[
Lists Home |
Date Index |
Thread Index
]
From: "Michael Kay" <michael.h.kay@ntlworld.com>
> > > There are plenty of areas in these specs that would really benefit
> > > from detailed review and feedback, but saying "I think I
> > could design
> > > a better language" is not helpful at this stage.
> >
> > Right. But how about "I don't want to have to deal with all
> > 40-odd XSDL
> > types in my implementation, please change the Basic
> > conformance level"?
>
> Specific comments such as "I think that an XQuery processor should throw
> an error if it encounters an attribute of type xs:NOTATION" are of
> course very welcome, if the rationale is given.
That seems to illustrate a problem fundamental to the committee design
process. In committee, members hash and rehash arguments seemingly endlessly
over every point in the specification, but by a political process that
depends on the personalities on the committee, their intransigence or lack
thereof, their personal and company agendas, and their knowledge or lack
thereof of the problem domain, eventually arrive at a compromise. New
members on the committee might tip the balance in another direction at any
level of detail, resulting in the discard or rethinking of the entire body
of work. It is difficult for committees to see such an event as progress. As
a consequence, committees develop a hard shell that repels challenges to
basic assumptions as a turtle's shell repels predators. In the end, nothing
is so precious to a committee as its hard-won consensus, as, in truth,
committee members often cannot even recall the chain of reasoning, politics
and personality that led them to a particular conclusion.
A reviewer of committee work, on the other hand, is often in the dark about
the consequences of committee decisions until the entire process has
unfolded. A "Holy shit!" comment at that stage cannot penetrate the
committee's shell, and the committee work product marches on of its own
momentum. The w3 process seems fundamentally flawed. That designers fall in
love with their own designs is a given, but the community at large should be
given an opportunity to weigh the merits of alternative designs before a
'standard' is cast in stone. Standards should be based on existing practice;
failing that, there should be ample opportunity for designs to compete.
Bob
|
