Re: Introduction to URIs (was RE: WebArch introduction, sort of)

On Wed, Jan 28, 2009 at 6:10 AM, Xiaoshu Wang <wangxiao@musc.edu> wrote:

> So, let's ask: what is the architecture of the Web? Is it something
> objective, like the physical laws that there is no way we cannot disobey? or
> subjective, like a design principles that we try to cook up so we can do
> more thing about it? I think it ought to be the latter.

I'm with you here. My feeling is the "web architecture" is best
thought of as a recommendation just like any other - an argument that
a particular solution solves a particular set of problems that need to
be solved, together with an explanation of why those problems are
important and a plea to use the recommended solution to solve them.
The AWWW document differs from other recommendations in that it's hard
for a reader to figure out up front just what those problems are or
why they're important; and as you say it is worded as if it is an
objective statement of the way the web actually works, rather than as
a goal in whose achievement the community is being enlisted. The whole
reason the "web architecture" effort exists is prescriptive or rather
persuasive - behavior in the public interest has to be nurtured, since
it's not always the easiest or most obvious way to act.

I can see why one would want to use objective language. The web itself
(the flawed implementation) follows AWWW design to a remarkable
extent. So when you say something good about "web architecture" you
can point to places where it's been followed and good things have
happened. But I know that personally I found the objective term and
tone very confusing on first encountering it, and have only gained
some understanding of the "web architecture" phenomenon since joining
the TAG. That it's so murky is unfortunate because the goals are
laudable and the strategies and tactics proposed to achieve them are
generally good. The objective tone is unfortunate exactly because it
alienates those who need to be persuaded - and those like you who want
to improve it.

For the "information resource" / "representation" / 200-restriction
question, which has been a free-for-all for too long, it might help to
focus if we (the TAG and its friends, I suppose) took an engineering
process approach, as you seem to suggest, and get goals, use cases,
requirements, installed base, design alternatives, etc. in line before
advancing a particular solution. As with any such effort, the first
step would be agreeing on a sort of charter to set the ground rules
for any discussion of the subject.

I wasn't around when AWWW and the httpRange-14 rule (which, you may
note, is very clearly prescriptive) were formulated, but they have the
appearance of committee compromises designed to settle issues that
everyone was tired of. In both cases I suspect it was very difficult
to get agreement and the imprecision reflects a consensus that further
consensus-building work would have low return on investment. If the
pain we're experiencing now (which is lack of standardization,
basically) becomes unbearable for someone, that person should become
an organizer. (I'm trying to chip away at the problem as a sort of
hobby, but at the rate I'm going it will be many years before there
are results...)

Best
Jonathan

Received on Wednesday, 28 January 2009 16:16:24 UTC