Slightly inflammitory but please bear with me. Also, I'm still a blog nomad, so I thought I'd nomad on over here. This came about from online conversations elsewhere.

Perl may or may not be hemmoriging programmers. In my experience, it is. For years, I've seen highly talented Perl programmers, one after another, leave for greener pastures.

As long as we don't understand how and why people pick languages, we'll assume that we're competing.

People used a lot of Perl in the late 90s and early 2000s. perl was widely ported to the OSes used at the time for Internet servers (SunOS, Solaris, IRIX, Linux, BSD, etc, etc). It supported the most common interface offered by the most common webservers, CGI. When Apache mod_ stuff started taking over from CGI to offer higher performance, mod_perl was much crankier to install than mod_php. The API offered matched Apache's low level request handling which wasn't very useful for most Web apps. PHP simplified it. PHP bundled mysql support.

It took the Perl camp a while to realize that PHP was winning, and then why it was winning, though it's easy to state these things after the fact. Here's the thing though: PHP is continuing to win right now, and I don't think most people realize it.

Perl programmers tend to be stuck playing catchup, and they haven't yet realized that people aren't building apps on Web frameworks any more. We're so happy with our wonderful Web frameworks and the efforts that started with Maypole that we haven't realized that the vast majority of sites are not being built on Web frameworks any more.

They're being built on CMSes, and WordPress in particular.

Job sites are WordPress plus a module. Support. Small HR sites. You name it, it's built on WordPress. CMSes offer people out of box what they need: a pretty layout, user registration and permission system, easily editable "static" content, and a shopping plaza for dynamic logic to add in.

People don't hire PHP or Perl programmers; they hire WordPress programmers. People don't have an intentional deliberation process when they decide which language to learn. Instead, they're good with computers, and someone asks them to install WordPress. They do so. Then they're asked to install modules and themes. They do so. Then they're asked to change things around. And then they're a PHP programmer.

WordPress creates PHP programmers, not PHP.

WordPress is the new CGI, and PHP is dominating it. Having apps out there that people install and extend is what brings in programmers, and we're losing. Creating products creates jobs and creates programmers. Creating Microsoft Windows created huge numbers of jobs. Creating the Atari 2600 created piles of companies, many of which are miraculously still around today.

So, to the Atari 2600 thing. The 2600 has a version of the famous 6502 CPU, the 6507. The 6502 is awesome processor in one way: it was insanely cheap to build, relatively. It killed the Motorola 6800 in terms of lower transistor count and price. I don't want to piss anyone off, but the 6502 has a one byte opcode, and only half of the possible opcodes are used. It had some good addressing modes, but each of the three general purpose registers are very specialized. The 6800 has more registers, and they're very general purpose. That makes programming the 6800 much easier. That's the CPU that came before the 6502. The Intellivision's 1600 is effectively a 16 bit CPU that looks amazing to program for (the Intellivision was actually my first system). Supposedly, it was modeled after the PDP11's instruction set, but somewhat simplified. It has, if memory serves, 8 registers, two of which are hooked up to special logic to serve as stacks but are otherwise all general purpose registers.

And guess what? None of that matters, at all. People don't generally now and didn't then think, "huh, I want to program the 1600 processor because it's better than the 6502, so my business is going to write games for the Intellivision". They wrote games for the Atari 2600, or the Coleco, or the Intellivision because they wanted to write games for that platform. What's inside doesn't matter at all. In fact, that so many companies wrote games for the 2600 instead of other systems that had vastly more RAM and vastly more advanced video hardware proves that what's on the inside doesn't matter. What matters is what the kids at home are playing. It's all about the platform. You have to create the platform or you have to jump on the right platform.