You Can Use Statistics To Prove Anything That's Remotely True (TIOBE 2008)
"Perl is dead", crows TIOBE's January 2008 index. The world belongs to Python.
You see what you want to see in statistics though.
For example, you could compare Perl, Python, PHP, and Ruby job trends. Don't drop those sigils yet.
Or compare Perl's delta to C's delta. Both lost ground in the TIOBE index, but C declined by almost twice as much.
Here's a fun one. TIOBE's editorial says that C# and Java will eventually be the two most popular languages. To do this, C# has to surpass Perl. That's a problem though; it gained more than Perl lost and still slipped a position and is still more popular than Perl.
Ultimately this isn't even good stats porn though. There's no analysis of why languages have gained or lost in popularity. Without that, there's no good way of deciding what these statistics mean. Without that, it seems silly to declare winners and losers and long-term trends. (One might also suspect that the actual release of Perl 5.10 and the buzz around that from the second half of December versus the "imminent" release of Python 3 may shift numbers from this point on.)
|I trust 'em. It's the return of COBOL and Fox Pro baby!!! Up 3 and 5 positions repectively.|
|Do I believe anything that says FoxBase/DBASE, Logo, Delphi are up-and-coming languages?|
Just a random note, look at http://jobs.arstechnica.com/. They started their own jobs page. 3 of the jobs are perl related. This could say a couple of things, good perl people are hard to find (they are). There could be just a lot of perl jobs out there and TIOBE is misreporting. Or programmers are moving away from perl, accentuating the jobs that need to be filled.
|And Google Trends? It show the real Search Volume.|
>There’s no analysis of why languages have gained or lost in popularity