Three cheers. :-0 :-0 :-0
It seems fairly clear to me that HTML's beauty is as a publication
protocol; but the top-level source of a document should <em>NOT</em>
use it. I compose documents in terms of semantics (i.e. "top-level
heading") not "big font, carriage return at the end"), and I feel quite
strongly that not only can I do without control over the formatting of
my documents, I actively don't want it.
Cranfield has an house style for paper documents designed by experts,
but it's not used much because each author must lay a document out by
hand, or pay a secretary to do so. Ideally, authors should be able to
produce a high level description in their favourite authoring tool, and
then translate to whatever formats are needed: a paper (e.g.
PostScript) copy in the "house style"; HTML which browsers can display
according to the (human) reader's preferences (which may be quite
unexpected for, say, a partially sighted user); entries for a
bibliographic database, and so on.
An interesting point occurs to me. Voice simulation for the blind is
becoming more and more sophisicated. It can't been too long before
spoken hypertext is available. Anyone considered how HTML might cope
with that? I suggest that only semantic markup will be of much use....
Peter Lister [email protected]
Computer Centre,
Cranfield Institute of Technology, Voice: +44 234 754200 ext 2828
Cranfield, Bedfordshire MK43 0AL UK Fax: +44 234 750875