RE: FW: LC Comments: Web Method Feature

I'm not sure I appreciate the distinction between "layering" and
"inheritance" in this context.  When you "layer" TCP over IP, you
"inherit" internetworking.  When you "layer" HTTP over TCP, you
"inherit" streamed data and connection semantics and internetworking.

The practice of "dressing up" objects to look like something else
seems more like a short term fix than an architecture choice, and
my understanding is that this group is searching for durable
design choices, not quick fixes.

You seem to be saying that non-networked software is rife with the
decay of the short term fix, and that therefore networking software
must follow suit.  In a sense you are predicting the failure of
the architecture group. (!) Is that a correct reading?

Walden


> -----Original Message-----
> From: Gavin Thomas Nicol [mailto:gtn@rbii.com]
> Sent: Friday, July 05, 2002 5:07 PM
> To: 'xml-dist-app@w3.org'
> Subject: Re: FW: LC Comments: Web Method Feature
> 
> 
> 
> On Friday 05 July 2002 2:43 pm, Walden Mathews wrote:
> > In object oriented design, we are admonished to think real 
> hard about
> > whether X "really isa" Y before we rush in with 
> inheritance.  Doesn't
> > that same kind of design care warrant here?
> 
> This is the difference though: what Mike is talking about is 
> layering, not 
> inheritance. It is very common to use a Facade pattern to 
> dress something up 
> as something else. That, and the old adage that "any problem 
> in CS can be 
> solved with another layer of abstraction" leads to some 
> degree of probability 
> that HTTP will not remain top dog/pure application 
> protocol... not that 
> that's a good thing mind you...
>

Received on Monday, 8 July 2002 10:08:40 UTC