Re: font-family: IPA

> [Original Message]
> From: Ian Hickson <ian@hixie.ch>
> To: Ernest Cline <ernestcline@mindspring.com>
> Cc: W3C CSS List <www-style@w3.org>
> Date: 5/9/2004 4:50:05 PM
> Subject: Re: font-family: IPA
>
>
> On Sun, 9 May 2004, Ernest Cline wrote:
> >
> > [no way to detect IPA fonts]

No way to detect using Panose.

> Not really much point adding it to CSS then.

I disagree because I believe that user agents should be able to
pick a reasonable default for this font and only those users who
know about IPA would be likely to seek to change that default.

> > It would probably be simpler to check the font to see if it contained
> > characters from the IPA Extensions block.
>
> Don't most "Unicode complete" fonts like Lucida Unicode and Arial Unicode
> contain glyphs in that block?

Yes, and they would be appropriate choices for an IPA keyword.
However, neither font can be assumed to be on a particular user's system.
So the only alternative would be to use instead of a keyword, a list of
fonts
and hope that one of them will do the trick.  However, if none of them were
on the user's system and yet there were fonts that did support the IPA,
you could end up with not just a weird mixture of fonts, but from the
standpoint
of the IPA, incorrect glyphs for certain letters.  There are many popular
default
fonts such as Times New Roman which have incorrect glyphs for one or more
characters (as far as the IPA is concerned that is.  For ordinary Latin
text, the
glyphs they use are within the range of acceptable variation.) 

> > As for font-variant instead of font-family, I don't think so.
> > smallcaps can be simulated if there isn't a smallcap version of the
> > font, but as you pointed out, there is no mechanism to know for certain
> > if the glyphs are appropriate for the IPA and even it there were not, no
> > algorithm to synthesize reasonable correct glyphs.
>
> I don't understand what synthesis has to do with it?

You can simulate smallcaps, italics, boldness, underline, etc. even if the
font doesn't contain the glyphs for that effect by starting with the base
roman font  and applying simple transformations to the glyphs.  You can't
do that with the proposed IPA keyword.  You would have to substitute
your own glyphs, essentially providing a built-in font that would have
absolutely no relationship to the specified one.  For rarely used characters
that might not matter much, but do that to LATIN SMALL LETTER A, and
I think you will manage to make most people P.O.-ed, and that is the letter
that would benefit most from this proposal.

The basic point is that there needs to be some mechanism (the exact one
doesn't matter too much, but I still believe that a new generic font-family
would be the most appropriate one) for an author to indicate that he wants
the content to be displayed using glyphs that conform to the IPA standard.

Received on Sunday, 9 May 2004 19:19:09 UTC