[
Lists Home |
Date Index |
Thread Index
]
I understand the example. I don't accept the
certainty of it.
I don't know who Jacobsen is, but regardless of
specification or audience, the author that requires
fidelity to the document-as-sent has to use a
format built for that, e.g, PDF. Legal documents
don't account for the judge's spectacles,
so the judge has legal secretaries, CEOs have
executive assistants (the egoBoo term for
secretary), and so on. In other words, the
webwonk can't require the user to be
in charge except insofar as to ensure
some means is available when they are. CSS or XSLT can be
one of those means until the rule is that
the document cannot be modified as a side
effect or direct effect of the means of
transmission.
That doesn't mean people
won't do as you say and build an unreadable
page, or use PDF when XML or HTML plus a
stylesheet would be more effective for a
greater number of people. Experience says
they will. What technology has to account
for is the exception to its own developers
worldview or experience. PDF is a means
to do something quite specific: FFF. It
can also be cheaper to produce but that is
a local effect and choice. I note that
PDF is zoomable, fonts are specifiable, and
nothing stops an author from building a document
that is WAI-compatible.
The meatspace world is and always will be at
least one order of magnitude more complicated
than any artifact of it (eg, the web).
Cherry rules made by designers with regards
to dominance in a communications relationship
will fail on occasion with certainty. Let
both buyer and seller beware. I read Nielsen's
pages on best design like I viewed Pat Paulsen's
run for the US Presidency:
satirical comedy with flashes of insight.
If one is too young and/or not an American, for Paulsen, see
The Smothers Brothers Comedy Hour, circa 1968.
len
-----Original Message-----
From: Dave Pawson [mailto:dpawson@nildram.co.uk]
At 12:30 10/07/2003 -0500, Bullard, Claude L (Len) wrote:
>My point here is that arguing for complete author
>control will push one away from XML anyway. User
>control pushes toward it. If I am sending XML
>to an XSLT-enabled receiver, my control is pretty
>much zero, so trust. If I send CSS inlined, I have
>more control but not perfect because it is not
>hard to edit a file or even XSLT it. If I send it with
>a reference to CSS, I am back to the same problem
>as with XSLT: trust. I can't be sure what is in
>the CSS with that name at the receiver. The only way
>I can send it without having to trust the receiver is to use a
>FFF (final fixed format).
\one of the reasons behind css was to allow the CASCADE.
WAI says the user disposes, the author proposes.
e.g. Len wants it 4pt, pale blue.
I can't read it, so my cascad overrides Lens and I
get 16point black. That's accessibility?
Bottom line is, I guess. The need to get content transitted
is even more important than the authors desire for it to be 'just so'?
If I can't read it, I can't say wow, look what a good job Lens made
of this content. I just dump it if I can't read it.
How many web sites have you come across where the authors
are just *so* desperate to cram as much in as possible that little
is legible? This on the assumption that it just *must* be on the front page?
Jacobson was pretty smart IMHO.
Yes/No?
|
