RE: Myth of loose coupling

Mike et al:

It would certainly help to elevate the debate at the metamodel level and
establish a mapping to technological concepts such as GET, URI and POST.
A picture being worth more than a thousand words on this list, I would
suggest to illustrate it with UML class diagrams.

Jean-Jacques 
 
 

>>-----Original Message-----
>>From: www-ws-arch-request@w3.org [mailto:www-ws-arch-request@w3.org]
On
>>Behalf Of Champion, Mike
>>Sent: Wednesday, January 08, 2003 9:57 AM
>>To: www-ws-arch@w3.org
>>Subject: RE: Myth of loose coupling
>>
>>
>>
>>
>>> -----Original Message-----
>>> From: Mark Baker [mailto:distobj@acm.org]
>>> Sent: Wednesday, January 08, 2003 9:39 AM
>>> To: Walden Mathews
>>> Cc: www-ws-arch@w3.org
>>> Subject: Re: Myth of loose coupling
>>>
>>>
>>>
>>> +1.  I was going to respond to Mike, but this says more or less what
I
>>> wanted to say.
>>>
>>> Though I detest the expression, there really is a "paradigm shift"
to
>>> understanding why GET+URI is superior to POST+method.  For most
people
>>> coming from a CORBA/DCOM-like background (like myself), it takes a
>>> "eureka moment", if you know what I mean.
>>
>>Well, I considered myself a RESTifarian until I realized that y'all
wanted
>>me accept the catechism that included the "uniform interface
constraint"
>>and
>>the categorical superiority of URIs over XML :-)   Again, these are
tools
>>that engineers can use to optimize solutions that are adapted to
specific
>>problems, IMHO.  I much appreciate the work that RESTifarians have
done to
>>educate people on the limitations of the RPC paradigm over the
Internet
>>and
>>on the desiriability of fully leveraging the Web rather than just
>>tunneling
>>a bunch of stuff through port 80.  I see that more of a kick in the
shins
>>than a religious epiphany, however.

Received on Wednesday, 8 January 2003 10:13:05 UTC