We've been getting a fair amount of comment spam, and we've got some heuristics on the server that allow us to detect and reject much of it automatically, but we've still had some bots that have managed to leak through. At a request from some of the bloggers at pluralsight.com, I've implemented a simple CAPTCHA solution to help us distinguish humans from spambots.
I'm dogfooding my own solution here, so if you have a moment, try it out by posting a comment on this thread.
This is opt-in for bloggers at pluralsight.com; for most of us the server-side heuristics are blocking spam just fine, so we won't need to turn on this feature. So far only Craig and Aaron have requested it, so you'll see it on their blogs as well.
It'll be interesting to see if any bots are able to break through our very simple image generation technique (I used the routine posted on codeproject.com for the image generation). Thanks to BrainJar for publishing this simple algorithm that works nicely in .NET.
Implementing this in .TEXT was rather interesting given that it doesn't support either session state or viewstate :-)
Posted
Apr 26 2005, 08:38 AM
by
keith-brown