Programming Language Popularity

These are the top languages used on SourceForge in December 2004:


        Projects Language

        -------- --------

CeePlusPlus
CeeLanguage
JavaLanguage
1. PhpLanguage
2. PerlLanguage
3. PythonLanguage
4. JavaScript
5. CsharpLanguage
6. VisualBasic
7. DelphiLanguage/Kylix
8. UnixShell
9. AssemblyLanguage
10. PL/SQL (StructuredQueryLanguage)

SourceForge, February 2006

Java *
C++
C
PHP
1. Perl
2. Python
3. C# *
4. JavaScript
5. Visual Basic
6. Delphi/Kylix
7. Unix Shell
8. Assembly
9. PL/SQL
  1. Tcl
  2. Objective C
  3. ASP
  4. Ruby
  5. VisualBasicDotNet *
  6. Pascal
  7. Lisp
  8. Object Pascal
  9. XSL (XsltLanguage/XpathLanguage/XslFormattingObjects) *
  10. JavaServerPages *
  11. Scheme

(Generated from http://sourceforge.net/softwaremap/trove_list.php?form_cat=160)''

SourceForge December 2007

TOTAL
Java
C++
C
PHP
1. Python *
2. Perl
3. C#
4. JavaScript
5. Unix Shell **
6. Delphi/Kylix *
7. Visual Basic
8. Assembly
9. PL/SQL
10. VisualBasicDotNet ****
11. JavaServerPages ********
12. Ruby ***
13. Objective C *
14. Tcl
  1. XSL (XsltLanguage/XpathLanguage/XslFormattingObjects) ****
  2. ASP.NET (AspDotNet)
  3. ActionScript ***NEW***
  4. ASP -
  5. Pascal
  6. LuaLanguage ***NEW***
  7. Object Pascal

These were the top languages used on Freshmeat in March 2003):

C
Perl
C++
Java
PHP
1. Python
2. Unix Shell
3. SQL
4. Ruby
5. C#

These were the top languages used on Freshmeat in December 2004 ( from http://freshmeat.net/browse/160/ ):


        Projects	Language

        --------	--------

C
Java **
C++ *
Perl
PHP
Python
1. Unix Shell
2. Tcl
3. SQL
4. JavaScript
5. Objective C
6. Other
7. Assembly
8. Ruby
9. C#
10. Other Scripting Engines
11. Scheme
  1. Lisp
  2. PL/SQL
  3. Delphi
  4. Fortran

This kind of list was used as evidence that CeePlusPlusReigns.

One should take into account some noise factors in the SourceForge data:

As it uses Sourceforge data, it is biased towards OpenSource projects and not necessarily indicative of popularity in proprietary systems.
It is up to the developers to report their languages. Many projects have none of this metadata.
Many projects report all languages used, of which one is primary and others are used for small amounts of glue. I don't think there were many projects that used primarily AssemblyLanguage, for instance.
Many projects report for which language it is used, rather than which language it was written in. (Example: most of the ForthLanguage projects were actually Forth compilers, not applications written in Forth.)
Many of the projects are VaporWare, MyToyProgram, or inactive for years (though this list has been culled since 2003).

Very interesting. I am astonished to see JavaScript slightly ahead of DelphiLanguage/Kylix, and more popular than all flavors of Lisp and all other flavors of Pascal combined.

-- DavidCary

In the end, this is only one data point. There are many other ProjectHosts, many language specific. Another measure (Yahoo directory links) is on ProgrammingLanguages.

And http://www.tiobe.com/tpci.htm has popularity trends over time. It uses Google for its data and appears to be aligned to the software job market, rather than the OpenSource community.

The "tpci" seems very unscientific as it uses keyword popularity as the indicator. There is no way to google for "C" without being completely overwhelmed by noise. It is also tainted by the fact that a language may have more "help pages" than a more popular language (but the language with more "help pages" would seem more popular).

None of the above have considered SpreadsheetModel as a ProgrammingLanguage. When I started to explore the MicrosoftExcelProgrammingLanguage, I found there is a huge secondary market related to education, consulting, third party tools. And the demand and pay scale for a good Excel consultant is much better than for the average website developer. I have made the wrong bet on traditional programming languages.

Moved from ProgrammingLanguages:

How to rate ProgrammingLanguagePopularity? Here is one way.

(Note, do not update this page, but use it to benchmark the current results.)

http://dir.groups.yahoo.com/dir/Computers___Internet/Programming_Languages

Rating on Sept 14,2004:


        ActiveX (41) 

        AppleScript (12) 

        Assembly (88) 

        Basic (164) 

        C# (201) 

        COBOL (91) 

        C and C++ (913) 

        Delphi (352) 

        Forth (14) 

        Fortran (11) 

        Java (1631) 

        JavaScript (178) 

        Lisp (21) 

        Object Oriented (219) 

        Pascal (69) 

        Perl (92) 

        Python (49) 

        RPG (146) 

        SQL (244) 

        Tcl Tk (20) 

        Visual Basic (774) 

        QBasic@ (52)


        Virtual Reality Modeling Language (VRML)@ (46)

VRML is not a programming language - at least not any more than HTML is.

DataAndCodeAreTheSameThing

Here is an attempt to gather data from several sources, while being transparent about the strengths and weaknesses of each one:

http://www.langpop.com

Top languages used in the GoogleCodeJam qualification round. So these are languages used in the small because people are familiar with them and they are suitable for quick problem solving (one day time limit to solve half a dozen problems).


        Lang	2008	2009	2010	2011	2012


        C++	10094	16288	15691	28108	25705

        Java	5537	6712	7684	11239	13778

        Python	2302	3342	4359	7199	9413

        C	1258	1325	1869	2484	3445

        C#	2062	2048	1843	2958	3177

        Ruby	393	595	623	1102	1217

        PHP	348	472	377	600	1048

        Perl	575	605	373	566	556

        Haskell	156	240	406	564	520

        Pascal	292	378	293	502	431

        VB	154	77	112	116	154

The subsequent rounds after most entrants are eliminated are dominated by C++ (75%), followed by Java and Python.

(From http://www.go-hero.net/jam)

See LanguagesOfChoice, PaulGraham's BeingPopularEssay, ProgrammingLanguageUsageStatistics, TheMostWidelyUsedProgrammingLanguageAtAnyLevel