INTRODUCTION

This file contains various miscellaneous pieces of information gleaned
by Google in July 2007, from a study using a parser based on r967 of
the HTML5 specification.


NUMBER OF ATTRIBUTES ON ELEMENTS

The following table shows the percentage of elements with a number of
attributes or less:

  elements with <= 0 attributes:  33.5%
  elements with <= 1 attributes:  66.1%
  elements with <= 2 attributes:  81.6%
  elements with <= 3 attributes:  90.2%
  elements with <= 4 attributes:  95.3%
  elements with <= 5 attributes:  98.3%
  elements with <= 6 attributes:  99.4%
  elements with <= 7 attributes:  99.8%
  elements with <= 8 attributes:  99.9%
  elements with <= 9 attributes: 100.0%

The following table shows the percentage of pages whose element with
the most attributes had a given number of attributes or less:

  max is <=   0 attributes:  0.3%
  max is <=   1 attributes:  0.8%
  max is <=   2 attributes:  3.1%
  max is <=   3 attributes:  6.8%
  max is <=   4 attributes: 12.8%
  max is <=   5 attributes: 27.9%
  max is <=   6 attributes: 52.1%
  max is <=   7 attributes: 70.5%
  max is <=   8 attributes: 83.8%
  max is <=   9 attributes: 90.5%
  max is <=  10 attributes: 95.9%
  max is <=  11 attributes: 97.8%
  max is <=  12 attributes: 99.0%
  max is <=  13 attributes: 99.5%
  max is <=  14 attributes: 99.6%
  max is <=  16 attributes: 99.7%
  max is <=  20 attributes: 99.8%
  max is <=  26 attributes: 99.9%
  ...
  max is <=  95 attributes: 99.99%
  ...
  max is <= 340 attributes: 99.998%
  ...
  max is <= 627 attributes: 99.9990%

Note that many pages have an element with an extremely large number of
attributes. In particular, many pages include a <meta name="keywords">
element where they have misquoted the "content" attribute, and so each
keyword ends up generating its own attribute. Such pages often have
hundreds or thousands of keywords.


METHODOLOGY

The sample data consisted of several billion documents from Google's
index. An attempt was made to eliminate duplicates from the
sample. Documents that did not appear, a priori, to be HTML documents
(e.g. PDF documents, Flash animations, Word documents) were not
included in the sample. Documents which, when parsed, resulted in DOMs
with more than 65535 elements were only parsed up to the creation of
the 65535th element.

It is hoped that this data will help implementors of HTML5 parsers to
optimise their branches to better serve actual Web content.

For more information, please contact Ian Hickson <ian@hixie.ch>.


LICENSE

Copyright 2007 Google, Inc.

Licensed under the Apache License, Version 2.0 (the "License"); you
may not use this file except in compliance with the License. You may
obtain a copy of the License at

   http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
implied. See the License for the specific language governing
permissions and limitations under the License.