Re: Current Downloadable Font Status....

On Mon, 25 Aug 1997, Gavin Nicol wrote:

> ><OBJECT PI-FONT="Bundesbahn" GLYPH-IDENTIFIER="47">Alt Gif goes here</OBJECT>
> >
> >For a font with only one glyph, GLYPH-IDENTIFIER could be left out.
> >Probably we would also need a size attribute, but I guess the right
> >way to provide it, as well as other presentation attributes, would
> >be via style.
> 
> Repulsive enough to not even be worth commenting on....

Gavin - Just a moment - What's the problem? There are quite a few
PI fonts (Bundesbahn is a well-known example) that contain many
symbols that are very useful on web pages or in other documents.

In essence, they are images, not characters. They are packaged
in fonts because this makes them available in collections and
provides excellent scaling, to a degree of quality that not
even line-based graphics (for which alone it may be worth to
have a font with a single character), and definitely not GIFs or
PNGs, can provide. They get used, currently with the <FONT> tag
that disguises them as text. I think we strongly agree that we
don't want the <FONT> tag, neither for this nor for anything else.

Putting them into the private area of the Unicode space doesn't
really solve the problem, because glyphs in the fonts currently
are in the rage 0-255, and so we need additional machinery to do
the mapping, which has to be coordinated, and so on. It's also
a philosophical problem, because these are really not characters
(well, there are enough similar things in Unicode to serve as
precedents, but that's really not a good argument).

What I want is a way to address images in a font which is
(mis)used as an image collection in the same way as other
images. The preferred way to do that is to use <OBJECT>.
Probably my syntax above, in an attempt to be explicit,
was too explicit. Of course, "PI-FONT" could most probably
be replaced by "SRC", and GLYPH-IDENTIFIER could become
COMPONENT-IDENTIFIER. And of course, the value for
COMPONENT-IDENTIFIER would allow names instead of numbers
(most font formats have glyph names for identifiers), and
these would strongly be preferred.

After having given more details on the background of my
proposal, I am looking forward to some more detailled and
less categorical feedback.

Regards,	Martin.

Received on Monday, 25 August 1997 15:04:40 UTC