Re: code, samp, kbd, var

Lachlan Hunt wrote:
>> it is necessary to have real statistics rather than to  simply describe
>> them as "widely used". Did WHATWG actually analyse their usage in a 
>> large random sample of extant web pages, and if so, what percentage of 
>> pages contained 1, 2, 3 or all of them ?
> 
> I believe Hixie's most recent study was over 3 billion documents, 
> presumably randomly selected from Google's cache.  I'll ask him for the 
> stats on those elements next time I speak to him.

Pages = approx % of pages using that element at least once
Per page = for pages using that element, average uses per page

  Element   | Pages  | Per page
-----------+--------+----------
  <td>      | 90%    | 90
  <ul>      | 30%    | 6
  <cite>    | 0.5%   | 6
  <code>    | 0.4%   | 20
  <blink>   | 0.1%   | 2
  <dfn>     | 0.1%   | 10
  <var>     | 0.05%  | 10
  <kbd>     | 0.04%  | 10
  <samp>    | 0.02%  | 20
  <ruby>    | 0.01%  | 70

Sample size: several billion pages.
Sample date: September 2006.
All numbers rounded to one significant figure.

To put that in perspective, 0.05% of 1,000,000,000 (1 billion) is 
50,000,000.

-- 
Lachlan Hunt
http://lachy.id.au/

Received on Tuesday, 15 May 2007 05:23:28 UTC