C keeps its lead over Java, but index's accuracy is questioned

An important PHP developer dismissed the validity of Tiobe's language survey, which has C continuing to displace Java as the most popular language

The C language, which displaced Java as the most popular language in the Tiobe Programming Community monthly assessment in April, has maintained its number 1 spot in the May report released this week. But a high-profile member of the PHP community is questioning the accuracy of the index itself.

The May release of the index, which gauges the popularity of programming languages, had C used by 17.346% of developers, as opposed to 16.599% using Java. The index is based on the number of skilled engineers worldwide, courses, and third-party vendors interested in the particular languages. Examinations of search engines including Google, Bing, and Yahoo are used to calculate ratings; other sites such as Wikipedia, Amazon, YouTube, and Baidu are factored in as well.

[ Open source Java is moving to Linux and AIX on the PowerPC architecture. For more analysis of software development, subscribe to InfoWorld's Developer World newsletter. ]

But Andi Gutmans, who is a key developer of PHP and CEO of PHP tools maker Zend Technologies, dismisses Tiobe's index, citing its tendency to fluctuate. "It's very questionable how they come up with those numbers," Gutmans said. Tiobe's index for May has PHP as the sixth-most-popular language, but used by just 5.711% of developers. Gutmans insists PHP usage is growing quickly. He cited other industry figures showing 50% of the Web running PHP.

Tiobe Managing Director Paul Jansen defended the index, saying most criticism comes from declining programming communities. "Of course, we use a very simple algorithm so it is easy to shoot at if [it] doesn't fit your expectations, but it shows something about the visibility of each language on the Internet over time," Jansen said. "It is not scientific but at least it is the best try I have seen so far. If Andi Gutmans has some suggestions [for] how to improve the Tiobe index, I really would like to know. We are willing to adjust the index if he shows up with a good suggestion."

Tiobe in its report said the top 10 languages in its index had not changed much in the past eight years, apart from Objective-C moving into a slot and Delphi moving out. New language adoption appears to be much harder than expected. "The main reason for this is probably that it is very difficult to migrate a large code base from one language to another one. So changes are slow. But even if we take this into account, there are no new languages that show a slow but constant [rise]," Jansen said in the report.

Tiobe cited Scala as an example of a language that has not had much movement over the years, even if it is becoming a hot topic in Internet discussions. Currently ranked 46th, Scala entered the Tiobe index at number 57 in 2006 and moved up to number 48 a year later.  "Not much happened to the language after that," Jansen said in the report. Other examples of languages not moving forward much in the index include F#,  ranked in the 40th spot now and 38th six years ago; Groovy, now 32nd and ranked 36th in 2008, and Erlang, which is now ranked 44th and was ranked 35th in 2006. "So where is that next big programming language? Let us know," said Jansen.

Also in Tiobe's top 10 for the month were: C++, in third place used by 9.825% of developers, followed by Objective-C with 8.309% and  C# at 6.823%. Ranking after PHP were Visual Basic with 5.457%; Python with 3.819%; Perl with 2.805%, and JavaScript with 2.135%.  

This article, "C keeps its lead over Java, but index's accuracy is questioned," was originally published at InfoWorld.com. Follow the latest developments in business technology news and get a digest of the key stories each day in the InfoWorld Daily newsletter. For the latest developments in business technology news, follow InfoWorld.com on Twitter.

Read more about application development in InfoWorld's Application Development Channel.

This story, "C keeps its lead over Java, but index's accuracy is questioned" was originally published by InfoWorld.

What’s wrong? The new clean desk test
Join the discussion
Be the first to comment on this article. Our Commenting Policies