Still no HTML

Many years ago, there were some standards set out for the language of the web: "HTML".

The idea is simple: if you define the language then everyone who writes it and everyone who reads it will understand each other. 

There were several good tools (mostly released free of charge) written to at least check that the HTML that you write is at least syntactically correct. That is, in non-technical language, when computers talk to each other, the computers use the same words and sentence structures so that they can understand each other.

So, why then in 2010 do big banks, with their massive budgets still skip the simplest of requirements+tests on their on-line banking and not stick to the standards, littering their HTML with obvious errors that are apparent on any browser (except perhaps the one that the web programmer happened to use)?