On Thu, 26 Sep 1996, John Kunze wrote:
> First, apologies from the user guide group for nt responding to the
> various syntax proposals (eg, Type vs Scheme, and placement of
> qualifiers in the META tag's content attribute). We'd been pursuing a
> different line of inquiry for some time and needed to see it through a
> bit further before pausing in our (need I say?) very busy lives to give
> the other proposals the responses they deserved.
That's a pity because the rest of us (also with busy lives) have been
burning midnight oils writing code and documents for what we thought was a
good solid consensus on the mailing list. Maybe I'm just too "IETFy" but
I'm used to working groups where rough consensus is drawn on the mailing
list and you're expected to speak up if something's niggling you or you've
got different ideas as to how things work. Even a couple of lines just
saying, "hey we've got a few different ideas here" with a rough outline of
the proposals would be cool. Suddenly having a completely new version of
Dublin Core appear out of nowhere, even if only in draft form, is a little
disconcerting to say the least.
> The User Guide is not a proper "user guide" yet. For one thing it's a
> rough draft. For another, it has a couple extraneous working sections
> temporarily thrown in simply because the guide is a convenient vehicle
> for supporting arguments (eg, describing the question for which
> metadata is the answer) and exploratory contextual scenarios (eg,
> Warwick framework).
Then maybe its better to rename the current document as a position paper
or discussion document?
> BTW, as mentioned before, a very short, quick-reference-style
> version of the guide is envisioned.
That's what I was expecting.
> In the guide, the
> convention of prepending "dc." to element names was taken straight out
> of the agreement reached at the Boston W3C searching and indexing
> meeting.
The use of "dc" as a schema identifier for META tags in HTML documents is
fine; I don't think anyone would have a problem with that.
> The guide's use of META for encoding metadata doesn't break
> any browsers (that I know of).
It does break any validation services that might be used to ensure
"future-proofed" HTML is generated and any browsers that are based on
strict SGML parsers (which are currently few and far between but might
not be in a couple of years time when there lots of different DTDs in use
on the web). Its also bad practise IMHO for something that I think many
of us would like to see become a defacto standard for metadata to go round
breaking other standards needlessly.
> It also says that the included notation
> is one of several, so an SGML notation is certainly not precluded.
>
> I believe Terry was concerned about the use of META outside a document
> that it was describing; apparently there's a rule that META cannot
> describe another document not included in the same HTML file (is that
> right?) -- the way out is to say that the file containing the META tag
> is not an HTML file (it just bears a remarkable resemblance to one).
Hmm, that's a bit icky for a recommendation in a formal user guide though.
Kind of smacks of a kludge. I'd much rather see reference to a proper
SGML DTD that can be used for archival purposes, etc, etc that isn't tied
to the wobbly path that HTML is following.
Tatty bye,
Jim'll
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Jon "Jim'll" Knight, Researcher, Sysop and General Dogsbody, Dept. Computer
Studies, Loughborough University of Technology, Leics., ENGLAND. LE11 3TU.
* I've found I now dream in Perl. More worryingly, I enjoy those dreams. *
|