The early computer pioneer Alan Turing thought a good test of a computer’s intelligence would be if you couldn’t tell whether you were dealing with a human or a computer during, say, an extended e-mail exchange. To pass the Turing Test, the computer would have to answer and ask questions, understand slang and have a lively sense of humor.

It would also have to be rude, unpleasant and irrational at times, or people would guess right away that they were talking to a computer and not a human.

To pass the Turing Test, a computer could not reply, “You misspelled ‘mother’” to the message “My mothre just died.” It would have to understand the message and respond in a sympathetic way.

Turing came up with his test 57 years ago, and I’m not sure whether any computers have met his challenge. However, most computers aren’t designed to fool humans that way any more than a wrench is designed to remove screws. Acting human is not a goal to which computers aspire, so the test is never really applicable.

I say the true test of a computer’s intelligence is how it handles spam. If a computer can tell that an e-mail from someone with the unlikely name of Briton Elveros with a subject line of “elstuoba” is junk mail and the monthly statement from my cable TV company is not, it passes. I will buy that program.

Based on this simple test, most programs are complete failures. While mine correctly put the “elstuoba” message in the trash, it also put an e-mail from an old friend and this month’s phone bill in the trash. I have to go through the trash box every few days to reclaim the misfiled mail, including e-mail from services I subscribe to.

How do I know Briton Elveros is a fake name? I don’t. But I’m pretty sure I don’t know anyone named Briton Elveros, so why should I read his e-mail, especially when he (or she) wants to talk about elstuoba? I knew it was bogus right away, but my computer probably had to exert a good deal of effort trying to figure out how fishy it was.