With the way that CPAN is currently set up there is no way for module authors to get any idea of how many people use their modules. Ok, if the module in question is DBI or LWP::Simple, then you might have some vague idea that a _lot_ of people are using your module, but most of us don't write modules that are that popular.
You might get the occasional email from someone asking how something works, or perhaps a bug or two raised in RT or (even more occasionally) an email saying "thanks for the great module" but generally you're working in a bit of a vacuum and you just keep pumping out releases in the hope that someone finds them useful.
Over the last couple of days, I've (accidently) discovered a good way to get a better feel for how many people are using your module. I released a couple of versions of Calendar::Simple that were broken in nasty ways. I realised what the first problem (in 1.15) was pretty quickly, which is why I released 1.16 very quickly afterwards. And then, within a few hours, I started getting reports of another (even bigger) problem in 1.16.
I didn't think that many people used Calendar::Simple. But now I know that there are at least a reasonable number of people who use it and get upset if it gets broken. Which is a good feeling.
And I've released version 1.17 which fixes all the reported problems.
Update: I spoke too soon!
I guess we should all break a couple of our module releases to see how many people care, then.
It reads like you "deliberately" broke your module and released it to see who would complain. I know that probably is not the case, but that is how it reads.
Re:That reads bad...
Aristotle on 2006-10-17T12:15:40
I’ve (accidently) discovered a good way to get a better feel for how many people are using your module.
I dunno about you, but that seems explicit enough to me.
If CPAN::Reporter becomes more widespread, then one way authors will get a sense of usage is to see how many test reports have been submitted against their modules. It's far from perfect, but it's another data point for how often a module gets installed relative to others. I quite enjoy watching my CPAN Testers RSS feed just to see what's been tested lately.
Ditto for CPANPLUS, of course, but it's less commonly used than CPAN.pm.
Re:Side benefit of CPAN::Reporter
offerk on 2006-10-17T13:40:14
Sorry to split hairs, but the OP was talking about how many people are *using* a module, while you are talking about how many times a module *was installed*. Two *very* different things...:)
Re:Side benefit of CPAN::Reporter
dagolden on 2006-10-17T14:55:14
I know. I said "get a sense of usage" and "far from perfect" to indicate that. Likewise, complaints about breaking a module doesn't really say how many people are using a module, just how many people (a) upgrade to the latest release and (b) report breakage.
Short of sending a "phone home" packet whenever a module is loaded (which would be very naughty!) -- there's no real way to know about usage. Even the Phalanx 100 is based largely on download counts, which implies but doesn't equal usage.
