Mobile Web Standards 2011-05


This dataset was obtained as part of the FS 2011 Mobile Systems Seminar presentation about mobile web standards. It aims to answer the following questions:

  • Which are the most commonly used mobile web standards at present?
  • Do web servers of popular web sites deliver different pages whether the client is a mobile or a desktop device?
  • Do popular web sites follow current standards and best practices?

In short, the 1000 most popular web sites (according to Alexa at the time of research) were crawled using different user agents. All pages from each website (with a limit of 5) was downloaded. Data was gathered, then aggregated. The work was carried by Francesco Luminati, supervised by Fabio Hecht, during the first semester of 2011. For more information on methodology and conclusions, please refer to the technical report on Mobile Systems V (to be published around August 2011).

Conclusions regarding mobile web standard usage were the following:

  • The most popular desktop markup protocol is XHTML 1.0.
  • The most popular mobile markup protocol is XHTML Mobile 1.0.
  • Mobile web standards are relatively less popular than desktop standards, even when the HTTP client identifies itself with a user agent of a mobile device.
  • JPEG and GIF are the most popular image formats.
  • Most external objects referenced by web pages are images, but the few existing Flash objects are responsible for most bytes downloaded, together with the markup itself.

Conclusions regarding mobile web best practices were the following:

  • Some web sites serve special smaller web pages to mobile users.
  • Considering total size (markup + all objects), most pages need to download more than 20 kbytes, thus not following best practices for mobile devices.
  • The vast majority of pages contain tables, which is not within best practices, some containing even more than 20 tables.
  • A negligible amount of pages use frames.
  • Most pages contain a significant number of markup validation errors, which is over 100 for more than 40% of the pages.
  • About 400 pages (nearly 10%) do not declare a doctype.

Download: totalized dataset in CSV format (zip format, 3.2 kbytes) (ZIP, 3 KB), full dataset in MySQL format (zip format, 10.5 Mbytes) (ZIP, 10 MB)

Below are a few selected graphs illustrating the aggregated information.

Please cite this dataset as

Mobile Web Standards 2011-05 Dataset,
Francesco Luminati, Fabio Hecht,

Keywords: mobile web standards survey dataset, usage of HTML, XHTML, on iPhone, Android, and Nokia devices, popular image formats, PNG, GIF, JPG, JPEG, average page size including JavaScript, Flash, CSS, images, markup, usage of tables and frames, markup validation.