The rates are not currently what is being measured. What is being
measure are the total results.
> Think of the number of mentions per day as a velocity. That's an
> interesting metric, to be sure, conceptually similar to our intuitive
> notion of popularity. So is the "acceleration" -- how fast is that
> velocity increasing? Perhaps even "jerk" would be interesting. What
> is of no interest whatsoever, however, is the "distance" traveled
> (total number of mentions), since the cars (languages) all started at
> radically different times. I can't think of anything useful that can
> be derived from a distance metric (given that we already know how long
> the languages have been around).
I really like that analogy - and I've used it myself.
However, I think you can find a lot of meaning in the 'distance',
although I think the idea of making 'acceleration' more visible would
be good.
Just because COBOL isn't used for a lot of new projects doesn't mean
that it's not "popular": indeed, in terms of job listings on
craigslist, it does better than a lot of the other "long tail"
languages. Same with books: if you want to learn about that language,
you can doubtless find more material than, say, Clojure.
Another example is Lisp: it's been around for a very long time, and
yet hasn't covered that much 'distance'. That tells me something.
The site does have a timeline feature:
http://www.langpop.com/timeline.html
But I think it needs some reworking.
--
David N. Welton
Nicholas