13 February 2007

Web browsers and language

I'm continually amazed at how forgiving web browsers are in allowing access to malformed information. HTML can contain unclosed elements or a mix of valid and invalid elements and attributes yet browsers will still display the page for the most part correct. This puts web browsers further from programming language interpreters and closer to natural language interpreters. For programming languages, as soon as any syntactic error is encountered, the program and remaining commands are abandoned. An interpreter can't degrade gracefully from division-by-zero or adding two numbers using the smiley-face character. In contrast, natural languages and its users are adept at eliding semantic and syntactic gaffes without completely abandoning the message. I can "don't want no ketchup" or ask "how many bottle should I buy?" and still be understood. For the most part.

[ posted by sstrader on 13 February 2007 at 6:17:21 PM in Language & Literature ]