Comment spam defeated at last
Finally, a few months ago, I found a way that worked 100% of the time. This raw text file shows what I'm up against, containing all server variables and full text of every comment I've gotten in the last couple of months.
Here's the code for the comment form below. Can you spot my solution? (No, it's not the "http:" part, which almost worked)
<div class="roundedcornr_top_473174"><div></div></div> <div class="roundedcornr_content_473174"> <div id=commentBox class=commentBox> <div class=pad> <h2>Post comment</h2> <form action="/blog/index.php" method=POST onsubmit="return validateCommentForm(this);"> Real Name: <input type=text name=displayname /><br> <span style="visibility:hidden"> Your Email (Not displayed): <input type=text name="email"/></span><br> Text only. No HTML. If you write "http:" your message will be ignored. <br> <textarea cols=60 name=comment rows=10 wrap=soft ></textarea><br> <input type=submit value="Post" /> <input type=hidden name=id value="75"> </form> </div> <div class=comment> </div> </div> </div>
Fun with Colour DifferenceAre you looking for a nifty way to choose colours that stand out? Are you the type of person who is not satisfied until you have mathematically proven that your choice is optimal?
VP trees: A data structure for finding stuff fastLet's say you have millions of pictures of faces tagged with names. Given a new photo, how do you find the name of person that the photo most resembles?
In the cases I mentioned, each record has hundreds or thousands of elements: the pixels in a photo, or patterns in a sound snippet, or web usage data. These records can be regarded as points in high dimensional space. When you look at a points in space, they tend to form clusters, and you can infer a lot by looking at ones nearby.