Friday, February 7, 2014

Perl vs Python 2013, some actual stats..

I saw a post just now of some stats about Perl in 2013, by Gabor Szabo:  The Popularity of Perl in 2013.

Anyway, in response to another ridiculous exchange I had with someone last weekend, I gathered the following stats from sources that actually indicate actual popularity and usage levels, so here they are:


Book Publication Rates

Perl books: On Amazon, three new books were published in 2013 about Perl, but only one of them was picked up by a publishing house that did its own printing run.  There were also new printings of several new translations of very successful Perl books into languages other than English.

- Perl for Bio Informatics I - was self-published by David Scott. It has zero reviews on Amazon and a purchase rank in the millions (this means millions of other books sold more copies).

- Perl Para Geocientíficos (Spanish Edition) - was self-published by Dorian A. Oria San Martín.  It has zero reviews on Amazon and a purchase rank in the millions.

- Perl One-Liners: 130 Programs That Get Things Done - was published by No Starch Press, written by Peteris Krumins. It has only one unsolicited/unpaid review that looks legit, and that reviewer gave it 3 stars. It has a book rank in the hundreds of thousands.

There are 2 Perl books scheduled for publication in 2014, but they aren't new titles; they are new editions.

Python books: Python has 31 titles listed for publication already in 2014.  Python had 128 titles published in 2013.

Perl job listings: Indeed.com aggregated 343 job listings with Perl in the title, and 7,962 job listings that mention Perl but don't mention Python, PHP, or Ruby.

Python job listings: Indeed.com aggregated 1,151 job listings with Python in the title, and 8,755 job listings that mention Python but don't mention Perl, PHP, or Ruby.

Perl Reddit threads: Perl has around 10,000 posts
Python Reddit threads: Python has around 62,000 posts

Commits per month: Of the 661,017 open-source projects tracked by ohloh.net, 1% of commits are in the Perl language each month, on a steady decline from 4% ten years ago, and 10% of commits are in the Python language each month, on a steady incline from 4% ten years ago.

Committers per month: Of the 3,270,420 open-source programmers tracked by ohloh.net, 1.3% of people pushed a commit in the Perl language each month, on a steady decline from 4% ten years ago, and 8% of people pushed a commit in the Python language each month, on a steady incline from 3% ten years ago.

Of the 21,121,142,262 lines of open source code tracked by ohloh.net, 0.5% of the Monthly Lines of Code Changed are in Perl and 3% of them are in Python (both starting from 2% ten years ago).

2% of the Monthly Projects With At Least 1 Line of Code Changed are in Perl (steady decline from 4.5% 10 years ago) and 7% of them are in Python (steady incline from 2.5% 10 years ago).

Module Count Rates (modulecounts.com):

CPAN is the only language repository that is currently heading in the negative direction (as of very recently) in number of modules.  The rest are steadily growing. You can see Ruby and Java are neck-and-neck at the top there with more than 70,000 modules each.



see yesterday's report from codeeval, a company that provides online programmer candidate screening tests:




Perl got 1.6% share of their programmer candidate tests in 2013, compared to Python at 30.3% and Java at 22.2%.

This seems pretty representative of Perl's usage in industry based on every other measurement (including job listings on indeed.com), and it feels about right when thinking about the attendance of the Perl events at Oscon last year... there were maybe 50 people at the State of the Onion and Perl lightning talks... and around 3000 attendees at Oscon in total.

Here's the Python perspective [to reiterate Gabor's link]: Python Is The Language of the Year

4 comments:

  1. Dammit Zoffix. I was happy when you cleaned up hundreds of unnecessary modules from CPAN, but now you're just making us look bad ;)

    ReplyDelete
  2. Yeah, the number of modules is, IMHO, a negative aspect to a language. How many of those are repeats? Given the history of CPAN I would say "a lot".

    ReplyDelete
  3. Given the number of repeats in the modules on CPAN, I wonder how many there are in the repositories of the other languages. I also would prefer it if we battled the other repositories on quality of code; versus raw numbers.

    ReplyDelete
  4. thanks for uploading those stats! that's wonderful news to the programming community.

    ReplyDelete