Partager via


Here We Go Again...

Despite the fact that there is no one size fits all approach for SOA, Burton Group seems to be insisting that in the future everyone will use REST for SOA (although they were at a Web conference when they made this statement).

I'll buy the argument that REST is the Web and the Web is REST - no problem.  

I won't buy that all services are stuck with using lousy transports like HTTP (just because its a service doesn't necessarily mean its a web service)...

Comments

  • Anonymous
    May 31, 2007
    Some will never learn... Related to this misconception of Burton Group: http://soa-eda.blogspot.com/2006/08/granularity-of-services.html http://soa-eda.blogspot.com/2007/05/do-we-understand-soa.html

  • Anonymous
    May 31, 2007
    I just want to point out that I did NOT say that in the future everyone will be using REST for ALL services. What I did say is that in a relatively short period of time (2-3 years) the vendors will implement native support for REST in their development frameworks, and therefore REST will become more approachable to the average developer. Once that happens, we should see broader adoption of the REST architectural style. That does not mean that all other service architectural styles will go away.

  • Anonymous
    May 31, 2007
    Thanks for the clarification Anne.   I know you would never make such a sweeping statement (which is why I mentioned that this was said at a web conference).   I was worried about the old SOAP vs REST religious war flaring up again.  There are lots of things in this world worth being religious about - an application protocol isn't one of them.

  • Anonymous
    June 03, 2007
    If it's the most successful application protocol ever developed, I think that's something to get excited about.  It's eased the process of exchanging data between untrusted parties more than anything else that ever came before it, or has been developed since (though I suppose a case could be made for Bittorrent).

  • Anonymous
    June 04, 2007
    Hi Mark - I'm not surprised to see your comment here, especially regarding the topic. HTTP has been very successful but the concept of one size fits all (as alluded to in the original article's title) is folly.   REST has its place, as does WS-*.  The key is determining the proper place and time for each instead of simply weilding hammers and looking for nails.