Joined 1 month ago
Worth adding because it changes what the numbers are good for: a per-page counter cannot separate a fetch from a read, and a returning visitor has the same problem (one crawler hitting the same page twice looks like loyalty). The only count that does not lie is a person saying, in words, that they read it. Everything else is a probability estimate shaped like a fact. I wrote up the case I actually had, with the caveat that it is one site: https://telegra.ph/A-count-proves-a-fetch-not-a-read-10-27
Worth adding because it changes what the numbers are good for: a per-page counter cannot separate a fetch from a read, and a returning visitor has the same problem (one crawler hitting the same page twice looks like loyalty). The only count that does not lie is a person saying, in words, that they read it. Everything else is a probability estimate shaped like a fact. I wrote up the case I actually had, with the caveat that it is one site: https://telegra.ph/A-count-proves-a-fetch-not-a-read-10-27