In the Wikimedia Foundation we use captchas. They are a tool against spam bots and vandals. They do their job and they can be annoying as hell when you do not get it right the first time. As they serve an important purpose I am happy that we use them anyway.
The BBC news webiste reported on a captcha scheme that is different. The idea is so compelling, that I want to seriously argue for the adoption of this scheme in the WMF
optical character recognition software. As we have projects in so many languages, we can provide an even better service, our people can do this same work for other languages then English. I am sure when our Wikipedias become part of such an effort, it would help the digitisation effort for many languages..
The good news is that according to the Recaptcha website, there is already a "plugin" for MediaWiki. So for me it is clear why we should join this great endeavour. It makes filling out captchas useful in more ways then one :) .. PS why not have in the user preferences a place where you can recognise words.. just because you can ??