> What bothers me is when folks try to explain REST using excessive 
> hyperbole and exaggeration, and negating the value of other 
> approaches, as if acknowledging the validity of them somehow would 
> diminish the importance of REST.  Which isn't the case, of course.

I presume that you're referring to this comment I made earlier.

 > In contrast, since WS-* is without constraint, it can't be said to
 > exhibit any desirable property--whether done well or poorly.


I'll admit to a little trolling, but this is not hyperbole, it is not a 
matter of opinion, and it is not a blanket statement based on religious 
fervor.  The fact is that there is a paper out there -- you know the one 
-- that analyzes all prevalent means of distributed computing and teases 
out the properties they exhibit and the design factors that brought 
about those properties.  Then, emphasizing "constraint and understanding 
of the system context," applies those design factors (constraints) to a 
new architectural style such that the end result is a system that 
provably! exhibits the properties desired.  Those properties being 
separation of concerns, scalability, reliability, visibility, 
performance, simplicity, and evolvability.

Since no such effort was ever undertaken with WS-*, then, ipso facto, it 
cannot be said to exhibit any desirable property.  Technically, this is 
not quite true.  Since WS-* is fundamentally RPCs with angle brackets, 
it can be said to exhibit the property inherent in a client-server 
architecture, i.e., separation of concerns.

Note, the above does not mean that WS-* can't be made to work.  Like 
most any technology, it quite obviously can. And that means that it's 
delivering some amount of value to the people that use it.  But neither 
of those things makes the above statement any less true.

It may seem like I'm splitting hairs, but it's important, and it's why 
this debate goes on.  As Anne said, there are a significant number of 
people on this list and elsewhere who continue to misunderstand what 
REST is.  And these people have influence in the world at large.  I, at 
least, consider it a worthwhile goal to enrich the Web; to indeed have a 
Web of services.  But, as should be clear, WS-* is not of the Web, for 
the Web is REST and WS-* is not RESTful.  Therefore, it is the best 
interest of the Web that WS-* not be use to cleave off thousands of 
private networks.  Informing and influencing WS-* thought leaders is one 
way of keeping that from happening.

Maybe you'll concede me the Web.  "Sure," you might say, "if you have a 
public facing service/resource, use REST.  But enterprise issues are 
different. And, besides, my intranet is not part of the Web."  To this I 
say, fine.  Knock yourself out.  But remember, the properties that made 
the Web so successful can be leveraged to make your enterprise more 
successful too: you can make information more available, systems less 
brittle, applications more scalable.  Just as the browser-oriented Web 
completely changed how internal applications are developed, so too can a 
machine-oriented Web.

Reply via email to