Re: zwj hack in MediaWiki (Was: Consonant allophones in Minza)

Henrik Theiling wrote:
> Hi!
>
> Benct Philip Jonsson writes:
>> ...
>>> So I would recommend to *not* use a ‍ to hack-fix a broken
>>> behaviour of a browser or font, but instead to use a working browser
>>> or font, or if that is out of the question (for whatever reason), live
>>> with a distorted display, and/or inform the maintainers of the browser
>>> or font of not being Unicode compliant.
>> I *dis*agree.  I want as correct display as possible *now*, and I
>> won't sit around until someone fixes the problem if I
>> can fake it now.  I think think that computers should adapt to humans
>> and not the other way around.
>
> Well, nothing else is what I am proposing.  But a little bit of
> patience for the humans who have programmed them would seem good, I
> think.  You can speed up by *informing* the right maintainers.
One problem is that the underlying font technology in one of the most
widely used operating systems (Uniscribe in Windows XP) doesn't even
support essential features for diacritic placement in the Latin script
unless you have just the right version of the DLL (which you can only
get by joining the VOLT discussion group unless they've changed this).
So ... either your readers need to download just the right version of
Uniscribe AND an appropriate font, or you can do things another way
which works more easily out of the box (even if it bends the rules a
bit). You can't blame the browser or the font for more fundamental
limitations of the system.

Another alternative would be a browser that supports Graphite or some
other font rendering system, if there is such a thing for Windows
systems. But again, you'd be asking everyone who reads the site to
download and install a new browser just to read the page; not everyone
is going to have that level of interest.

Windows Vista seems to have better support for Unicode, and even their
built-in fonts have some good diacritic support, but asking readers to
"upgrade" to Vista is the last thing I'd want. Still, I've done some
experimentation with OpenType tables, and if I ever get around to it
(not likely any time soon), I was planning on putting some basic support
for diacritics (e.g. removing dots over lowercase i's and j's) in the
next version of the Thryomanes font.