HTML Statistics

10 March 2006 Code

Everyone loves statistics, especially us web designers. From checking your traffic stats 5 times a day to Technorati this or that, everyone likes to compare stats.

So a couple of okes did a web study on sites that validated, which type of HTML tags were most common or how many different ID tags were used. These studies were based on less than 1000 websites or pages. So the results are more interesting than conclusive. December 2005 roles around and in steps Big Daddy and did “an analysis of a sample of slightly over a billion documents”. Woah…

Some interesting facts

Most pages have 19 elements specified. <html>, <head>, <a>, <table> etc
The top 4 elements used are <head>, <html>, <title> and <body> - funny thing is that 3 of the 4 elements are optional in HTML.
The most popular class names used are footer, menu, title, small, text and nav - and all these class names have all been proposed in HTML5.
Interesting that “msonormal” was the 13th most used class name which is a Microsoft Office generated tag. Just shows you how many people still use that stone-age way of coding sites.
More pages use the completely worthless <meta name=”revisit-after”> than use the <em> element!
The <body> element is probably the most abused element. Out of the top 20 attributes (bgcolor=”", topmargin=”", link=”" etc) 9 are completely invalid, and 5 have been deprecated for 8 years, which is half the lifetime of the web!
Tables. There was a scary amount of table tags without any td tags. Empty tables?!
The <br> element is used more than the <p> element.
Some of the <a> elements rel=”" attributes values have mostly been spurred on of late by Wordpress (rel=”external”) and the Web 2.0 social tagging phenomenon (rel=”tag”).
The <script> element has some funny attributes associated with it. The language=”" attribute is the most used, but has been deprecated for years. Funny thing is that people also can’t spell it - langauge=”", langugage=”" and languaje=”".

Go read some of the stats for yourself at http://code.google.com/webstats/index.html. Darn interesting.

Please bookmark this post: