Google releases GO, DART – Does Cloud Computing have anything to do with it?

Posted by Udayan Banerjee on April 27, 2012 · 5 Comments 

Do we need programming languages?

You may think that the answer is no. But, if you go by the recent trend you may need to change your mind.

• Why is Google working 2 (GO, DART) new programming languages?
• Why has IBM (X10), Cray (Chapel) and Red Hat (Ceylon) working on creating a new programming language of its own?
• Why have the attendees of the QCon London have selected 5 (HTML5, DART, Scala, Clojure, Node.js) new languages as the most important software development trends for 2012?
• What does Neil McAllister mean when he says that these 10 (DART, Ceylon, GO, F#, OPA, Fantom, Zimbu, X10, Haxe, Chapel) programming languages that could shake up IT?
• Why did Anders Hejlsberg of Microsoft, the creator of Turbo Pascal, Delphi and C#, release TypeScript?

Are new programming languages getting created just to satisfy the creative urge of somebody or has the technology evolution created the need for new way of programming and hence these new languages?

If you look back …

There are 3 periods in history when there was a burst of new programming languages. Each of these periods is linked to a critical point in technology evolution.

Period	Driver	Languages
1957-64	Invention of compiler and the era of 3GL	(1) Fortran, (2) Algol, (3) Lisp, (4) Cobol, (5) RPG, (6) APL, (7) Simula, (8) Basic and (9) PL/1
1978-84	Invention of RDBMS and the era of the 4GL	(1) SQL, (2) dBase, (3) C++, (4) Oracle Forms and PL/SQL, (5) Informix 4GL, (6) Gupta SQL, (7) Unify Accell and (8) Ingress
1990-95	O-O & Multi-tier programming and the era of WWW	(1) HTML, (2) Haskel, (3) Python, (4) Power Script (Power Builder), (5) Visual Basic, (6) Lua, (7) Ruby, (8) Object Pascal (Delphi), (9) Java, (10) Java Script and (11) PHP

Other important languages which was created outside these time period are: (1) Pascal – 1970, (2) C – 1972, (3) Prolog – 1972, (4) Smalltalk – 1972, (5) Erlang – 1986, (6) Perl – 1987 and (7) C# – 2001. As you can see, they are very few.

Are we entering another period which on hindsight will be classified as another period of technology transition? Only time can tell.

Does cloud computing got anything to do with it?

It most certainly does. But that is not the whole story – there is more to it.

What could the Technology Drivers be?

If you analyze the primary motivation behind these languages, you will see several common themes emerging.

(1) Distributed/Parallel computing:

Programmability of parallel computers (Chapel from Cray, X10 from IBM), Concurrent programming (Clojure, Fantom, Go) and single language for complete cloud stack (Opa)

Cloud computing is all about distributing your process across multiple CPU and running them in parallel or concurrently. Current programming languages are not very well suited for that.

Hence this attempt to create languages tailored for parallel processing.

(2) Multi-paradigm programming:

Languages which support object oriented programming as well as functional programming (Clojure, F#, Fantom, Scala)

Functional programs are relatively easy to parallelize. However, pure functional languages have not been very successful.

Hence this attempt to create multi-paradigm programming language.

(3) Multi-platform programming:

Languages that has implementation for both JVM and CLR (Clojure, Fantom, Scala) and several other platforms (Clojure & Fantom on JavaScript engine, Scala on Android, Haxe on several platforms)

Well this is a dream which we have been chasing for decades. Will it ever be a reality? We never know.

Hence this attempt to create multi-platform language!

(4) Programming at Scale:

Readability, Modularity (Ceylon), Speed of compilation (Go, Zimbu), Suitable for large team (Scala), Meta-programmability and extensibility (Ceylon, Scala), Concise (Scala, Zimbu)

Cloud provides almost unlimited computing power which encourages us to build larger and more complex application. Building such application will require large teams and the code has to be easily maintainable.

Also, the focus on frequent integration requires the process of compilation to be speeded up.

Hence this attempt to create programming language for large application.

(5) Client side programming:

Replacement for JavaScript (Dart), same language for both client and server side (Naxe, Opa, Zimbu)

Cloud computing requires client side code to seamlessly run on multiple client devices. Though JavaScript has become a standard, nobody really likes it!

Hence this attempt to create languages which is better than JavaScript and which can run on multiple client platforms.

Quick overview of the 12 new languages mentioned earlier

Here are the 12 languages listed in an alphabetic order. (I am not including HTML5 in this list as it is strictly not a new language and I have excluded Node.js as it is only a server side JavaScript engine)

Language	Who is behind it?	Primary Driver	Licensing
	Red Hat – Home page	Readability, Predictability, Tool-ability, Modularity, Meta-programmability.Runs on JVM	GPL v2
	Cray – Home page	Programmability of parallel computersRun on Cray supercomputers and various high-performance clusters.Portable to most Unix-style systems, Mac OS X and Windows	BSD
	Rich Hickey – Home page	Concurrency using Functional programming paradigm.Runs on JVM, CLR, and JavaScript engines	EPL
	Google – Home page	A replacement for JavaScript on the browserFaster, easier to maintain, and less susceptible to subtle bugs.Dart VM needs to be compiled – can run on Linux, Mac and Windows	New BSD
	Microsoft – Home page	Multi-paradigm: Functional + Imperative + Object-oriented.Runs on CLR and Mono10 Reasons Why It`s a Hot Programming Language for Developers	Apache
	Brian Frank, Andy Frank – Home page	Portability, support for functional programming and concurrency.Runs on JVM and CLR. Is also compiles to javaScript. Future targets might include Objective-C for the iPhone	Academic Free License
	Google – Home page	Compiled with the ease of programming of a dynamic language, concurrency and communication, speed of compilation.Compiler available for Linux, Mac OS X, Windows	BSD style + patent grant
	Nicolas Cannasse – Home page	Multi-platform support.Compiler for JavaScript, Flash, NekoVM, PHP, C++. C# and Java support is expected	GPL v2
	MLstate – Home page	Targeted for cloud computing. Client-side UI, server-side logic, and database I/O are all implemented in a single languageRuntime environment own Web server and DBMS.Runs on 64bit Linux and Mac	AGPL
	EPFL – Home page	Scalability for multicore and distributed computing. For large team. Multi-paradigm: Functional and O-O. Extensible.Runs on JVM, Android, CLR10 Reasons Developers Need to Check It Out	BSD
	IBM – Home page	Designed specifically for parallel programming, performance and scale.Runs on IBM AIX, Linux, Mac OS X, Windows	EPL
	Bram Moolenaar – Home page	Aims to be fast, concise, portable, and easy-to-read and support  GUI application to an OS kernel.Compiles to ANSI C	Apache v2

Finally…

What is most interesting is that all the 12 languages are available under some form of open source license.

Interesting articles on history of programming languages:

• The History of Programming Languages [Infographic]
• Programming Language History
• A Brief, Incomplete, and Mostly Wrong History of Programming Languages
• Modern books on multiple programming languages

Statistics on Language usage

• TIOBE Programming Community Index
• Programming Language Popularity
• Most Popular Programming Languages

Udayan Banerjee on Google+

Filed under Cloud Computing, Trend · Tagged with Ceylon, Chapel, Clojure, Cloud Computing, Dart, F#, Fantom, Go, Haxe, Opa, Scala, X10, Zimbu