Should I make this open-source?
jaudu.net
jaudu.net
The idea behind a traditional captcha is to present a task that is easy for a human to solve (ha-ha) but hard for a computer.
As far as I can tell, your method does not rely on this principle. Which means that you use some machine-readable scheme that relies on obfuscation to provide this function.
While that method may work individually, it does stop working as soon as your method gains any kind of traction, because then spammer will target exactly your scheme and any advantage is null, void and gone.
Even if it would work only for a brief time it surely would make their job harder and the overall process much more slower.
But then again, this is just my opinion and I've convinced myself it to work - for that brief time only if nothing else.
The problem with spammers and crackers are not people like you, but rather the ones ho hammer sites using thousands of bots running behind proxies, which send raw POST data to the action url of the site's form. Even with CAPTCHA the bots are able to successfully send the POST data using OCR, which recognization rate varies from all around 10% to high as 80% depending on the CAPTCHA system used.
For this reason, even if your method would work it is lesser bad for the website owner.
Every solution has its weaknesses, but it is rather subjective which has the least of them.
Jhsto: you might want to include a noscript tag warning visitors about this. Of course it might be better to have this done server side as anything client side will eventually get worked around.