Anne, would you agree that a common component model at the meta-level of 
definition (though SCA, probably, is not at such level of abstraction) were 
much more usefully and made SOA infrastructure interoperability easier than an 
ESPERANTO-like language such as Web Services (universal but quite limited)?
- Michael

Anne Thomas Manes <[EMAIL PROTECTED]> wrote:                                  
I'm still on the fence about SCA. Does SOA need a common component model? 
Doesn't a common component model unnecessarily constrain the types of service 
you can create? Perhaps it makes it easier to package and deploy a composite 
application, especially in the Java world, but I'm not convinced that every 
service enablement environment has to directly support SCA for SCA-based 
systems to consume them. SCA-based systems can consume any web service 
described by WSDL--including any .NET web service. 

.NET doesn't do packaging the same way as Java -- you have assemblies rather 
than jars or wars or ears. Perhaps a jar-like packaging model would improve 
management of .NET, but I don't hear users complaining too loudly about .NET 
packaging.  

I also want to point out that I don't recall and can't imagine myself ever 
saying the following:


Anne Thomas Manes, research director at the Burton Group Inc., cited BizTalk as 
the number one offering in Microsoft's SOA efforts. 


The native communication framework in .NET (WCF) is the number one offering in 
Microsoft's SOA efforts. 

Anne

On 4/21/07, Michael Poulin  <[EMAIL PROTECTED]> wrote:                          
           
Well, while it's interesting to know who has longer beard, it does not help in 
understanding why are the beards of different colors. Now people talk about a 
common component model for SOA but how much is it different from the common 
model for data exchange for SOA? That is, if Microsoft was so active with 
regard to XML/WS-Security/SOAP, why it goes aside from the "Gang of 18 th"?
   
  Such separation is not new if one recalls early days of Java. The only 
problem is in the analysts and architects who play technology-agnostic games 
pretending that SOA allows mix and match just because it is SOA. How a 
wonderful SOA mediation solution (a la ESB) on .NET looks inside the Java-based 
Services environment?  
   
  Yes, such combination can work but only on quite limited set of technologies 
(like XML and,  probably, Web Services). What's about other service 
technologies like Java/Jini and CORBA? How is it possible to enforce technology 
limitations in the technology-agnostic business-oriented SOA? It seems that 
instead of technological integrity under the SOA umbrella we are pushed (again) 
into segregated, technology specific worlds. This is the direct threat to SOA. 
   
  If even 18 organizations that compete in many markets between each other 
could agree on SCA (finally), what's prevented Microsoft from being among them, 
especially if it moves that well into SOA? Oh, maybe it is a .SOA insteadÂ… 
  
- Michael

Steve Jones <[EMAIL PROTECTED] > wrote:                             Now I could 
be completely wrong, but I've always seen WCF as a development aid while SCA is 
a design and deployment aid.  What I mean by this is that WCF makes the job of 
consuming and developing individual technology (.NET) services much simpler 
(especially consumption) but SCA focuses on pan-technology services (BPEL, 
Workflow, Rules, EJB, WS, etc) both in terms of design and consumption. So WCF 
makes it easier to abstract the protocols away and do simpler service to 
service communication, WCF is a developer focused technology that helps make 
code simpler.  SCA however makes it easier to design, deploy and manage 
enterprise class service solutions, SCA is an architect and operation focused 
technology that helps make projects simpler.  

>From a historical  point SCA was first released in IBM Process Server which 
>debuted in 2005, so wasn't "keeping up" with WCF which wasn't released until 
>.NET 3.0.


 On 18/04/07,  Eric Newcomer <[EMAIL PROTECTED]> wrote:                         
            
 Nothing has happened with it so far at OASIS.  Also I believe one of the 
motivations for SCA was to keep up with WCF... 
  
  Eric


 ----- Original Message ----
From: John Evdemon <  [EMAIL PROTECTED]>
To: "[email protected] " < 
[email protected]>
Sent: Wednesday, April 18, 2007 1:44:00 PM 
Subject: RE: [service-orientated-architecture] Seeley on the MS Approach to SOA

   
 Many of the capabilities in WCF seem to be available in some form within SCA 
(at least that was my take on it after I read some of the  SCA papers – I 
haven't kept up with it since it went to OASIS). 
  
   
    From:  service-orientated- architecture@ yahoogroups. com [mailto: service- 
orientated- architecture@ yahoogroups. com] On Behalf Of  Stefan Tilkov
Sent: Wednesday, April 18, 2007 2:44 AM
To: service-orientated- architecture@ yahoogroups. com
Subject:  Re: [service-orientated -architecture] Seeley on the MS Approach to 
SOA 


  
    On Apr 17, 2007, at 6:39 PM, Gervas  Douglas wrote:

> Microsoft doesn't support the Service
> Component Architecture (SCA) and Service Data Objects (SDO)
> specifications, which offer similar functionality to .NET. 

I wonder what this is referring to - what would qualify as the .NET  
equivalent to SCA and SDO? Not that I'm a big believer in these two 
specs, just curious.

Stefan
--
Stefan Tilkov,   http://www.innoq. com/blog/ st/

 







  



          
---------------------------------
Ahhh...imagining that irresistible "new car" smell?
 Check out   new cars at Yahoo! Autos. 
     
               
        



 
     
             
         

---------------------------------
Ahhh...imagining that irresistible "new car" smell?
 Check out  new cars at Yahoo! Autos. 
     


               
        




 
     
                       

       
---------------------------------
Ahhh...imagining that irresistible "new car" smell?
 Check outnew cars at Yahoo! Autos.

Reply via email to