Article:
  Is Perl Still Relevant?
Subject:   Sales vs. demand vs. use and that graph
Date:   2005-07-22 22:00:27
From:   bioinfotools
Two thoughts:


1. A quick glance at the graph that Tim links to in his article, shows that it doesn't show error margins or confidence bars. You really can't interpret data without confidence estimates, etc. Tsk, tsk, Tim, should know better! :-)


I wonder if once they are added in the only conclusion might be that Visual Basic is declining (which it certainly is in that graph!). All the other languages are nearly remaining constant, which is interesting in itself.


Java does seem to have the erratic (wavering?) support, though and its interesting how strong C/C++ is despite the brickbrats people hurl at it. And .NET has never really taken off to be the big all-sweeping thing Microsoft probably wanted it to be, it would seem.



2. Book sales, demand by employers and use by choice will all differ.


FWIW, perl is very strong in bioinformatics, although I've come to hate aspects of the language.


There is also a historical element in that many projects are linked to a language for legacy reasons.


Without boring on, I suspect the reality is more complex than book sales alone would indicate.

Full Threads Newest First

Showing messages 1 through 3 of 3.

  • Tim O'Reilly photo Sales vs. demand vs. use and that graph
    2005-07-23 05:59:33  Tim O'Reilly | O'Reilly AuthorO'Reilly Blogger [View]

    Why would there be error bars or confidence levels on this graph? It's actual reported sales data, not an estimate.

    Reality is of course more complex than book sales alone would indicate. But just as book sales of Perl in 1996 told me that people were under-valuing the language -- it was the number one computer book in revenue of 1996, but the language got no coverage in the computer trade press -- led me to launch my Perl conference and other Perl activism, and eventually my open source activism in general, so now, the same data set tells me that Perl is in something of a decline.

    As I said in the article, the story is far from over. And in fact, you can see the reversal in the Java slope that happened with the release of Java 1.5. So a new release will bring new life to the book market.

    The data is the data. Make of it what you will.
    • Sales vs. demand vs. use and that graph
      2005-07-23 16:58:42  bioinfotools [View]

      Sorry if I'm not making myself clear. The thing that got me looking at the graph was trying to distinguish fluctuations from trends.
    • Sales vs. demand vs. use and that graph
      2005-09-13 10:43:13  adamdonahue2 [View]

      Book sales are an unreliable indicator of "popularity" in so far as the share of Perl users may be more likely to find their information via other means. People buy books to help them learn a language. Perhaps Perl programmers, as a group, are more erudite when it comes to programming than users of, say, Visual Basic.

      I'd also think that the /type/ of books being bought would be a factor here. Are they introductory texts, reference books, or cookbooks?

      Finally, "popularity" is only partly the question. Deployment is important, too. If Yahoo's, eBay's, and other large companies' infrastructures are built on Perl, it's highly unlikely the language will "fall out of favor."

      Obviously there are many flaws in using book sales as an indication of Perl's relevance.

      Adam