Re: Hosted source format?

To facilitate our discussion during the WG Zoom call, I think it would be
productive to make a clear distinction between the "stored" font format vs.
what is actually "exposed" to web font clients - either via @font-face
declarations or, in our new use case, as an IFT subset..

IIRC, Monotype license used to clearly differentiate the two cases (server
storage vs. web fonts formats), where licensed fonts could be stored on a
server in their original source formats, but the "exposed" font formats
(i.e., what would be allowed to be used with CSS @font-face declarations)
were limited to WOFF / WOFF2 - the license would prohibit transfering a
font to a web browser / client in TTF or OTF source format.

Not imposing restrictions on storing a TTF / OTF font on a server (e.g. for
the purpose of generating subsets of the font) makes perfect sense, for as
long as the access to the font is limited to the server and
other workstations licensed to use it. Yes, one could've stored fonts on a
web server only in WOFF2 format, and  would have to unpack a font every
time a subset needs to be generated, but IMO, unless one has limitations on
storage capacity and wants to minimize font file storage size, it doesn't
make much sense, and only adds processing time without any tangible benefit.

Vlad




On Mon, Jul 15, 2024 at 10:32 PM Chris Lilley <chris@w3.org> wrote:

> Hi John,
> On 2024-07-11 19:03, John Hudson wrote:
>
> At that time, he presumed the hosted source format for IFT would be an
> uncompressed TTF or OTF file, but I pointed out that license agreements
> might be worded in such a way that webfonts need to be *hosted* as WOFF2
> or WOFF. Garret thought it should be possible for the hosted source font to
> be pre-compressed as WOFF2 or WOFF, then decompressed, subsetted, and
> recompressed as WOFF2 for delivery. I wanted to check on the status of this
> query, and wasn’t able to find a definite answer in the working draft text.
>
> That used to be the case, in particular for the old range-request method.
>
> With the new draft, the incremental version of the font is a regular sfnt
> font with two extra tables holding the links to the patches. This means
> that WOFF compression will yield a bit-identical font; and WOFF2
> compression can be used, and will yield a functionally identical font,
> provided certain conditions are met when creating the incremental version:
>   https://www.w3.org/TR/IFT/#ift-and-woff2
>
> --
> Chris Lilley
> @svgeesus
> Technical Director @ W3C
> W3C Strategy Team, Core Web Design
> W3C Architecture & Technology Team, Core Web & Media
>
>