Article:		Is Perl Still Relevant?
Subject:		Some objective, empirical evidence
Date:		2005-07-14 08:33:48
From:		haywood
Results of a job posting search on www.dice.com: (Posts in last 10 days, all locations, no restrictions): Perl: 2700 Python: 282 Perl AND Python: 188 Ruby: 20 Seems to me Perl is pretty relevant... Keep in mind that Mr. O'Reilly's job is to sell books. If Python and Ruby (and probably soon Groovy) are the hot topics de jure, thats what he's going to sell, and it will be expressed in his stats. For those of us who make a living actually writing scripts, I'd suggest the DICE stats are more relevant. Book sales are more a leading (and I'd suggest very leading) indicator. Perl has matured. Like C, its become so de rigeur that employers tend to assume anyone they hire will know at least some Perl. Its just not sexy any more. And since it doesn't have major corporations flogging for it in the way that e.g., Java, C#, or VB.NET do, popular press tends to disregard it. Conveniently, the most current issue of Sysadmin magazine - possibly the most popular mag among people who actually keep all those racks of web servers and DBMS's running - includes the results of their annual editorial survey: "Which languages do you use most often? Shell (92%), Perl(87%), PHP(61%), C(61%),Python(37%)" - syslog, Sysadmin Vol. 14, Number 7 In addition, many of the featured articles (Database access was the topic of the month) were essentially "Here's how to solve this problem using Perl" articles...even tho only the Wizard's monthly Perl column actually included Perl in the title. While I'd love to see faster progress on Perl6, I'd suggest the rumors of Perl's death are greatly exaggerated. If the perl community seems less inclined to astroturf for their favorite language, its probably because, due to the high demand expressed in those DICE numbers, we're too busy actually doing things with it.

Showing messages 1 through 4 of 4.

• Some objective, empirical evidence
  2005-07-18 11:51:20  AdrienLamothe [View]
  
  
  Well, a huge majority of people eat fast food, which doesn't make such food particularly nutritious. My experience is that most companies don't avail themselves of the best solutions, and that there is a time lag before a good tool achieves widespread use. I believe that most people purchase programming books for their jobs or to augment their skills to acquire a job.
  
  I like both Perl and Python, but when faced with the choice of choosing which to use for a critical piece of an e-commerce site I had to go with Python because of the strong exception handling (breadth and syntactic ease.) Losing data was unacceptable, and too many things can go wrong over a network, which made comprehensive error-trapping necessary.
  
  We need more decision-makers to read sites like this, so they understand the choices available and can make better informed decisions, rather than play it safe and follow the herd.
  
• Some objective, empirical evidence
  2005-07-23 18:12:55  bioinfotools [View]
  
  
  I agree with your point! (Well put, too.) As I was saying earlier, rather less clearly, book sales, demand by employers and use by choice will all differ. perhaps what we need is for employers to have more faith in the "experts" they hire to choose these things!
  
  One other thing while I remember: any relationship between fluctuations in Java text sales and the start of university semesters?
• Some objective, empirical evidence
  2005-07-26 23:40:09  AdrienLamothe [View]
  
  
  So, to differ from the argumentum ad populum, a language is relevant if it effectively provides results. COBOL is still effectively producing results, so it is certainly relevant to those using it. An otherwise effective language can lose relevance if not maintained to the point where it won't run on the operating systems people use. People are more likely to change operating systems than re-write a successful custom application program. Open source languages minimize this risk by allowing virtually anyone to alter and re-compile them; adherance to programming standards reduce the chances that the source code will require alteration.
  

• Some objective, empirical evidence
  2005-07-14 09:40:24  Tim O'Reilly | [View]
  
  
  Great points, Haywood. You're right that book sales and job postings may not be in step, and the fact that there's an increase in interest in other languages by book buyers is just one data point. We're actually trying to do some studies now on the relationship between job postings and book sales. This topic is very interesting to us as well, because, as you say, we make most of our money by selling books. But I want to be clear that we don't have any agenda here -- we're an equal opportunity promoter when it comes to technology.
  
  I'll also note that there haven't been a lot of new perl books for a while -- and we just released two wonderful ones: Advanced Perl Programming by Simon Cozens, and Perl Best Practices, by Damian Conway. I'm very much hoping that those two books move the needle significantly, and we see a nice uptick in the Perl book trend line as a result!