Re: Arch Doc: 1 August 2003 Editor's Draft

> Hmm... the core terms from the introduction have been
> substantially changed...
>
> |The World Wide Web (WWW, or simply Web) is an information system
> |that relates information sources and services, referred to
> |collectively as resources ...
>
> That defines resources as information sources and services
> and nothing else; that's a substantive change from the earlier text,
> which I thought was quite mature:
>
> |Objects in the networked information system called resources
> |are identified by Uniform Resource Identifiers (URIs).
>   -- http://www.w3.org/2001/tag/2003/webarch-20030716
>
> it placed no constraints on what a resource is.

Actually, it places the same constraints -- they just get missed by
the noise.  All I did was rewrite the sentence to reflect what it was
actually saying.  In my message to the TAG, I said it was wrong because
it defined the Web as an information system rather than an information
space.  The changes that Ian applied do not change that, nor will
removing them improve it.

> Where did this change come from? The changelog says...
>
> 1. Introduction: Incorporated text suggested by Roy Fielding."
>   -- http://www.w3.org/2001/tag/webarch/changes#changes-20030801
>
> i.e. this change came from work-in-progress from one tag member,
> sent only to the tag private list, with no endorsement from
> anybody else, that I can see.
>
> That seems like a big step backward from the level
> of maturity section 1 had as of the Vancouver meeting.

That portrayal is incorrect.  We didn't discuss it in Vancouver
because we had already discussed it in Irvine, at which point both
myself and TimBL said that the section needed to be rewritten.
It is still in a very immature state.

> Hmm... I'm disappointed to see that the record of the Vancouver
> meeting doesn't record
>   -- my request that the review of each section
>   not assume that everybody agreed to everything unless
>   they said otherwise; that we solicit endorsements
>   as well as criticisms
>
>   -- my endorsement (and Tim Bray's, if not lots
>   of others') of the text of the intro section,
>   and the definitions of the core terms in particular.

Was that before I arrived?  I have no recollection of it.  In any case,
I introduced my changes during that meeting.

> So as a reviewer like any other reviewr, I ask:
> please undo this change and put back the intro text
> that was there before.
>
> And I ask the chairs to work with the editor to reduce
> churn. Let's not have half a suggestion by one TAG member
> turn into substantive changes to text that has been
> endorsed previously by multiple TAG members. Let's please
> be especially careful with the definitions of the
> marked-up terms.

Bah!  In that case, let's also be careful to waste our time with
procedural details whenever Dan disagrees with a change, rather
than actually attempt to fix some of the utter nonsense in the
webarch document that makes it nearly impossible to review.

I rescind my abstention and vote against this draft being
published as a TAG draft until the abstract and introduction
is rewritten to be at least halfway sensible by avoiding the
creation of new terms that nobody outside the W3C use to
describe software architecture, ceasing to redefine existing terms
such that the WWW information system and Semantic Web cannot be
distinguished from the Web of relationships between resources,
and including human behavior as being within the scope of the
system that is being described.  And while you are at it, please
feel free to improve the grammar and prose.

Below are some relevant suggestions that were posted before in
private mail.  The problem with defining the Web as an information
system rather than an information space is that people start
making global assumptions about traits that are only applicable
to one system that uses the Web of resources.  Thus, I don't like
the definition of Web given in the current document and merely
rephrased here -- it needs to be changed to refer to the Web as
the set of explicitly related resources and some new term has to
be created to refer to the WWW information system.  That will allow
the differences between the WWW, SW, and WS architectures to be
properly explained in the document, rather than continuing pointless
arguments about how WWW constraints might impact the SW.

Note that most of these changes are simply placing quotes around
URIs that appear in text so that they are properly recognized and
delimited from the surrounding punctuation.  Feel free to replace
them with angle brackets, as recommended by all URI specifications
since 1994, but quotes are easier when dealing with XML transforms.

....Roy