phil ringnalda

Dave almost got it:

What if comments, by default, were deleted after 24 hours? What if the owner of the site had to check a box in order for a comment not to be deleted? That way if a comment had lasting value, the owner of the site could make sure it sticks around.

But that’s not really quite enough. Googlebot’s got a serious jones for my comments, and can’t imagine staying away more than 24 hours, and no matter how much or little benefit they get from just having their link indexed once, I don’t want it ever indexed, and I don’t want people who subscribe to my comments feed to have to wade through the spam, whether or not it will evaporate after a while. So, solutions:

Paul Freeman’s absolutely right that we need to moderate the URL, not the comment. Rather than moderate everything, and have comment conversations only happen while you are awake and online, or moderate-by-default but only after its already been crawled, let the comments go through, but don’t link any URLs, either the commenter’s URL or links in the comment body, until you’ve approved it. That takes care of the unwanted indexing, but still leaves you with dozens of comments saying “Check out my site for hot shemalez!!1!”

Luckily, so far, most spambots come in to older comment threads that have had a chance to build up some PageRank and get established in search engines. The conventional wisdom says that means you should turn off comments on all entries more than a few days old, but I disagree. Even if I never get another useful comment on an older entry, knowing that a new version of Movable Type means Joshua Kaufman will be reminding me that old hacks need to be updated pleases me so much that I’ll keep hundreds of other threads open, just in case. I feel the same way about only allowing comments on a select few entries: I (and you, whether you know it or not) just don’t know what someone’s going to want to comment on, or whether they’ll want to within a week, or a month. Fortunately, there’s a solution to that, too, though I’m not sure whether I stole it from Chad Everett, or he stole it from me, or we both stole it from someone else: if the entry is more than x days old, and the last approved comment is more than n days old, where x is probably 7 and n is 2 or 3, then the commenter shouldn’t be expecting quick conversation, and the whole comment gets moderated until you approve it for publication. The spambot authors could work around that, by either trolling update sites for new entries to spam or just navigating up from old indexed entries to more recent ones, but so far, for the most part, they haven’t shown much sign of enterprise. At least while it’s a Club that only a few people have, it ought to work out pretty nicely. Add in an expiration time on moderated comments, so that if you haven’t approved them after whatever length of time suits you they just evaporate, and all you have left to directly deal with are plain text comments on your last few entries.