hutcheson - 1:35 am on Oct 3, 2007 (gmt 0)

The standard policy is "unique content." For each site, the question is "what information does this site add?"

And it's not a "yes-or-no" question -- it is a gradient. Some sites (and, yes, directories are particularly prone to this) subtract from the sum of human knowledge merely by existing. A directory has to be fairly good just to be totally worthless. (Few ever get that good: most remain an intellectual and economic drag on the internet.)

Here's a way of quantifying a directory site's value. Imagine a user, looking at all the BETTER sites on the internet first. (1) What are the chances of him NOT finding what he wants in any of them? (pretty small?) (2) What are the chances of him perservering to look for the information here despite not finding it online elsewhere? (not so great?) And what are the chances of him finding what he wants in the site being reviewed? (pretty small?) Multiply them together. (VERY small? infinitesimal? negligable?)

Now compare that value (and it's likely to be tiny) with the chances that the user looking at this site (and failing to find his content) will exhaust his patience and cause him to quit looking: and thus not find his content when he COULD have found it, if this site had only died first.

And remember that volunteers place a very high value on their own time. You can't keep paying them to do scutwork or busywork. You have to show them they're being productive. You have to let them be as productive as they can--and here, that means not wasting time looking for lumps of coal in old slagpiles. It means editors will naturally focus on where they can make a difference. And sites that seem to be nothing more than yet more "mee, too" "resources" are therefore going to be neglected, in facor of work that builds value for users.

All of this is most certainly fuzzy logic. Different editors have different boredom tolerances (and this is a good thing, this difference is what makes a community function!). So yes, one editor could let "B&B directories" categories rot forever. Another pokes around once or twice a year to see if anything really new has come along. Another brings his Hoover vacuum/Black and Decker Blender into the category, just to enjoy the feeling of cleaning up the worst of the mess (so editor #2 will be more likely to find something useful when he comes back next year.) So you can't draw a hard-and-fast rule. You'd check out the content already listed--the number and quality of sites. You check out the content not listed--the number and quality. And you'd ask what's good for the users.

Certainly, two or three good competitors is a good thing. And equally certainly, fifty "comparative shopping" book shopping sites for a half-dozen real bookstores, is a most evil, pernicious, and deplorable condition. Visible (and identifiable) uniqueness--in authoritativeness, perspective, comprehensiveness, kinds of useful information,--three or four of these conjoined would tilt the scales far towards a listing. Lack of two or three of them would tilt the scales far the other way.

I've said this before, and I believe it: aiming to be the second-worst directory listed in a category is doomed to failure: after all, the worst directory will probably disappear next time the category gets extensive work anyway. Aiming to be the average directory, is doomed to failure--there's an average site already listed. Aiming to be the best, and almost succeeding, is the only reliable way of getting a listing.

The fairness issue usually arises here. But it ought not to. What we all want to do is encourage new, different forms of content--by focusing effort on finding that, rather than no pilpulistic evaluation of what are, after all, cookie-cutter sites. And, in any case, all that matters is being fair to the users who don't want to sift through a dozen directories to sift through a dozen different random samples of the actual data set. If we can do that, we've accomplished something.