Re: A way forward

On Thu, Jul 23, 2009 at 6:02 PM, John Daggett<jdaggett@mozilla.com> wrote:
> Right now linking to TTF/OTF fonts represents what will soon be an
> interoperable solution for all browsers other than Internet Explorer.
> Some authors may consider this to be enough, using @font-face only as
> a progressive enhancement for their sites.  For a more interoperable
> solution, authors can also choose to serve EOT versions of fonts to
> Internet Explorer users.

And doing so requires encoding a font twice and linking it twice,
which defeats the whole purpose of interop.

> As I understand it EOT-Lite boils down to a header prepended to the
> front of a font, with no MTX data compression and a null root string.
> I'm assuming the silly XOR'ing of the data (TTEMBED_XORENCRYPTDATA)
> has been omitted.  None of the remaining data in that header seems
> like it's useful, the data is either already in the font or it's
> defined in the @font-face rule.  You might as well just prepend a null
> four bytes to the font data, that would have the equivalent level of
> protection, you wouldn't be able to use the font file as a desktop
> font.
>
> If Microsoft can ship an update to support CFF fonts in an EOT format
> in older browsers on older operating systems they certainly could ship
> an update to support a simple format like this, I don't see a why
> other browser vendors should bend over backwards because Internet
> Explorer has long product cycles.

By your own admission, supporting EOT-Lite wouldn't be "bending over
backwards" - it's just a bit of mucking about in the header.

The reason you would slightly cater to Microsoft in this case is
because it helps authors.

> Either the .webfont format or Jonathan Kew's ZOT format seem fine to
> me, but I think Mozilla would only support an additional format that
> other browser vendors were also willing to support, including
> Microsoft.  And I don't see any other browser vendor eager to support
> any variant of EOT (with or without the spicy mustard) other than
> Microsoft.

Not through lack of trying.  Kew posted a question back on July 2nd
that was essentially about EOT without rootstrings but with MTX and
CORS (at least, the generic format he described would include what I
just said).  He posed it to both browser developers and font
foundries.  Not a single browser developer responded.

It's unfortunate that when these sorts of direct questions are posed,
we get silence.  ;_;  Having a firm array of everyone's positions on
the various proposals would likely make this whole thing *much*
easier.

~TJ

Received on Thursday, 23 July 2009 23:24:47 UTC