Google Image Search
Image by schoschie
While working on something related, I un-minified and cleaned up the source code of an example Google image search, and man, that is some ugly shit. Obviously, it works, but it’s really ugly.
The HTML is HTML 4 with markup errors and font tags, and the actual search results are built up completely using
Since Google usually know what they are doing, there must be a reason for this, and I’m not quite sure what it is. Or maybe they wrote it in 2000 and someone said »if it ain’t broke, don’t fix it«, and it stuck until today.
(* I recently took another look at some of the IOCCC entries which gave me a reality check as to the extremes of obfuscation – www.ioccc.org)
Update 20090615: After having studied some more Google web page sources and reading the other article, the reason is performance and efficiency. Google doesn’t care much about standards as long as the content is accessible in any browser, and every byte they can save counts. To them, 10 more bytes in a search result probably means 1,000 new servers.