Welcome to the Jungle (Redux)

Finally it seems that Ruby on Rails is starting to show its real face. Two years of RoR development put into the trash. Wow! Being seven times more productive than Java that means fourteen years of development! Seriously, it's not common to hear of failures of a new and cool software stuff -you can look stupid- so I guess this is the peak of the iceberg of Ruby On Rails projects. In my opinion, Ruby on Rails failed in this project because when dealing with the very small details of a project (that 5% of the project that takes 90% of the time) RoR convention over configuration approach fails. This, and of course integration with legacy systems (considering a database a legacy system). It's clear that any other general purpose language and framework (JEE, .NET) can solve the integration problem better and faster.

A few days ago I made a search in Infojobs.es (the most popular Job Board in Spain) for open positions for Ruby On Rails developers. The result was... one! I was very surprised. Only one? With all the hype around RoR? Only one job offer?
It's obvious that USA is more developed regarding technology than Spain, but the question is: How far are they? Can we measure the distance? Honestly I think it's very difficult and probably comparing two countries is not the right way. But anyway I did the following exercise. I selected a list of key words of software development, and I compared the number of open positions found in dice.com for USA and infojobs.es for Spain. Since absolute numbers are not comparable, I summed the number of open positions per key word and compared the percentages.
These are the key words and the number of open positions:

Keyword	dice.com	dice.com %	infojobs	infojobs %
JAVA	17098	32.87%	1682	32.41%
JAVASCRIPT	6437	12.38%	581	11.19%
J2EE	8304	15.96%	1077	20.75%
.NET	13269	25.51%	935	18.02%
PYTHON	1224	2.35%	49	0.94%
RUBY	602	1.16%	12	0.23%
PHP	2262	4.35%	459	8.84%
RUBY ON RAILS	257	0.49%	3	0.06%
ROR	34	0.07%	1	0.02%
STRUTS	2502	4.81%	391	7.53%
GRAILS	10	0.02%	0	0.00%
ERLANG	3	0.01%	0	0.00%
GROOVY	13	0.02%	0	0.00%
	52015	100.00%	5190	100.00%

I'm not going to talk about the absolute numbers, but the number of open positions in USA for that key words is ten times the number in Spain. Spain is 45 millions and USA is 300 millions. It means that there is an open position every 8000 Spaniards, and one open position every 5800 Americans.
I think we get the following stuff:

1. Java is about 32% in both countries. Does it mean that a mature technology is deployed everywhere in the same way?
2. Javascript 12%. Same thing.
3. J2EE is 21% in Spain and 16% in USA.
   
4. .NET 18% in Spain and 26% in USA. From points 3) and 4) we can conclude that Spaniards are smart guys ;-)
5. Python is not popular in Spain compared to USA. 0.95% vs 2.35%.
   
6. Spain doubles USA in PHP popularity: 8.84% vs 4.35%
7. Ruby, 1.16% USA vs 0.23% Spain. Ruby On Rails 0.49% USA vs 0.06% Spain. RoR 0.07% USA vs 0.02 Spain. Ruby is 5 to 6 times more popular in the States.
8. Struts popularity almost doubles in Spain: 7.53% vs. 4.81%
9. Grails, Erlang and Groovy are nonexistent in Spain, and almost insignificant in USA.

• Mature technologies seem to be deployed all around the globe equally.
• Python is not popular in Spain. Why? Can anybody explain it?
  
• PHP is popular in Spain. I guess PHP popularity is not declining yet because the hype of the new scripting languages has not arrived yet, plus the Python thing.
• Ruby and RoR have not caught Spanish developers yet. I guess the problem is the barrier of the language. It takes some time to translate manuals, tutorials and may be the community of Spaniards developers is not big enough to provide good support. Also, RoR is the language of choice of a lot of start ups companies because its productivity. Spain does not have a lot of 'start up' culture (same thing for Python?).
• Struts is still very popular. But in Spain is more dominant. Again, Struts is a mature technology compared to RoR, and RoR is not relevant in Spain yet.
  
• Grails, Erlang and Groovy are almost irrelevant on both sides of the Atlantic ocean. But it's a good idea to follow the trends.

I have just read this post in Engadget about rumours of the Google Phone. I will highlight this: '...the phone will run on a Linux variant (which is nothing new), and will be Java Virtual Machine-capable. Additionally, the OS of the phone will be Java-based (as well as the all phone apps itself), and performance is said to be "very responsive."'
It seems the GPhone will look like a Palm Treo or Blackberry with a screen smaller than the IPhone.
If these rumours are true, Google will become a big supporter of Java for interactive devices something the device market needs desperately to set free of the Windows Mobile and CF.NET slavery.

Evans data on his semi-annual Web Services Development Survey conducted in June says that companies planning the deployment of SOA with a Java platform has increased slightly, but the companies deploying on .NET have declined by almost 20%. .NET is still ahead with a 31% and Java in second place with 28%. 25% will support both.
It seems the reason is the increase of the activity of the Open Source world, specially Eclipse, around SOA.
Read the press release here.

When I read this post asking Microsoft to make .NET cross-platform, I wondered myself why they did not use Java -the one and only truly cross-platform environment-. Apart from the the obvious, Microsoft has very good reason to avoid making .NET cross-platform:

1. The jewel of the crown is Windows, so opening .NET to OS X and Linux is like shooting on your feet.
2. As Steve Ballmer said he loves 'developers'... because they build 'applications'...for the Windows OS. Fuck the Linux communists and the OS X snobs! If you get the same functionality on a free OS, why are you going to pay for Windows?
3. Performance and cross-platform do not match. The good performance of .NET comes from its architecture, using the Windows OS APIs when dealing with I/O and display. It took years to Java to improve his performance, and now they are taking an approach similar to .NET.
4. Most of the people think that .NET is a rip off of Java. May be if .NET does not take the cross-platform way to avoid legal IP problems with SUN... who knows!
5. Microsoft does not care about .NET. The real sources of incomes are Windows and Office licenses. .NET is the key lock to keep developers in the Windows platform. If the key lock is opened, then they should throw away the key lock and buy a new one.
6. Making .NET cross-platform is expensive, as SUN knows. If .NET does not give them direct revenue, there is no reason to invest on it. The business model is nonexistent for Microsoft.