[
Lists Home |
Date Index |
Thread Index
]
On Tue, 26 Oct 2004 09:27:05 -0500, Bullard, Claude L (Len)
<len.bullard@intergraph.com> wrote:
> How much code
> do we want to obsolete so that applications that
> were nearly done are now codeBits
That's one question that needs to be answered (or at least guessed at)
before doing anything: how much short term pain would actually be
caused? My own guess is that most end user application builders have
avoided the crufty stuff, whether or not it is legal in the specs.
The people complaining at the Sells' conference are the poor suckers
trying to implement the specs because their customers want 'standards'
in the abstract but are not clamoring for the nasty bits of the actual
specs. The purpose of any refactoring would be to cut at the
inflection points beyond which a given feature causes more complexity
pain than empirical benefit in the real world. I grant that will be
hard to determine!
> Why
> not toss out this whole 'pointy' thing and get
> back to a clean one pass parse based on proper
> data definitions, white space, end of lines,
> and curlies (let's Do C!)?
Sooner or later someone is going to do just that. The question is
whether we want to do selective breeding to keep the specs in synch
with changing realities or wait for punctuated equilibrium to toss it
out and start over.
> There comes a point where the business execs and
> the data owners look back and say "good enough"
> and push back because the costs of reinnovation
> are restarts in too many places.
OK, automobiles, television, most home appliances, homebuilding
technology ... lots of things have been essentially "good enough" for
decades. I'm trying to think of a computer-related technology that
exhibits this (mainframes? COBOL?). In the technology industry,
who's not busy bein' reborn is busy becomin' a low margin commodity.
>
> Are there
> any non-XML geeks,
Maybe not, but if "Developers Hate XML", who will stand up for it when
its equilibrium gets punctuated?
|
