Re: WOFF2 file extension and media type

Thank you Chris and Vlad.

For Google Fonts, we are serving the font files with a .woff2 extension and
with a mime type of "font/woff2". So a big +1 for recommending the ".woff2"
extension in the specification.

As part of our early testing/verification ~50% of the Google Fonts
collection are currently being served as WOFF 2.0 for Chrome 36 (which is
in Beta).  Things are looking good so far.

This email is a great reminder that I need to double down on capturing the
benefits of a new top-level font/* mime type.  It would be great to start
this process in parallel...


On Wed, Jun 25, 2014 at 7:44 AM, Levantovsky, Vladimir <
Vladimir.Levantovsky@monotype.com> wrote:

> Thank you Chris!
>
> All, this is a straw poll for ".woff2" file extension - any objections or
> concerns / limitations of usability?
>
> Thank you,
> Vlad
>
>
> -----Original Message-----
> From: Chris Lilley [mailto:chris@w3.org]
> Sent: Wednesday, June 25, 2014 10:12 AM
> To: Levantovsky, Vladimir
> Cc: w3c-webfonts-wg (public-webfonts-wg@w3.org)
> Subject: Re: WOFF2 file extension and media type
>
> Hello Vladimir,
>
> Wednesday, June 25, 2014, 4:03:43 PM, you wrote:
>
>
> > I've been contacted privately by a developer who asked me about two
> > things - what would be the appropriate font file extension for WOFF2
> > and what is the media type to be used when the fonts in WOFF2 format are
> served.
>
> I agree that we should define this.
>
> > We haven't yet discussed the file extension at all - the assumed file
> > extension was .woff2 although I am not sure if having
> > five-character-long file extension is okay in all circumstances.
>
> The main restriction on file extensions was those older operating systems
> which required an extension to be three or less characters.
> (MS-DOS and older versions of Windows (1,2,3,3.1,95,98,ME).
>
> This was a problem in the early days, .html thus became .htm on DOS and
> non-NT versions of Windows. Since Windows 2000, the three letter extension
> is not a factor any more.
>
> >  As
> > an alternative, I suggest to consider .wof2  but this is merely a
> > proposal -
>
> .woff is four letters and has not been a problem, so my suggestion is to
> formalize the assumed .woff2 and it is simple and easy.
>
> > I'd rather have us to decide on this and announce it publicly as soon
> > as possible - the folks are waiting and the guy who contacted me with
> > these questions is not alone, I am sure.
>
> Agreed.
>
> > As to the media type - we did discuss it in the past and while we all
> > seem to agree that having the top-level 'font' type would be ideal -
> > we are not there yet. Applying for the top-level type would be a huge
> > undertaking and something that  may not bring the results soon enough
> > for developers who need it now.
>
> Yes.
>
> >  I am wondering if we
> > should (at least in the near term) consider something similar to what
> > we did for WOFF 1.0 and register a media type as part of the
> > "application" sub-tree, e.g. "application/font-woff2"  Comments?
>
> I agree that we should do this, and doing so does not preclude any later
> change if font/* goes ahead. Developers need this now.
>
> I will take an action to draft a MIME registration appendix, similar to
> the one for WOFF 1.0.
>
>
>
>
> --
> Best regards,
>  Chris                            mailto:chris@w3.org
>
>