Anne Thomas Manes wrote:
> Let's use order entry as an example.
> 
> A chatty interface (fine grained) provides the following methods/operations:
> - createOrder
> - addItem
> - processOrder
> 
> And a fatty interface (coarse grained) provides the following
> method/operation:
> - submitOrder
> 
> The chatty scenario is reminiscent of distributed object technologies, where
> the client performs remote operations on the service-based Order object.

No, it's not "reminiscent of distributed object technologies".  A "chatty 
interface" is in fact a particular implementation that might work just fine in 
particular cases.  "Distributed object technologies" are not restricted to this 
particular implementation.  Bad or non-scalable designs can happen with any 
technology.

The problem with many "distibuted object technologies" of the past were that 
they tried hold the 8 fallacies of distributed computing as truths.  Systems 
that try and hide remote computing from the developers aid in helping them make 
poor designs.

What WS-* technolgies and "course grained" exchanges have created is a focus on 
idempotent operations.  By only making a single call and getting a single 
result, the operations are stateless from the clients perspective, and can be 
idempotent from the clients perspective if the server protects the operations 
associated with the inbound request with transactional semantics.

There is a case for every type of design that "works".  Some scale, many don't. 
  But "distributed object technologies" are not any more likely to create a 
"chatty interface" than WS-* or other technologies.

It's the design and the efforts of the engineers (or programmers or ignorant 
person placed in charge) that create designs which are in appropriate for the 
situation at hand.  Some technologies, such as mobile code, provide more 
opportunities to tune the implementation over time.

Gregg Wonderly





 
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