Re: Size increase for streamable font pre-processing

I can gather data on how much the TTF/OTF's increase in size as a result of
the processing.

For the simulation I did two separate runs:

   - The first was with unprocessed fonts, simulating everything but range
   request.
   - The second was with processed fonts (via Myle's tool), simulating only
   range request.

Re: hinting that's correct, the optimizer tool currently drops hints. So to
make things fair I also dropped hints from the fonts used in simulating all
of the non-range request methods.

The data set zip I sent out contains two copies of the font library "fonts"
(hints dropped) and "fonts.optimized" (hints dropped and run through the
optimizer tool). These are the exact fonts I used in the simulation runs.

On Wed, Oct 7, 2020 at 9:34 AM Chris Lilley <chris@w3.org> wrote:

> For fonts processed by the Streamable Fonts tool
> https://github.com/litherum/StreamableFonts
>
> is there data on how much larger they get as a result of
> desubroutinisation and other processing?
>
> Ideally a range with percentiles, rather than just an average.
>
> It isn't clear to me (sorry) if, in the simulation, the fonts were first
> processed this way and then served by the two approaches (so they
> started from the same baseline) or whether this was only applied to the
> fonts served with Glyph Byterange, and the ones served with Patch and
> Subset were the original fonts (thus, a bt smaller to begin with).
>
> Also, I recall reading that the fonts served had hinting removed, is
> that true? I can't find where I read it now, sorry.
>
> --
> Chris Lilley
> @svgeesus
> Technical Director @ W3C
> W3C Strategy Team, Core Web Design
> W3C Architecture & Technology Team, Core Web & Media
>
>
>