Aaaargh, more commenting problems

I owe many people some apologies: this site has been quietly eating your comments. There are filters set up to catch and discard spam comments, and they work very well. I’m getting thousand or so junk comments a week that are not making it through to be displayed.

Unfortunately, about a 5% of the junked comments are false positives. I’ve gone through and tried to restore the ones I could find, but that represents a far too substantial loss of blameless comments.

I’ve set up the comments section now to optionally allow TypeKey registration. I’m hoping that the software is smart enough to realize that if you’ve gone through TypeKey, you are not spamming, and that that will improve the accuracy of spam detection—so use it, if you don’t mind TypeKey. I might go further still and require TypeKey at some point; I suspect most people would rather jump through a few more hoops and have better reliability, than type something up and have the mysterious and ineffable spam filter decide to quietly shuffle your words off to the holding pen before erasing them a few days later.

Let me know any objections or suggestions you might have.


  1. Steve LaBonne says

    OK, so I’m hoping this fix will now finally allow me to comment from my work computer (yes, I’m officially allowed to do a bit of online goofing off as long as my work gets done.) Here goes…

  2. says

    The problem with TypeKey registration is that I’ll need to constantly clear up my cookies every time I move from one SB blog to another. That is why I so rarely comment on those SB blogs that require registration. Sometimes I want to have my cookies and history intact while browsing around. I don’t mind cleaning them up once a day, but every five minutes…

  3. says

    Well, maybe not. It seems to work one way and not the other: the more restrictive the commenting requirement, the less likely the information given in the comment field will be insufficient to satisfy the flaky software. It would make it easier to comment here, for those who have problems, and screw over some of those other scienceblogs.

    You haven’t been slumming and commenting over there, have you? Shouldn’t Pharyngula be enough for everyone?

  4. says

    man, to think of the massive amount of pithy, well-reasoned, airtight comments that i lost. and only the crazy, semiliterate ones made it through.

    a pity!

  5. oldhippie says

    I really sumpathise. I wanted on my web to allow people to post their own comments. I don’t get many so cannot afford a continuous watch. It became a total nightmare. Hundreds upon hundreds of pointless web links would be uploaded. It was not a commercial operation because I checked with some of the webs concerned, and also they would vary in nature from from ferry lines and electric companies to porn sites, so there seemed like a virus program people design just to mess up people’s interactive webs.

  6. Marine Geologist says

    “houldn’t Pharyngula be enough for everyone?”

    Well, actually PZ, I don’t know about the commenting but Tara’s picture is a whole lot easier to look at than yours.:-)

  7. wamba says

    I’ve noticed comments get filtered if they have several links in them. That means that extremely well referenced posts are treated as spam. Go figure. Why not put wamba on a ‘white list’?

  8. says

    Why, Orac? Do you have a blog or something?

    Some would call it that. I call it a repository for my self-involved ramblings.

    Oh, wait. That’s as good a definition of a blog as any I’ve heard…

  9. says

    Whaddya know, I already had a TypeKey account sitting around. And a Blogger profile too. . . dear gods, I’m slipping over into the twenty-first century. Whatever happened to that starry-eyed pizza eater who could program with the MS-DOS DEBUG assembler and score a hundred thousand points on Atari 400 Centipede?

    I think I’ll appreciate the bumped-up link limit. I lost a deeply insightful post on the fallacies of quantum consciousness (or something like that) the other day, because I had unthinkingly provided references for my arguments.

  10. says

    PZ, you might tell the boys at Seed to look into Akismet. It was originally a WordPress plugin, but they’ve now released a Moveable Type plugin too.

    It’s supposed to work extremely well, with an extremely low false positive rate.

  11. says

    I notice that the requirement has been “retconned” into earlier postings. That confused me at first, though I went through the registration, figuring it was new policy I’d get to eventually.

    OT: Have any of the bioscience people here read David Hull’s Science as Process?