- From: Chris Lilley <chris@w3.org>
- Date: Fri, 2 Sep 2016 18:11:47 -0300
- To: public-webfonts-wg@w3.org
On 2016-09-02 14:13, Roderick Sheeter wrote:
> For {encoder, decoder, format} it seems to me we can use the reference
> impl as implementation 1 and FontTools as implementation 2. Currently
> they wouldn't both pass all tests but we have a clear path to make it so.
>
> For User Agent, the reference impl can be implementation 1 but to call
> FontTools a User Agent seems a bit of a stretch.
>
> So where does UA#2 come from?
Yeah that is a troubling question. On the one hand, all browser
implementations use the same code (for Brotli and for WOFF2 decode).
They then use different code to actually display the fonts. So there is
only one independent implementation.
On the other hand, all the browsers have (or will soon have, assuming
iOS Webkit gets there soon) support, and it is interoperable :)
I think that this may well be a clear case for special pleading. Two
independent implementations is a good condition to have, but if *all the
browsers* support it then we should just declare victory.
All browsers use the independent JPEG group implementation for
JPEG/JFIF, as an example; same for libPNG.
--
Chris Lilley
@svgeesus
Technical Director, W3C Interaction Domain
Received on Friday, 2 September 2016 21:11:54 UTC