1. About four years back, the programmer we hired straight out of college started us off using PHP. He was a very competent Perl programmer, but he said access to session info and to the database was a lot easier in PHP when doing something small. The only thing that stopped further growth of PHP development in our shop was an old BSDI Unix box that didn't handle threads properly. Without that old box, scripts in PHP (and maybe Ruby) would be a significant part of our code. There's a strong draw to knowing about and participating in the thing that's "happening". That, coupled with even moderate technical merit, is enough to draw people's attention to something.
2. At a 4th of July picnic this year, an Electrical Engineering professor (whose interest is other than programming) was suprised to learn from me that Perl was a large and full featured language, and that we were planning to translate our older core software into Perl. She thought Perl was mostly just for web scripting, and pretty much the "old guard" for doing that .

The number of people using a language decreases when losses outnumber gains; when the number who wander off and start using something else that looks more intriguing outnumber those who try it, like it, and stay with it.

Perl needs a new, snazzy web application server right now. Something that has the performance of mod_perl but doesn't expose Apache's internals; something as easy to install as PHP or RoR, with a great name and a great API. (A toned-down version of mod_perl coupled with T2T and Mason would suffice.)

This is good advice which addresses a real need. It's worthwhile focusing some attention on this aspect of what will make perl stronger in the long term, however long it takes for Perl6 to be usable in production code.

I disagree with the OP on Perl6, both in content and style. The rewrite of the language is needed, whatever you think of the role of Parrot. Also, saying that it's taking too long just irritates those who are generously contributing their time and creativity (to whom the rest of us should be very thankful), and it does nothing to recruit more hands to the cause. It's better to say something like "Perl6 needs to get out there soon to ensure that your favorite language remains strong and healthy. Help wherever you can. Here's how ..."

Eric, I never suggested Perl was only a web scripting language. My point was that web scripting is what made Perl mainstream. Before the web it was just a fringe language. If the Perl community had only realized this and tried to accommodate web developers better, then Perl might still be attracting lots of new programmers.

There needs to be a compelling reason for people to learn Perl, otherwise the language will begin to rot. It doesn't have to be web scripting, but it needs to be something. Perl just being a great language isn't enough. Python has almost completely taken over the desktop scripting market that Perl was once making great inroads into. When people want to write small to medium sized GUI programs, they turn to Python, not Perl. PHP is now on the command line too. Ruby has surged in popularity in the past year or two, almost entirely because of its new web development platform Ror. These other languages are growing. Perl is shrinking.

"Python has almost completely taken over the desktop scripting market that Perl was once making great inroads into".....seriously? I mean maybe it has while i've been hiden away in my turtle shell. I certainly know it is common to come across VB and C++ solutions, maybe even Java, but i've never been searching for something and come across python at all. That's hardly proof that it isn't taking over, just that if it is, I've failed completely to notice. Is there even any way to prove that type of thing? If so what kind of metric would you use and can you compare a couple of language?


___________
Eric Hodges

Yesterday, I was surprised to find out that my employer has Ruby and Python code running somewhere on the production servers. Doing what, I don't know.

⠤⠤ ⠙⠊⠕⠞⠁⠇⠑⠧⠊

I see where anonymonk got that impression, I think: Python really looks like the new Universal Solvent being used to dissolve all system administration problems by a growing contingent of faithful devotees in the realm of open source unix. It has gotten to a point where people are reinventing wheels left and right (I tend to guess because somewhere in the backs of their brains they think those wheels will roll better if they're made of pythons instead of perls). Start paying close attention to what options come up while searching for candidate applications to solve problems using a software management system such as Debian's APT, and you'll probably notice that a lot of stuff is described not only by its functionality and feature set, but by the fact that it's written in Python. Perl doesn't tend to get that attention: utilities written in Perl are described by way of functionality and feature set, and that's it. Often, to find out what language was used to write it a Perl utility, you have to apt-get source the thing and look at the shebang line. More to the point, the Perl stuff mostly tends to be old, currently maintained and updated but not newly conceived, material.

There's a tremendous body of code written in Perl, though, and Python is still playing catch-up. There's so much system administration stuff written in Perl that it may very well be the case that we're not getting massive growth now because almost everything has already been written (in Perl), and nobody needs to write it (again). Thus, Python looks like it's thriving while Perl looks like it's dying from certain perspectives — perspectives that don't take into account the fact that much of the Python development is duplicating old work in Perl.

That's not the whole of it, of course, but I think that accounts for some of the situation, and some of the perceptions that have arisen.

print substr("Just another Perl hacker", 0, -2);

- apotheon CopyWrite Chad Perrin

Maybe the keyword is scripting, I don't know. However, it may just be the posters choice of platform - python is very common as a desktop language writing all sorts of (mainly small) apps under Linux. In Ubuntu, apps are either compiled or Python, as is much Gnome apps in general. Perl or Ruby are sadly rarely seen at all. Numbering totally, against Win and Mac, it would be insignificant, of course.