Perl is dead. PHP sucks. Long live Python!

[Alex Libman 15]

[Alex Libman 15]

Man, TIOBE Index [WP] not updated for October yet.  There ought to be a law... 

In other news, milestone perl 5.11.0 released yesterday, while perl 6 continues to be a joke.  Python is on 2.6.3 and 3.1.1 (bug-fixes). 

[Alex Libman 15]

OK, the TIOBE Index was finally updated for October.

PHP continues to grow strong, while Perl, Python, and Ruby continue to yield a little.  Also in decline are business app kings Java and VB.  Plain C is holding strong, but Objective-C continues to gradually replace C++ as its OO alternative.

[AL the Inconspicuous]

OK, the TIOBE Index was finally updated for November.

C is about to surpass Java (and that's excluding C++, Objective-C, etc).

PHP remains the scripting champ.

Another language that's doing well is C#.

[AL the Inconspicuous]

Moved from another thread:

Yeah, I'm watching the Clang / LLVM projects and the attempts to integrate them into BSD with great anticipation, but it will probably take a couple of years before all ports can be reliably compiled and the GNU compiler / toolchain can be abandoned completely.

By the way, it's starting to look like Python will be the major scripting language of a GNU-free UNIX OS.  Once again, the corporate hero there is Google with its unladen-swallow project (which also uses LLVM).  Python is also the easiest way to script libtorrent / python-ogg, because all major front-ends for those BSD libraries are either copyleft themselves or have copyleft dependencies.

I'm a big fan of Ousterhout's dichotomy - a language should either be system-oriented or scripting oriented.  Unfortunately there's also a dichotomy between generations: the current (C for systems, Python for scripting) and Google's vision for the future (Go for systems, Web client / server-side JavaScript for scripting) - so that's 4 languages to be concerned with.  Everything else (Java, Obj-C / C++, perl, ruby, PHP, lua, tcl, bash scripts, etc) should be phased out, the sooner the better.  The older generation will be around for a long time, until JavaScript gradually gets more libraries than Python / PHP, and all systems software (including the kernels) are rewritten in Go.  Then - singularity! 

I wouldn't be disappointed with C and Python ruling the world together. 

Not gonna happen.

• C has been the God of systems programming for so long that no one programs in assembly any more, which is quite a feat and has done great things for program portability, and no one can even conceive that serious systems software could have been written in something like Fortran instead.  C has also outlived many scripting languages that were supposed to complement it, the most memorable one being Lisp.  Unfortunately it has failed to evolve intelligently over the years (C++ sucks!), which is why we saw the rise of so many neither-town-nor-country languages like Pascal, Java, and C#.  It can definitely incorporate a lot more syntax sugar without slowing it down, and for modern systems good built-in parallelism support is a must.  Improving on C would immediately make all systems software rewritten in the new language more secure and scalable by default, and programmer productivity savings would further translate into more time being spent on testing and enhancement.  That's why I've been following the DMD and similar efforts with great excitement over the years, only to see all of them stagnate.  Now Google's Go offers a new hope for an eventual C-killer, and I can't think of any reason why I wouldn't want it to eventually succeed.

• JavaScript / ECMAScript has been THE dynamic client-side language standard of the Web since its beginnings, and there are many very good engine implementations from different teams, with speed competition only starting to heat up:  Google's V8 is currently the fastest, Apple's JavaScriptCore, Adobe's ActionScript used in Flash, Mozilla's SpiderMonkey, Opera's Presto, Nokia's QtScript, etc - and Microsoft is once again throwing its money into R&D for IE9.  And we're not just talking about the scripting language itself here, we're talking about the whole layout engine / Web paradigm (established HTML 5 / CSS / AJAX) as well as new paradigms in browser add-ons (ex. Gears), desktop widgets, standard libraries for multi-language VM's (Java / CLI / someday Parrot [2]), and so on.  Yes, it's not thought of as a stand-alone language yet, but with Google's help it could rise in that area even faster than Ruby did, because writing library bindings for yet another language isn't that hard.  Plus as much as I love Python's innovative use of whitespace, most people do not, and it does make sense that the dominant scripting language and the dominant systems language should share more syntactical habits.

[Sam Gunn (since nobody got Admiral Naismith)]

Moved from another thread:

Yeah, I'm watching the Clang / LLVM projects and the attempts to integrate them into BSD with great anticipation, but it will probably take a couple of years before all ports can be reliably compiled and the GNU compiler / toolchain can be abandoned completely.

By the way, it's starting to look like Python will be the major scripting language of a GNU-free UNIX OS.  Once again, the corporate hero there is Google with its unladen-swallow project (which also uses LLVM).  Python is also the easiest way to script libtorrent / python-ogg, because all major front-ends for those BSD libraries are either copyleft themselves or have copyleft dependencies.

I'm a big fan of Ousterhout's dichotomy - a language should either be system-oriented or scripting oriented.  Unfortunately there's also a dichotomy between generations: the current (C for systems, Python for scripting) and Google's vision for the future (Go for systems, Web client / server-side JavaScript for scripting) - so that's 4 languages to be concerned with.  Everything else (Java, Obj-C / C++, perl, ruby, PHP, lua, tcl, bash scripts, etc) should be phased out, the sooner the better.  The older generation will be around for a long time, until JavaScript gradually gets more libraries than Python / PHP, and all systems software (including the kernels) are rewritten in Go.  Then - singularity! 

I wouldn't be disappointed with C and Python ruling the world together. 

Not gonna happen.

• C has been the God of systems programming for so long that no one programs in assembly any more, which is quite a feat and has done great things for program portability, and no one can even conceive that serious systems software could have been written in something like Fortran instead.  C has also outlived many scripting languages that were supposed to complement it, the most memorable one being Lisp.  Unfortunately it has failed to evolve intelligently over the years (C++ sucks!), which is why we saw the rise of so many neither-town-nor-country languages like Pascal, Java, and C#.  It can definitely incorporate a lot more syntax sugar without slowing it down, and for modern systems good built-in parallelism support is a must.  Improving on C would immediately make all systems software rewritten in the new language more secure and scalable by default, and programmer productivity savings would further translate into more time being spent on testing and enhancement.  That's why I've been following the DMD and similar efforts with great excitement over the years, only to see all of them stagnate.  Now Google's Go offers a new hope for an eventual C-killer, and I can't think of any reason why I wouldn't want it to eventually succeed.

• JavaScript / ECMAScript has been THE dynamic client-side language standard of the Web since its beginnings, and there are many very good engine implementations from different teams, with speed competition only starting to heat up:  Google's V8 is currently the fastest, Apple's JavaScriptCore, Adobe's ActionScript used in Flash, Mozilla's SpiderMonkey, Opera's Presto, Nokia's QtScript, etc - and Microsoft is once again throwing its money into R&D for IE9.  And we're not just talking about the scripting language itself here, we're talking about the whole layout engine / Web paradigm (established HTML 5 / CSS / AJAX) as well as new paradigms in browser add-ons (ex. Gears), desktop widgets, standard libraries for multi-language VM's (Java / CLI / someday Parrot [2]), and so on.  Yes, it's not thought of as a stand-alone language yet, but with Google's help it could rise in that area even faster than Ruby did, because writing library bindings for yet another language isn't that hard.  Plus as much as I love Python's innovative use of whitespace, most people do not, and it does make sense that the dominant scripting language and the dominant systems language should share more syntactical habits.

C STILL KILLS JAVA.  FUCK JAVA.

[AL the Inconspicuous]

C STILL KILLS JAVA.  FUCK JAVA.

Yeah, I think we all agree on that one.  I am particularly disappointed with Java's encroachment into the OpenOffice.org project, its dominance with advanced FLOSS IDE's, and its resistance to reform (ex. Groovy had potential but never got off the ground).  The middle-of-the-road languages / systems are fail on both sides, and Java even managed to fall behind .NET in many aspects as well!

[AL the Inconspicuous]

An update to the TIOBE Programming Community Index for December 2009 --

December Headline: All time highs for C#, Objective-C and ActionScript

This month company-related programming languages are in the spotlight. Microsoft's C# language reached its highest ranking ever. Chances are high C# will become TIOBE's programming language of the year 2009. Another language that set a personal record was Apple's Objective-C. It is getting closer to the top 10 every month and has now officially reached status "A". But also Adobe's ActionScript had a high score. It re-entered the top 20. Other breaking news is that software giant Google released its first own programming language last month. The language is called Go. It will be monitored in the TIOBE index as of next month.

Other observations from looking on that index is that Java, Perl, and C++ continue to decline, while Python has regained some territory since last month.

[NHArticleTen]

http://www.linuxjournal.com/article/3882

.

[AL the Inconspicuous]

Monthly update from TIOBE Software -- March Headline: Objective-C's popularity still rising, Go falling behind --

An interesting change can be observed this month. Objective-C scores another all time high, whereas Go drops below 1% for the first time. None of the promising languages mentioned last month made it into the top 20. Good old Fortran re-entered the chart instead, replacing Lisp/Scheme. MATLAB also made an interesting jump. It is now close to becoming a status "A" programming language.

(In case you missed it, Google Go was the "TIOBE Programming Language of 2009".)


Python also continues declining...  I guess that means... long live PHP!  Speaking of which, Facebook is doing some great things with PHP compilation and optimization.

[Level 20 Anklebiter]

Stackless is still a viable derivative to CPython. In fact, I'm considering its use in my graduate project (over Haskell or Ocaml).

[AL the Inconspicuous]

I occasionally go through brief phases where I fall in love with a language like Haskell and consider marrying it and living on an obscure Haskell-based island somewhere, but fortunately I always quickly come to my senses and let the market dictate what I code. 

[Level 20 Anklebiter]

I occasionally go through brief phases where I fall in love with a language like Haskell and consider marrying it and living on an obscure Haskell-based island somewhere, but fortunately I always quickly come to my senses and let the market dictate what I code. 

Functional languages have a niche in the market, particularly languages like Erlang and Lisp. Those two pretty much are the defacto functional language paradigm. Haskell and O/caml are just upstarts with nicer syntax.

[AL the Inconspicuous]

I don't like functional languages.  I understand the concurrency and other benefits, but I still don't like em, not even SQL.  I want to tell the computer exactly what to do.  If I want speed I code in C and do all the threading work myself.

[Level 20 Anklebiter]

Well it's the future, but I suspect that functional languages will be made more friendly and given more multi-paradigm goodness like how Python and other languages have (even C++ has concurrency via BOOST's ASIO I believe).