Since my last use.perl post about Perl's perception issue seemed to generate a moderate amount of interest, I decided to do a bit more research about, well, research. In order for it to be effective, a clear goal needs to be established. I think a first pass at said goal would be "creating the perception of Perl as a modern, useful language." If that alone could be done, most of the other issues might take care of themselves. It will be a tough thing, though. I still talk to people today who don't know Perl has OO programming. We've had OO for fifteen years, but when people still don't know this, we've failed miserably (of course, this might not be a widespread perception problem).
Once a goal is defined, there are two questions which need to be asked before market research is feasible:
In other words, this is something which might need to be a joint effort between P5P and either TPF or Enlightened Perl. Polaris research has a handy guide explaining some market research basics. This is something which could be done. Whether we'd do anything with it or not is the real question.
At times I feel like some people in the Perl community are hoping Perl 5 dies a quiet, dignified death. Do we care about Perl 5, or are all eyes on Rakudo? I certainly hope people still feel there's merit to Perl 5, but if the Perl community gives up on it, why shouldn't the rest of the world?
"Do we care about Perl 5, or are all eyes on Rakudo?"
Do we care about a language that runs important infrastructure in practically every large company on earth? Hell yes. I wrote some of that. Do we care about a language that is not yet ready for production? Maybe. Well, a little at least because there's a small bunch of people doing impressive amounts of design and implementation work.
*My* future wrt. Perl dialects is with Perl 5. That's not going to change for another ten years at least.
Steffen (who is part of the non-silent minority)
Do we care about Perl 5, or are all eyes on Rakudo?
I care only for Perl5, and most of what I have to say about Perl6, or Rakudo, is not welcomed by the much vocal Perl6 groupies, hence I usually stay quiet when they blog about it. But then again, it feels like I am lightyears away from them anyway, with my requirements for stability and reliability, and with my core software upgrade cycles anyway, so the only thing Perl5 and Perl6 have in common is the prefix of four letters to their name.
Re:Perl6 is more marketing than substance
ajs on 2009-09-18T16:15:58
I know this was a while ago, but I feel it requires a response.Do we care about Perl 5, or are all eyes on Rakudo?
I care only for Perl5,
Really? Do you care about Perl 5's given operator or smart matching? Do you think those are "Perl 5 constructs?" Do you really need to care?
most of what I have to say about Perl6, or Rakudo, is not welcomed by the much vocal Perl6 groupies, hence I usually stay quiet when they blog about it. But then again, it feels like I am lightyears away from them anyway, with my requirements for stability and reliability,
Users of Perl 4 (and I was among them) said exactly the same thing about this crazy Perl 5 thing that was going to add an object system and make data structures far too complicated to support in the "real world."
the only thing Perl5 and Perl6 have in common is the prefix of four letters to their name.
Hmmm... no. There are a plethora of features of Perl 5 that remain in exactly the same form in Perl 6. They also share many conceptual similarities ranging from the use of expression context to the use of twiggles to future-proof existing variable names. I think that a lot of people read the specs as they were being written and focused on all of the things that they couldn't imagine doing in Perl 5 instead of the fact that programs will largely look like Perl 5 code. Here's an example:
sub is_prime(Int $n is copy) {
$n = abs($n) if $n < 0;
given $n {
when 0 { return False }
when 1 { return False }
default {
for 2, (3.. sqrt($n) :by(2)) -> $f {
return True if $n == $f;
return False if $n % $f == 0;
}
return True;
}
}
}This is Perl 6. The words True and False and the ordering of iterator and list in the for loop are different. There is a type and formal parameters. Other than that, this is bog-standard Perl 5 and perfectly valid Perl 6. If that doesn't bear any resemblance to Perl 5 then what language do you think is bears a resemblance to? It's certainly not more like C, shell, AWK, Python, Lisp, Ruby, Forth, Java, etc.
I'm still very much interesting in promoting Perl 5. I don't see Perl 6 impacting my business for at least five years.
I'm very interested in getting involved in this project. Let's get together in Lisbon and discuss it.
Re: Defining the Marketing Problem
Ovid on 2009-07-26T17:22:07
Excellent! I would love to get a "marketing" BOF together
:) Re: Defining the Marketing Problem
davorg on 2009-07-27T13:30:40
I think most people, including me, are going to "Marketing Campaign" and think that means spending money on advertising.
It doesn't have to.
I, personally, think there's a lot of value in just doing the survey part. Just knowing which issues are hurting us the most makes it easier for us to fix the right problems.
Imagine what would happen if we did a "Try Perl at perl.com" or "Try Perl at perl.org" today? If we were to spend a hypothetical million dollars on Magazine, TV and Radio, would we actually be in any kind of position to do that?
Right now, I'd say no.
Re:We don't necesarily need a marketing campaign
Ovid on 2009-07-26T17:25:55
Marketing could mean money, but it doesn't have to. Frankly, most Perl people aren't marketing people, so we really don't know what's involved, but something as silly as putting up a "I'm Java/I'm Perl" videos on Youtube might take off. Lots of marketing seems to be throwing lots of stuff at the wall and seeing what sticks. I'm quite happy for us to try and figure out marketing which doesn't cost money, but I also don't think that spending money for long term gain would necessarily be bad. It would just be bad if we don't know why we're spending it.
I am, however, quite serious that we while exploring this idea is worthwhile, it less worthwhile if the marketing/p5p follow-ups don't happen.
Re:We don't necesarily need a marketing campaign
mpeters on 2009-07-27T14:07:31
I've been thinking about what we could do as programmers to help the marketing side of things and I think I have an idea. Lots of people come into a new language because they want to hack on some system/program written in that language and the cooler and "prettier" the app, the more people that use it and want to hack on it.
So I think the best thing that us programmers can do for Perl's image (5 & 6) is to write some killer programs. Not cool CPAN modules (which are only visible inside our little bubble) but full fledged apps. We'd also need real designers doing real design (I know we try as programmers, but we fall really short in this area) and TPF could probably help out here too. There's no reason that TPF grants have to go to Perl programmers. They could also go to non-programmers helping out with Perl programs.
Now, I don't have the killer app idea in mind yet, but I'm pretty sure it's not blogging software
:) Re:We don't necesarily need a marketing campaign
Aristotle on 2009-07-27T20:05:11
We’d also need real designers doing real design
This precise utterance is actually the reason why Perl applications don’t have good design.
I appear to be one of few Perl programmers who have any direct appreciation (if not particularly great skill) at design. In contrast, both PHP and Rails are infested with these people. I have seen reams of postings about typography out of the Ruby people I follow, f.ex., and not a single thing about it from a Perl person, if memory serves.
And that means whatever full-fledged web apps these people write, they’re going to look at least passable, out of the box; even if the developers themselves are not great designers, they’ll know to build their apps from things that give them good defaults, and they’ll be able to tell better design or a small improvement when it comes along.
In contrast, the Perl community reminds me of what Paul Graham once wrote about American car companies: because the executives had no taste themselves, they didn’t know how to tell who to hire to make tastefully designed cars, either. If you have no sense for design, you can’t tell who is a good designer. And Perl does not have a culture of appreciating design. This “we need real designers so they can put a good coat of painting on our crud” mentality is perfectly symptomatic of that.
Frankly, I feel pessimistic about fixing this. You can’t suddenly imbue a culture with a sense for design when it has never cared about design or had any appreciation for it.
And I agree that it’s a huge reason for the marketing problems that Perl is having: Perl stuff looks ugly.
The few Perl things that don’t, don’t advertise Perl as being a primary ingredient of theirs. The only exception I can readily think of are the SixApart applications, but I bet that even then, only MovableType is widely associated with Perl – and MT’s sun has long set.
The question is not "Is p5p willing to add new features to Perl 5?". That's simplistic.
One question is "Is someone willing to implement a new feature some people want?"
Another question is "Is someone willing to shepherd such a patch through the inevitable arguments on p5p?"
A very good question is "Is some pumpking willing to commit such a patch?"
An excellent question is "When will that patch be available in a stable and supported release?"
There are many answers to these questions. In my opinion, they all rely on the fact that release managers drive the development of Perl 5 in the absence of a language designer.
