Tommy Ollsen has had some trouble with spam comments on his blog. I can’t sympathise, because nobody visits here. 🙂 But in patience for the day, some accessibility and usability concerns…
I’ve always been very confused by blogs which have allowed comments, but not when you come across them (months later). It’s not intuitive or usable, particularly where discussion boards are based around the whole idea of years-long threads. So if you do go ahead with, make it very explicit that there’s a reason for no-comment, otherwise you get users looking around the page for the comment form, or thinking they’ve been banned.
If you’re going to close comments, make it clear on the page – it would particularly ne usefulwhen located in the same area/style division as the comment box would usually be, so that if you’re already aware of the site design it’s clear that something’s different on this page.
There’s an argument for closing comments, but there’s also a flipside. I’ve come across various important discussions (Mezzo/Malarkey spring to mind) by searching for them, and where there are still “gaps” in the conversation. That’s why the “bumping” feature of forums is a very useful feature, in that it can reignite debates which haven’t properly concluded. However, as a blog/journal it’s clearly a different matter, but you could argue that a “Ten recent comments” list could include all comments from all entries, rather than just the open ones, so that any user would see that the old/archived ones were still in use, and thus reflect/comment further on the new entries.
Consider whether it is worth pairing the timeout (based on the original post date + x days) with some sort of activity figure (if comments made in past 28 days > 2) which flags the entry for human intervention, to work out whether the conversation really is going on for months and months (and so where it’d be a shame for the timeout to apply), or whether a sneaky comment spam has just held it open.
In the end, I think at the moment – without resorting to some extreme filtering/AI techniques – in order to allow the blog to be sufficiently useful over a wide range of people and over a reasonable length of time, it’s going to have to be the compromise involving a machine making broad decisions and flagging up things it’s concerned about to a human, in the same way as simple email spam filters work. Sad but true.