Herbjörn, the book of GOF is not called 'Object-oriented Pattrens'; this is why 
it does not obfuscate object-orientation. Thanks to David Ch., I will look 
closer at the Erl's pattrens.

Right now, here are two examples of those patterns: 
1) "Version Identification (Orchard, Riley) 

How can
consumers be made aware of service contract
version 
information?"
2)"Legacy Wrapper (Erl, Roy) 

How can wrapper
services with non-standard contracts be
prevented 
from spreading indirect consumer-to-implementation coupling?"

This looks like Erl's "service contract " has almost nothning to do with OASIS 
SOA RM's service contract. Moreover, it seems that Erl's "service contract " is 
defined by WSDL, i.e. it is a Web Service. As we know, Web Service is just an 
interface and service contract, according to the standard, is much more than WS 
interface. Thus, are we talking about Web Services or SOA service in "Version 
Identification " pattern? After all, why a consumer might be interested 
explicitly in the interface version rather than in the service version? To the 
consumer, it does not matter what has been chnaged in the service; the consumer 
cares only about the answer to 'Should I change anything on my side and do I 
want to use this service after the change?' I believe that versioning separate 
parts of the service results in the consumer's nightmare, i.e. the service that 
exposes such versioning is a dead/unused service. But, one moment, this pattern 
is abot Web Service, not
 SOA service. Thus, it is OK.

Another example - Legacy Wrapper -  is a pattern for purely interface-based 
problem derived from Web Services again. Can anybody explain me how a legacy 
system wrapped with "non-standard contracts " or even "standard contracts ", 
i.e. interfaces, can become a Service while the legacy system itself does not 
demonstrate any implementaion of SO principles? Since when a car door placed on 
a cart made it a car? This is what I call obfuscation of service orientation.

- Michael




________________________________
From: herbjorn.wilhelmsen <[email protected]>
To: [email protected]
Sent: Friday, January 16, 2009 9:00:55 PM
Subject: [service-orientated-architecture] Re: Vaughan reviews Erl's Tome


>  SOA Design Patterns is published in 2009, when OASIS SOA RM is 
known
for long enough to understand (accept or deny) that Service Inventory
is not a pattern of service orientation, it is a pattern of technology
inventory, IMO.
>
> - Michael

Some of the patterns that are presented in the book are applicable in
other areas as well. This is in my opinion no reason for excluding
those patterns. And if you would want to exclude such patterns I have 
a question for you: What patterns in the SOA book is only applicable 
in the field of SOA? I would be very surprised if you found
even a single pattern with that characteristic.

Comparing with the GOF patterns book: As you yourself said quite a 
few of their object-oriented design patterns are not object specific, 
meaning that they can be applied to other areas than object-oriented 
development. Do you think that the GOF design patterns book 
obfuscated the meaning of object-orientation?

As I see it the important thing is that the patterns are appicable in 
the field of SOA and that the patterns are described specifically 
with SOA in mind.

- Herbjörn Wilhelmsen

    


      

Reply via email to