Beefy Boxes and Bandwidth Generously Provided by pair Networks Cowboy Neal with Hat
laziness, impatience, and hubris
 
PerlMonks  

Perl is sinking (TIOBE): all time low for Perl

by larus (Acolyte)
on Mar 08, 2009 at 12:47 UTC ( #749122=perlmeditation: print w/ replies, xml ) Need Help??

Perl is going steeply down in the TIOBE list:
http://www.tiobe.com/index.php/content/paperinfo/tpci/index.html
What we should do :-) ?

Comment on Perl is sinking (TIOBE): all time low for Perl
Download Code
Re: Perl is sinking (TIOBE): all time low for Perl
by moritz (Cardinal) on Mar 08, 2009 at 13:11 UTC
    What we should do :-) ?

    Don't trust a page that titles "All time high for JavaScript" and won't even display text when Javascript is disabled ;-)

    Ok, let's try to stay serious for a moment... have you read what they actually measure?

    From what I understand they simply search "perl programming", "C++ programming" etc. with some search engines, and look at the results. So what does that actually mean? It's a quantitative metric, but does it actually mean something?

    IMHO it just means that some people have written some websites that programming language; so all it does is measuring the PR efforts. And yes, Perl can improve in that respect. But I hesitate to draw any other conclusions from these charts.

    What we should do :-) ?

    There are two answers here:

    • Blog about Perl, or do other things to improve the public image of Perl
    • Ignore shitty metrics.
      ...the PR efforts. And yes, Perl can improve in that respect.

      Exactly. Perl does not need to--it is the marionette's puppeteer--only those who look for it will find it, and once they do, they realize how much it actually does :). (What it is manipulating is up for decision.)

      And you didn't even know bears could type.

        For the sake of learning, I recently built a front-web page access a mySQL web page. The backend ETL part was implemented with Perl. PHP was used to query the front-end. The PHP code passed on JSON structures to my jQuery and Javascript with which my web page was built.

        Honestly, I did not see any advantage to using PHP. I have read that PHP has a better interface than CGI but I could not really tell.

        There were two reasons for selecting PHP. The first is that being new to mySQL, the book "Head First PHP and mySQL" was an excellent tutorial on mySQL and PHP seems to integrate well. PHP seemed very C-like and Perl so picking up the language was not difficult.

        Second, with so much talk about PHP, I felt that learning some PHP could only help my resume.

        However, the power or PHP does not compare to Perl. And for Web 2.0 web pages, Javascript and libraries like jQuery are far superior than depending on PHP.

        Comparing PHP to Perl, I can state the following: 1. The raw parsing power of Perl is unmatched. 2. The lack of an equivalent STRICT for variable declarations results in numerous debugging headaches. How can one build a robust application without having a strongly typed language? 3. Perl's variable declarations prefixes are so nice as they allow you to spot a scalar from an array from a hash. 4. Hashes are more straightforward. 5. Pointers/References ... No serious programming can be done without pointers, and PHP's roundabout use of pointers is not friendly.

        Of course, I am still learning about PHP and it does have a huge libary of functions. But the CPAN library is still unmatched.

        Now my backend for the mySQL database was done in Perl Almighty. There was a total of some 4200 NCL lines of code encompassing some 25 modules. The 25 modules were built into a single executable using ActiveState. My application FTP's into some 15 sites collecting information continuously. And Perl does a great job at FTPing! What a breeze to drop my application on multiple machines with none having Perl installed on them.

        The Perl DBI is first class. PHP seems to have copied the Perl DBI.

        I used caller() for my first time to help point out exactly where errors occured. My error_log_file - thanks to caller() - has been of such enormous help.

        My standalone tools in Perl have put many an engineer in shock and awe after seeing how Perl tools have shaved hundreds and hundreds of hours of manual labor from their work. Quite a few engineers around me have purchased Perl books after witnessing the speed with which I relieved them of their headaces. Of course, many give up quickly because without a strong software foundation, you can only do so much.

        It would been a nightmare to replicate this effort with PHP.

        After working for some 10 years with Perl, I still have to see a language that can match Perl's power. I have looked at Python briefly and did not like it. TCL is archaic. In the past 5 years, I have written over 60,000 lines of NCL code with it. Perl 5.10 has really made this language better. Having a SWITCH statement with SMART_MATCHING is great. But hash references are probably the most powerful thing in this language after regular expressions.

        Perl is an easy language to start with but takes time, patience, and opportunities to grow with it. Each new problem that I solve enlightens me more. PHP has grown because for Web Programmers, they can easily embed HTML and PHP. But with AJAX making waves, you are better off using Javascript and Javascript libraries for your HTML.

        Even if TIOBE is right as to number of users, that does not make Perl a second class language.

Re: Perl is sinking (TIOBE): all time low for Perl
by zentara (Archbishop) on Mar 08, 2009 at 15:23 UTC
Re: Perl is sinking (TIOBE): all time low for Perl
by Anonymous Monk on Mar 08, 2009 at 15:37 UTC
    TIOBE's methodology is wrong, read their definition. It is based on google, yet google has their own methodology for categorizing results and content, it is way more accurate than what TIOBE uses and it turns out it totally contradicts all the data the TIOBE has! Google says: http://tinyurl.com/d4wnsu

    Google shows that C# is way above PHP, Python, Perl, Ruby yet TIOBE doesn't?

    Look at what the TIOBE ranking is based on: The ratings are calculated by counting hits of the most popular search engines. The search query that is used is +"<language> programming"

    This is a totally obviously WRONG methodology. Effectively TIOBE is slandering just about every language they rank.

    Do not trust TIOBE, it is wrong and it even disagrees 100% with the results that GOOGLE has once the category of programming has been selected.

    TIOBE is bad methodology FOR ALL LANGUAGES.

      And even big G's rankings appear to omit a search term which might drive up Perl's numbers:
      CPAN

      Check the dropdown box for "related to perl" at AM's Google link, http://tinyurl.com/d4wnsu which doesn't even offer "CPAN" (Of course, that might mean we refer people there with a caution against "reinventing the wheel" so effectively that that search is irrelevant?)

      The default display shows the red language at a very low (and only slowly rising) "general interest" level. Based on analysis with the Mark I eyeball, the deltarate (relative to category) may have been highest in mid-2005, while the all-time, so-far high "relative to category" appears to have been during late-2006, after which it trends lower to midway into the third quarter of 2008. In the overall "interest" tab, the same language appears to have pretty much flatlined since mid-2005. Does any of that strike you as probative of anything substantive?

      Many, perhaps most, single-issue methodologies for ranking anything as amorphous as "popularity" have serious flaws. IMO, that includes what -- for this particular issue --- may be better than most: the job search technique mentioned below by Gavin.

      Update: Fixed (s/slow/low/) typo.

Re: Perl is sinking (TIOBE): all time low for Perl
by perrin (Chancellor) on Mar 08, 2009 at 17:07 UTC
    I guess you'd better learn Logo, since TIOBE clearly shows it is the language of the future.

      You mean I wasted the past six months brushing up on Delphi?

      Finally! All those years of LOGO really paid off in the end.

      I keep telling everyone about it, but did they listen? Noooooo....
      LOGO - I thought that was the language of the past (childhood). Isn't that the language used in primary school computer labs to teach children programming using Turtle Graphics?
      In the fall of 1995 I started teaching my son's second grade class how to program using the LOGO programming -- Learning LOGO

      Now, if we could only teach all managers to hire second graders...

      Best, beth

Re: Perl is sinking (*****): all time low for Perl
by BrowserUk (Pope) on Mar 08, 2009 at 17:37 UTC

    Posts like this always remind of that Jim Cary film where the guy lives his life as part of a reality show.

    What we should do :-) ?

    Tomorrow morning, when it's time to decide what type of underwear to put on, consult google.


    Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
    "Science is about questioning the status quo. Questioning authority".
    In the absence of evidence, opinion is indistinguishable from prejudice.

      Another words its "Pants"

      or in the Geordie vernacular "A bag of shite"

Re: Perl is sinking (TIOBE): all time low for Perl
by ELISHEVA (Prior) on Mar 08, 2009 at 18:27 UTC

    The real question here is "who listens to TIOBE?". Do any of us know? I don't. If people are listening to TIOBE, it really doesn't matter if we disapprove of the methodology and can show how it disagrees with Google's own rankings. It will bite us, unless we are proactive. Of course, if no one is listening, then TIOBE's rankings don't matter at all.

    TIOBE is claiming that they can predict which languages are a good investment. Which languages will you be able to build your career around five years out? Which languages will you be able to build teams around not just today but five years hence when you have five versions of your project in maintenance mode? Do we really want college students and managers thinking they are jeopardizing their careers by investing in Perl? If they trust TIOBE, that is what they will think.

    Perl's success of course depends on the quality of the language and the quality of developers attracted to it. This in marketing terms is an "order qualifier": the thing that even lets us make the Perl pitch. But to actually make the sale we need more than a good language and a good community: we need managerial mind share and new blood - neither of which we will get if we simply dismiss those like TIOBE who, however awful their methodology, view those very managers and students as their target audience.

    TIOBE may or may not be important, but TIOBE rankings raise issues of deep importance. Moritz is right - this isn't the death of Perl, but it isn't exactly a PR success story either.

    So what can we do?

    • Find out who is listening to TIOBE. And if people we care about are listening,
    • Find out what other blogs, news sources, etc they also listen to
    • Get our critique of TIOBE's methodology out before the people who are listening to TIOBE by going to all the other places they listen to.

    Best, beth

      This thread in general, and your post in particular with its 10 references, have probably increased the Google ranking of this obscure and unimportant site way up. Even though (I think) most of the references have been negative.

      It's that old cliche coming true again: any publicity is good publicity.


      Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.
      "Science is about questioning the status quo. Questioning authority".
      In the absence of evidence, opinion is indistinguishable from prejudice.
        I agree. When we refer to it, let's shuffle the letters around or call it something else as a codeword. Maybe E-OBIT ?
Re: Perl is sinking (TIOBE): all time low for Perl
by larus (Acolyte) on Mar 08, 2009 at 18:59 UTC
    Ok, TIOBE's way is the wrong way to go. We do not have to count on TIOBE. Is there any better ways to measure the "success" or popularity of programming languages? I think that lots of people trust TIOBE's index, although it has its flaws and disadvantages.

      A good method is to look at job vacancies and see what languages are required and how much the salary is.

        A good method is to look at job vacancies and see what languages are required and how much the salary is.

        Looking at job vacancies gives some languages higher ratings than are probably realistic, due to the verbosity of the language. So, for example, a project that I would probably stick two programmers on if they were building it Perl, when built in Java results in 14 job openings.


        www.jasonkohles.com
        We're not surrounded, we're in a target-rich environment!
      I think that lots of people trust TIOBE's index, although it has its flaws and disadvantages.

      On the basis of what? Do you? :-)

      Personally, I have my doubts. Here's some food for thought:

      • Looking at the first few pages of Google listing for TIOBE I note that virtually all listings are either to (a) tiobe.com (b) simple statements of fact about the latest results - often by vested interests, i.e. language blogs tooting their language's successful position in the rankings - or (c) significant critiques of the listing. In fact, the first non-tiobe.com Google listing (at #3) is one such critique.
      • More telling, after scanning several pages of Google listings I failed to find even one article showing that someone was actually making decisions based on the TIOBE index. Compare that to metrics like COCOMO.
      • I sometimes like to scan my alumni print journal database to get a thumb-in-the-air sense of someone's off-line brand recognition. According to this database (a repository of 1300+ trade journals and news magazines targeted at managers), TIOBE has only four articles: three in Dr. Dobbs (2 from 2008, 1 from 2005) and one in ComputerWeek (2008). The content reflects the same trend found in the google hits: a self-promotional interview, a critical letter to the editor, and simple reporting of results. By contrast, the Gartner Group has 1407 articles, many of which use the Gartner Group results as a building block in their own independent analysis.

      Best, beth

Re: Perl is sinking (TIOBE): all time low for Perl
by sundialsvc4 (Monsignor) on Mar 09, 2009 at 02:55 UTC

    I have never heard of TIOBE before, and I won't choose to pay any attention to it starting now.

    Journalists love “the next big thing.” Even if they don't understand what is, they do know that “the next big thing, whatever it is” will probably help sell issues and maybe even subscriptions. If you confront someone with something they don't know too much about (an easy thing to do ... with everyone), and then you proclaim that “if you don't learn about this, then It Is The End Of Your World As You Know It.”

    Hogwash, sure. But it does sell issues and subscriptions.

    Become good at what you do. Always work to become better. But also become good at letting other people know why what you do is important to them. I'm not really “talking technical” here, because the business value of what you do really isn't that closely tied to how you do it.

    In other words, “you could provide good business value in COBOL. Or (ugh...) PHP. Plenty of folks are (still) out there doing just that. They can describe what their well-running software does for the company without mentioning any language-names at all.

Re: Perl is sinking (*****): all time low for Perl
by mr_mischief (Prior) on Mar 09, 2009 at 20:51 UTC
    There are, as already mentioned, several issues with the index you name. Even so, Perl only slipped 3 places by their count and is still at 5 times the threshold they use to determine "mainstream programming languages". They say "it is strongly advised to stick to mainstream languages for industrial, mission-critical software systems" on the page defining the index.

    Even if this is taken as a valid metric, it is only claimed to be one metric. The people who publish it specifically state it does not represent the best programming language or the amount of lines written in any particular language. Any intelligent discussion of a language takes far more into account than one metric.

Log In?
Username:
Password:

What's my password?
Create A New User
Node Status?
node history
Node Type: perlmeditation [id://749122]
Approved by planetscape
help
Chatterbox?
and the web crawler heard nothing...

How do I use this? | Other CB clients
Other Users?
Others meditating upon the Monastery: (12)
As of 2014-04-18 21:42 GMT
Sections?
Information?
Find Nodes?
Leftovers?
    Voting Booth?

    April first is:







    Results (472 votes), past polls