RE: Announcing new font compression project

On Monday, April 02, 2012 12:25 PM Chris Lilley wrote:

> JK> I think there would be merit in separating - both for discussion and
> JK> in implementation - the two logical stages that are being done here.
>=20
> I agree. 'Lossless' can have various meanings, depending on what one
> considers important.
>=20

I agree, this is something we need to discuss. The proposed compression is =
"lossless" in a perceptual sense as the functionality of a font that has be=
en compressed and decompressed is preserved 100%. However, due to data opti=
mization steps that are taken by the compressor to reduce data size, the bi=
nary input and output files aren't likely to be identical. In essence, the =
functionality of the compressor can be loosely compared to an optimizing co=
mpiler - the same source code compiled with different optimization levels (=
e.g. with or without DEBUG option, etc.) would produce executable binary fi=
les that are not binary-identical but have 100% guaranteed, identical funct=
ionality.

<snip/>

> WOFF 2.0 compression requires an optimised font as a starting point and
> is then lossless.
>=20

I am not sure we'd have to go this far and require optimized font as a star=
ting point. The optimization will come naturally as an added benefit, but I=
 agree it needs to be discussed and probably captured in the spec in suffic=
ient details.

> There might be merit in defining a 'visually lossless' concept but I
> suspect that might be hard to define rigorously.
>=20

Another alternative would be a 'functionally lossless' concept and it may b=
e easier to define.

Thank you,
Vlad

>=20
>=20
>=20
>=20
>=20
>=20
> --
>  Chris Lilley   Technical Director, Interaction Domain
>  W3C Graphics Activity Lead, Fonts Activity Lead  Co-Chair, W3C
> Hypertext CG  Member, CSS, WebFonts, SVG Working Groups
>=20