W3C home > Mailing lists > Public > w3c-wai-ig@w3.org > October to December 2014

Clarification on normative glossary definition of "Large scale (text)"

From: Patrick H. Lauke <redux@splintered.co.uk>
Date: Mon, 06 Oct 2014 23:57:46 +0100
Message-ID: <54331E6A.70205@splintered.co.uk>
To: w3c-wai-ig@w3.org
Currently, "Large scale (text)" is defined as a measure point sizes.

"at least 18 point or 14 point bold"

Note 3 does aknowledge that "The actual size of the character that a 
user sees is dependent both on the author-defined size and the user's 
display or user-agent settings" and suggests that "The point size should 
be obtained from the user agent, or calculated based on font metrics as 
the user agent does, when evaluating this success criterion".

However, to me this does not take into account the fact that "pt" as a 
measure is anchored (in practically all environments, with the exception 
of, I believe, KDE and Konqueror) on the the measure of one CSS pixel 
http://www.w3.org/TR/css3-values/#absolute-lengths, which means that 1pt 
in CSS does not necessarily match 1pt in the traditional physical 
measurement sense. Further, this does not seem to take into account 
differences in viewing distance, or the fact that browsers/OSs/devices 
may not actually be using a measure that's close to the ideal viewport's 
*reference pixel*. This can be even further complicated on mobile/tablet 
devices depending on whether or not a page has been given a meta 
viewport directive.

In short, a measure of "18 point" can mean almost anything (even before 
a user may have  changed browser settings or actively used some form of 
zoom/resizing) depending on browser, OS, device. Does the definition 
require some modification or expansion? Should it, for instance, 
explicitly state that it's talking about CSS pt measures and, referring 
to the CSS 3 values spec, point out that it assumes that 1 CSS px 
matches the ideal *reference pixel* and that the user is reading at the 
average assumed viewing distance for that particular device? (which 
would still leave some room for interpretation, and would make any 
attempt at calculating the *real* pt value of any text measurement quite 

Or should the definition just drop the use of point measurements 
altogether, as even if those were actually meaningful in all 
environments, they would need to be taken as only informative measures 
since they're dependent on font shape/metrics anyway?

p.s.: this rumination brought on by this thread on WebAIM 
Patrick H. Lauke

www.splintered.co.uk | https://github.com/patrickhlauke
http://flickr.com/photos/redux/ | http://redux.deviantart.com
twitter: @patrick_h_lauke | skype: patrick_h_lauke
Received on Monday, 6 October 2014 22:58:12 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 13 October 2015 16:21:53 UTC