Posted by Alan J. Flavell on 12/22/05 23:18

On Thu, 22 Dec 2005, Jukka K. Korpela wrote:

> Advocates of WEFT, when asked to present real-life
> examples, show nothing or show dummy demos - and babble endlessly.

That had been pretty much my conclusion, until Gérard Talbot showed me
how it could be used to support Canadian Syllabics in MSIE. I'm
hoping to write it up at some point, but, as I said before:

Meantime, interested readers could consult the discussion which
includes this posting

http://groups.google.co.uk/group/comp.infosystems.www.authoring.site-design/msg/2f48e144a18a19ea

Users of web-compatible browsers are still free to install and use
appropriate fonts - they are neither helped nor hindered by this
IE-specific option. I'm not *madly* keen in investing this extra work
specifically for IE users, but it sure beats the fake Latin-1 fonts
which some native-Canadian communities were offering as a means to
display their subset of syllabics.

This would also be good for other writing systems which IE tends to
botch[1], such as Persian, Urdu, polytonic Greek, and so on (review
the ones fingered by Andreas Prilop from time to time for examples).

[1] for the reason discussed at
http://ppewww.ph.gla.ac.uk/~flavell/charset/browsers-fonts.html#msie

Basically, MSIE decides which font to use for each group of characters
(e.g Arabic), based on what the font says it supports. If one or more
of the needed characters happens to be missing from the chosen font,
it makes no effort to locate it elsewhere. Persian and Urdu have a
few characters which aren't in the mainstream Arabic "alphabet" and so
may be missing, but IE doesn't care: if the font says it supports
Arabic, then it uses it, basta. Anyone interested in the details can
refer to the web page. Doubtless the usual suspects will show up to
tell us that the problem will be solved if we only nominate the
appropriate named fonts in the CSS, so I'd better say right now that
they're wrong, and I won't bother to contradict them one by one when
they show up.

Thus, MSIE is the commonly-used browser which, to my way of thinking,
is *most* in need of being helped out by the WEFT technique. If you
think MSIE ought to show up its own shortcomings, then I'd be only too
happy for you to pay no attention to this idea; but, if you care more
that all your readers - even those who are so misguided as to use MSIE
as a web browser - get something readable, then I grudgingly present
this option for your consideration.

cheers

[Back to original message]