Spam Barriers (Redux)
Today we are posting a revised version of the word verification system we released yesterday. With this version we have resolved a number of the problems from the initial launch - the most important of which was the inability of some users to solve the CAPTCHAs presented.
There should also be fewer false positives. However, as I mentioned earlier, with any automatic classification of spam there will be some legitimate content that gets classified incorrectly.
It's important to know that if you are prompted to solve the CAPTCHA, it doesn't mean that there is anything wrong with your blog. Because of the number of variables our classifier uses, there's no easy way for us to pinpoint why your blog may have tripped the word verification (publicizing this information also serves to defeat the classifier).
The number of false positives will affect only a small percentage of the overall Blogger community. However, I know that for those of you asked to answer the word verification that this is a true inconvenience and for that I apologize.
We will be continually improving the classifier to reduce the number of false positives. We're also working on ways so that once a blog with word verification has been established as legitimate, the blogger will no longer need to solve the CAPTCHA.
It's important that we find ways to put reasonable barriers in place to further prevent the automated creation of spam content. This is not just to prevent the contamination of search indexes with spammy search results, but to ensure the quality of Blogger's service for everyone.
0 comments:
Post a Comment