W3C home > Mailing lists > Public > www-style@w3.org > May 2003

Re: Values in font-size-adjust example

From: todd fahrner <fahrner@pobox.com>
Date: Wed, 14 May 2003 13:06:20 -0700
To: www-style@w3.org
Message-Id: <87457546-8647-11D7-BE7D-003065BD01D0@pobox.com>

On Monday, May 12, 2003, at 19:21 US/Pacific, L. David Baron wrote:

> On Tuesday 2003-05-13 04:01 +0200, Christoph Päper wrote:
>> Verdana:         0.58 -- 1117 / 2048 ~= 0.545
>
>> Times New Roman: 0.46 --  916 / 2048 ~= 0.447
>
>> Are the values wrong, inaccurate, derived from different versions, or 
>> is the
>> calculation method flawed?
>
> I'm not sure.  I did similar computations back in February 1999, and 
> got
> similar numbers for the two fonts above based on measuring the 
> lowercase
> x (see screenshot [1]).  I emailed Todd Fahrner, who I think was the
> original author of the table in question, and he replied, with a link 
> to
> a screenshot that no longer exists, which contained what I suspect was
> an answer to my question that satisfied me at the time (probably with
> the font's metrics drawn in showing where the "official" x-height was).

I measured the x-heights of the sample fonts at 100 pixels per em. I 
suspect strongly that the hinting for that rasterization causes the 
values to vary from the values you derived at 2048 ppem. Note also that 
x-height is an inexact notion given that (good) letterforms typically 
depart from simple metric consistency for optical finesse. Carrying 
these figures out to 3 places seems overprecise, especially for coarse 
screen display - the medium for which the property is most badly needed 
in my opinion.

Note also that the property has been removed from CSS2.1 in view of the 
fact that there are no (known) implementations. It seems that a 
decisive majority of authors are willing to assume the presence of 
Microsoft's (formerly free) TTF fonts client side.
Received on Wednesday, 14 May 2003 16:06:29 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:54:21 GMT