“PHP Eats Rails for Breakfast”

» 21 October 2006 » In PHP, Tech »

That is indeed the title of the article on Ohloh, a site that collects information on open source projects and gathers statistics by analyzing the source code of those projects. So far they’ve indexed over 3,000 projects and their conclusion seems to be that among Web scripting languages, PHP is the undisputed champion (as measured by the LOC count).

Measured purely by the number of new lines of code, PHP leaves all other web-scripting languages in the dust, and continues to grow. Quite simply, one-fifth of all open source code being written today is written in PHP.

This, combined with their observation that the relative number of developers working in PHP is not increasing and that the number of new projects being started has actually decreased, leads to an interesting premise: the trend for for PHP-based projects points towards mature code bases. Meaning that the developers prefer to work with existing applications or frameworks and increase their output through updating these applications rather than starting new ones. In general, that seems to be a good trend, if one can rely on the data gathered by Ohloh. I have not done an exhaustive examination, but there are some puzzling things in the data on PHP project itself. For example, how do they calculate “man-years” for each developer? My number is 5.5, which puts me fourth on the list behind Rasmus, Derick, and Frank. I know I’ve been working on PHP longer than that. And then, Andi and Zeev are actually below me with 5.3 and 5.2 man-years respectively. The account with the most man-years seems to be our automatic changelog committer script, even though it committed a total 0 lines of usable code. 🙂

Even without Ohloh’s data, one could venture to say that the code base of projects using PHP is larger than those of Ruby or Python. A lot of that is due to the accessibility of PHP, I think, but does pure lines-of-code count give indication to the quality of said code? Of course not. I think Ohloh is providing a useful service by tracking this kind of information, but I would take its results with a few grains of salt.

Trackback URL

  1. andrei
    michael schurter
    21/10/2006 at 8:53 pm Permalink

    I think you’re using statistics of dubious accuracy to arrive at some incorrect conclusions.

    Not only do LOC != quality, but LOC != maturity or popularity either. LOC is one of the oldest and most meaningless code statistics ever collected.

    About the only thing is useful for gauging is how verbose a language is.

  2. andrei
    Andrei
    21/10/2006 at 9:04 pm Permalink

    Michael,

    I believe you are incorrectly attributing these conclusions to me. I merely related some of the text of the Ohloh article. If you read my last paragraph again, you will see that I question the LOC statistics myself. “Does pure lines-of-code count give indication to the quality of said code? Of course not.”

  3. andrei
    michael schurter
    22/10/2006 at 5:59 pm Permalink

    Sorry. 🙂

Additional comments powered by BackType