
systems that qualify for the SOA label have been defined and proven in
production to be composed of very small pieces indeed with very
fine-grained interfaces.

As it happens that *same* technology (Erlang) has been used to build
those larger, internet scale components and interfaces. Nokia in fact
has a set of them in their product line. I would rather have a
technology that can go as fine-grained or as coarse-grained as
desired.

SOA is the future they say, but everyone seems to assume "big" is the
desired goal rather than the unfortunate consequence. I guess if one
size *has* to fit all then you have to go big. Well, then I think
there are better definitions of big than WS-*.

It's all kind of unfortunate really that XML-RPC begat SOAP begat 
WS-* and now we've managed to assume its appropriateness even its
Superiority with little critical evaluation. Marketecture.

> Look at Jim Waldo's approach to SOA using Jini; you'll definitely
> find formal aspects there.

I have, thanks. To the extent that SOA is defined, I think Jini makes
a very good SOA. Certainly more defined and capable (and standardized)
than anything else I have seen falling into the SOA bucket. People may
prefer WS-* / XML as the glue than Java. I am not crazy about Java,
but I do think it makes better glue. And then Jini's service model
does not rule out compatibility with WS-* / XML. I like that it can
accommodate variety and evolution in protocols, data, and granularity.

-Patrick





 
Yahoo! Groups Links

<*> To visit your group on the web, go to:
    http://groups.yahoo.com/group/service-orientated-architecture/

<*> To unsubscribe from this group, send an email to:
    [EMAIL PROTECTED]

<*> Your use of Yahoo! Groups is subject to:
    http://docs.yahoo.com/info/terms/
 


Reply via email to