|
|
Posted by Alan J. Flavell on 09/26/46 11:35
On Fri, 23 Dec 2005, Gérard Talbot wrote:
> Jake wrote :
>
> > WEFT-generated fonts work just fine with IE users; very practical, no
> > problems.
>
> It can lead to problems as Alan F. made me realized that I had
> created a webfont from a font (nunacom font) which had no support
> for Unicode.
The key point is that they had built their custom font, in effect, as
a symbol font, in which normal ASCII (or Windows-1252) characters have
been swapped out in favour of the required custom characters. In an
HTML context, this technique is just as *bogus* in an embedded font as
it is in an installed font. My discussion of the general principles is
at http://ppewww.ph.gla.ac.uk/~flavell/charset/fontface-harmful.html
There is no need for this! (as you already know, but I'm writing for
the other readers). WEFT3 is quite capable of producing proper
unicode-based embedded fonts for this purpose.
> The thing is that such webfont may work for creating synthetic
> webfont in MSIE 5+ but then, the Mozilla/Opera/Firefox/Safari user
> who may have installed a truly Unicode font (like code2000) may not
> be able to read the document despite using a unicode-capable font:
> this is very wrong. This become illogical, inconsequent, incoherent.
Those are the practical consequences, indeed; as I say, that method is
fundamentally bogus, and ought not to be used at all. There might
have been some excuse in the distant past, when unicode support was
not widespread in browsers; but the continued promotion of these
techniques (as was found at some Canadian Syllabics sites, such as the
one you mention) is deplorable. And it leads to a legacy corpus of
bogus HTML content, which its owners may not be competent to convert
into "real"(tm) HTML.
> My recommendation to Microsoft: include a true, complete Unicode
> font in their next os release and have one which can be downloaded
> for free from windowsupdate.com so that none of this could happen
> again.
It's a reasonable request, but I think a more-versatile solution would
be to let IE recognise when a character is missing in the selected
font, and to have some algorithm for finding it in another font, thus
producing the desirable result that is seen in Mozilla-family
browsers. Ideally there should be an algorithm for choosing the
character from a font which is cosmetically compatible with the
selected font, of course, but, when push comes to shove, even a
visually incompatible version of a glyph is better than displaying a
missing-glyph indicator (empty box or whatever). In the final
analysis, this still doesn't help users who don't have *any* font
containing the required character; but in practical terms it would go
a long way to remedying this shortcoming in IE.
This isn't only an issue for Canadian Syllabics, of course:
mathematical operators is a topic that gets a regular airing on HTML
authoring groups (in English and in German, to my knowledge, and
probably elsewhere too). Then there's Persian (Farsi), Urdu, Yiddish,
and other languages which use a few language-specific glyphs which
aren't used in the corresponding mainstream writing system (Arabic,
Hebrew, etc.) and tend to be missing from general-purpose fonts.
But I should be starting on writing that web page, instead of posting
too much detail here ;-)
Festive greetings
Navigation:
[Reply to this message]
|