W3C home > Mailing lists > Public > public-low-vision-a11y-tf@w3.org > April 2017

Experiment Design

From: Wayne Dick <wayneedick@gmail.com>
Date: Sat, 29 Apr 2017 13:05:20 -0700
Message-ID: <CAJeQ8SAOKcmk6+Okj2QANEmN=ejUVEUsBboixLd01fA5esDuDg@mail.gmail.com>
To: public-low-vision-a11y-tf <public-low-vision-a11y-tf@w3.org>
Last night I downloaded the JSON file that gives load instructions for all
800+ Google Web Fonts.  My program can compute the width a font takes with
great accuracy. I propose the following experiment.

Compute the font "Typewriter Set" width and average character size for
every Google Web Font and 16px. For Chrome and Firefox. They are different.

Compute a mean for both distributions.

Compute a new distributions of Typewriter Set width divided by the
distribution mean. Surprisingly this number is very stable across Crome and
Firefox to one digit past the decimal place. It also gives a revealing
statistic: A distance from the mean. For example, Verdana will probably be
between 1.2 and 1.3.

That should give us a range of font widths that can be safely substituted.
We can even choose our range so that at least one font lies on the upper

What do you think? It should take one week to program roughly.

I added a normalize computation to yesterdays app. I know it is rough but
it works. I do not have a mean now. Instead I normalize around the pivot
width(Arial). When the time comes we can pick a pivot more intelligently.

Received on Saturday, 29 April 2017 20:06:34 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:23:25 UTC