Research firm Evans Data recently did a survey to over 400 developers in Europe, Middle East and Africa (EMEA), and found some interesting results.

The number of developers using PHP for development dropped by more than 25% in the last year and the number of developers indicating they would not evaluate or use PHP for future development projects grew by almost 40% in the same time period. Perl usage in EMEA has dropped by more than 20% and those developers with no intentions to evaluate or use Perl grew by 20%. Python usage also saw a 25% reduction in current usage and developers without any intention to use or evaluate Python grew by 17% in the last year.

"PHP, Perl and Python use on a global basis peaked one to two years ago and has started to decline based on a number of factors. This decline is more exaggerated in EMEA and APAC than in North America," said John Andrews, Evans Data's Chief Operating Officer. "One of the key factors to this loss of developer mindshare has been the inability of these languages to penetrate the enterprise space."

Another interesting point was that three out of five developers, 61%, in EMEA have made use of open source software modules in their development but only a third, 33%, have contributed back to the open source community. Roughly the same number of developers (60%) in other parts of the world also use open source.

If less people are using LAMP (Linux Apache MySql PHP/Perl), then what are they migrating to? LAMP people won't easily switch to .NET, leaving Java as the only other major platform.

It's obvious everyone is switching to ROR. Let the flame war begin :)

I'll apologize in advance for the bad joke.

peter

...so that we could answer the question...
could be usefull to get something specific for .Net too.

side note : scripting languages could re-gain in popularity when their integration with Java will happen.

Laurent.

...so that we could answer the question...could be usefull to get something specific for .Net too.side note : scripting languages could re-gain in popularity when their integration with Java will happen.Laurent.

PHP does have java integration.
But the thing is if you want to use php with java - why not use jsp with java.

PHP does have java integration. But the thing is if you want to use php with java - why not use jsp with java.

There are lots of PHP applications out there.

You may want to extend an existing PHP application with some functionality implemented in Java. In this case, calling Java logic from PHP application would require less effort than converting PHP pages to JSP.

--
Igor Zavialov, Factoreal Corp.
Financial Data and Technical Analysis solutions.

.Net is pretty much entranched in MS Windows, while the
"LAM" in "LAMP" stand for "Linux/Apache/MySQL", three
technologies which pretty much compete directly with
Microsoft's stack.

There is project Mono and its implementation of C# but as
was demonstrated a few days ago, Microsoft could shake its
chain any time to prevent it from being a real competition
to its own platform.

And as others indicated, though the Java specification itself
is not as open as FOSS people like, it's much easier to implement on top of non-MS platforms than .Net (as an
understatement).

Evans Data is letting itself down. How can they publish a 230 page report on the responses of a tiny sampling?

EMEA is a sales territory that has no correlation with the subject matter (developer preferences/activities). A sampling that includes developers from (rich) Northern Europe and the poorest parts of Asia and Africa is really going to tell you nothing at all.

The sampling method is so poor that even if they published this on an annual basis it would not be valid for trend analysis (a classic fallback when the survey is too small, not demographically correct, etc)


Evans used to publish solid market analysis, what happened?



PJ Murray

CodeFutures Software

Code Generation for Java Persistence

Excuse me, but why not? What would make it easier for them to move to Java?

Excuse me, but why not? What would make it easier for them to move to Java?

"LAMP" not only specifies a set of technologies, it also usually indicates an Open Source mindset, which can mean a resistance to Java.

not sure how an "open source" mindset can lead away from java since there are complete solutions that compete with RoR + LAMP that are Open Source in Java

Struts / Spring / Hibernate / Tomcat / JBoss / HQL ... etc
(and yes this solution can compete right with RoR)

I think a factor is what was mentioned above, is very minimal use of php in the enterprise (which actually produces the jobs), which means php programmers make less than java programmers because they are in demand less....

not slamming php or any language, but Its a matter of no one is behind it like Java. (Sun,Bea,Oracle,IBM) Those companies sell java to Enterprise...

not sure how an "open source" mindset can lead away from java since there are complete solutions that compete with RoR + LAMP that are Open Source

It is because Java itself is not available open source.

I am not against java being open source, but not sure how many people actually need to change the JVM code.... If you did I'd suspect that most of the time... what was being done was too low level to be solving business problems and more for the fun of it. (yes there may be a few rare exceptions)...

not sure how an "open source" mindset can lead away from java since there are complete solutions that compete with RoR + LAMP that are Open Source

It is because Java itself is not available open source.

Just because the language implementation itself is not open sourced is irrelevant for all but a few zealots.

What matters is that Java runs on most any platform out there -- and this goes for all of it, not just the non-GUI parts (e.g. in Mono) and has many valuable open source offerings based on and usable from it.

Just because the language implementation itself is not open sourced is irrelevant for all but a few zealots.

I mostly agree; I was simply trying to explain why LAMP supporters (not being one of them myself) might not consider Java.

Excuse me, but why not? What would make it easier for them to move to Java?

Hey, where's Rolf? On vacation? This is perfect troll food!

Jens

The document states that PHP Dominates in the Windows segment... really?

http://news.com.com/Java+devotee+BEA+eyes+scripting+languages/2100-1012_3-5820062.html?tag=cd.lede

.V

http://news.com.com/Java+devotee+BEA+eyes+scripting+languages/2100-1012_3-5820062.html?tag=cd.lede.V

I don't think it is FUD, but I think it is poor statistics. A sample of 400 distributed between Europe, the Middle East and Africa is very small, and, having some personal experience in statistical methods, I would not take this report seriously. If 'over 400' was several tens of thousands, I might have a different view.

http://news.com.com/Java+devotee+BEA+eyes+scripting+languages/2100-1012_3-5820062.html?tag=cd.lede.V

I don't think it is FUD, but I think it is poor statistics. A sample of 400 distributed between Europe, the Middle East and Africa is very small, and, having some personal experience in statistical methods, I would not take this report seriously. If 'over 400' was several tens of thousands, I might have a different view.

Hmmmm, most people who have "personal experience in statistical methods" know that the sample size doesn't need to be anywhere near "tens of thousands" to be statistically significant. The key is really getting a true random sample that is representative of the overall population you're interested in. With that said, you may still be correct about the poor statistics - it's hard to know without knowing the specifics of how they chose the 400.

This is true. Assuming a normal distribution (which is often a dangerous assumption), the rule of thumb is a sample size of 30. The reason that they choose as many as 400 is to allow splicing and dicing of the population, hence a 200+ page report. Still their random sampling itself is questionable.

Hmmmm, most people who have "personal experience in statistical methods" know that the sample size doesn't need to be anywhere near "tens of thousands" to be statistically significant. The key is really getting a true random sample that is representative of the overall population you're interested in. With that said, you may still be correct about the poor statistics - it's hard to know without knowing the specifics of how they chose the 400.

That is what I meant. As someone who has "personal experience in statistical methods" I realise that with a true normal distribution, a sample size of 400 is very good indeed, but there is no way this is a normal distribution. IT markets and practices vary widely over Europe alone, let alone the Middle East and Africa.

I agree, a sample size of 400 for Europe, Middle East and Africa is very small, too small to indicate an accurate trend.

For example, where did they get this set of programmers? If they did it at a Java conference, then the results are biased.

Is there more details on this survey?

Do those developers migrate to Java/.NET? From 400 developers, the result means nothing. LAMP developer declining that is because more and more fool people claim that they are developer. With nice IDE, Java and .NET are fool-prove programming language, but not bullet-prove. PHP, PERL and Python are short of those tools. I personally love PHP better than Java, even I am making living on Java/.NET now.

I have no problems beliving the usage of Perl is dropping.

My impression is that the usage of Python is so small that
it is hard to see any trend.

But I can not see PHP in decline.

I see more and more hobby PHP programmers.

I see more and more companies using PHP CMS's/forums/communities.

And why not ? If you want something ligthweigth then PHP
is an excellent choice !

(just like VB6 may be more suited for some tasks than
C++ with MFC)

Every new release of PHPBB usually contains a strong warning about some major vulnerability that was fixed and if don't update, there is a risk of a script kiddy overwriting all of your web files. The forum software's usability is top notch, but these vulnerabilities are too scary.

Every new release of PHPBB usually contains a strong warning about some major vulnerability that was fixed and if don't update, there is a risk of a script kiddy overwriting all of your web files. The forum software's usability is top notch, but these vulnerabilities are too scary.

Whats your point?

This is different from most commodity software, how?

Regardless of what Evans Data claims, the statistical sample is biased. It consists only of people (perhaps not even in technology) who responded to Evans Data spam for an opportunity to win $500 or a $1000 for taking surveys. That is not random sampling of the population. We should not leave out of the survey all technology people that do not respond to spam. They may have a very different opinion of Perl and Python. At best, we can say that the number of unverified technology respondents that answer spam e-mails for a chance to win money have said that they are no longer considering the use of Perl.

Regardless of what Evans Data claims, the statistical sample is biased. It consists only of people (perhaps not even in technology) who responded to Evans Data spam for an opportunity to win $500 or a $1000 for taking surveys. That is not random sampling of the population. We should not leave out of the survey all technology people that do not respond to spam. They may have a very different opinion of Perl and Python. At best, we can say that the number of unverified technology respondents that answer spam e-mails for a chance to win money have said that they are no longer considering the use of Perl.

Yeah, it seems pretty worthless data on the face of it. I also can't see what figues they are using as a point of reference, in order for them to extract a trend that can say usage of something is actually "dropping". The corresponding 2004 report doesnt seem to be readily available.

In my recent interview to IndicThreads about comparison of PHP and Java CMS I said the following phrase "Java is not a panacea. In order to dig a small hole it is easier and quicker to use small tool - spade. If you need to dig a trench you are forced to use bulldozer."

Also this interview covers the position of Russia in BRIC (Brazil, Russia, India and China) as a most IT-active part of EMEA (Europe, Middle East and Africa).

Do you really think that BRIC countries are part of EMEA? HA HA... spend some time on geography!

Brasil is in South America (not part of EMEA)
China is in Eastern Asia (not part of EMEA)
Russia is both Asian and European so it probably may be considered part of EMEA
India is in Asia (not part of EMEA)

ROTFL!

+1

Maybe Hani Suleiman could write about these *very interesting* interviews...

I'm flaming, I know.

As stated before, this seems like just bad statistics and strange interpretation. Perhaps three years ago I first saw an article about java dying, still hasn't happened. Currently there just isn't substitute for php; plain jsp provides very little extra compared to php.

I don't know if this survey is reliable or not but you can easily see how languages like PHP are looking more and more like Java. Just look at the evolution path of PHP and now, at version 5 it's object oriented with pretty much all the features that Java has. It still does not have the overhead that Java sometimes has, but it is no longer significantly easier to pick up or learn. Throw in some backwards compatibility doodads and you might be puzzled.

Then look at some of the PHP projects. They are using templating engines, scripting (vs. JSTL expression language), and all the rest. It just starts to look and feel so much like any typical Java web app. Hmmm, there might be some sense to move to Java in these cases.

The Evans report makes no mention of Java, though we're left to assume that teams abandoning scripting languages are switching to Java. If the industry is focusing on Java for web apps, that would make Java a modern COBOL, a commodity skill with a commodity wage. Though unlike COBOL, the Java language and virtual machine specifications begin with warnings of Sun's patents.

Another interesting point was that three out of five developers, 61%, in EMEA have made use of open source software modules in their development but only a third, 33%, have contributed back to the open source community.

"Only" a third? I'm shocked it's as high as a third! I think that's outstanding news, if it's correct!

Peace,

Cameron Purdy
Tangosol Coherence: Clustered Shared Memory for Java

Not that shocking since the study defines contribution as "contribute code or suggestions." If a developer just gave one suggestion or comment to a news group over the course of their life, they are considered a contributor to the community. A more accurate question would be whether they have contributed code over the last three years (or even decade). I am sure that number would be more within expectations.

Not that shocking since the study defines contribution as "contribute code or suggestions." If a developer just gave one suggestion or comment to a news group over the course of their life, they are considered a contributor to the community.

That stretches the meaning of 'contributor'. I agree that someone who submits a bug ticket is contributing as a tester. But someone who only asks for help is not a contributor.

The Java Vs PHP results of the survey seem at odds with what is happening 'on the Ground'.

In Dublin , Ireland there is increasing adoption of PHP , often with some sort of 'Enterprise Back End' such as Oracle. This is driven by the facts that it is easier to get PHP people than Java , and at a much lower cost. In this scenario , any time wasted by the lack of 'PHP Enterprise Edition' can be made up with the cost savings on the project.

Maybe the survey reflects a Start in PHP , Move into Java / J2EE intention on the part of the respondants?

(Before anybody flames this , I'm a J2EE consultant , but have been really impressed by the speed of development of PHP and the quality of open source PHP like http://www.sugarcrm.com/ )

Paul Browne

http://www.firstpartners.net/rp

Python seems to be gaining ground in the game industry, both for scripting events etc. and for the actual implementation. AFAIK EVE Online and Disney's ToonTown Online, both multiplayer games, are partly written in Python.

What I want to know what was the point in reinventing the wheel by making Groovy.

Apart from mobile phone and "parlor" games, I seem to recall only Red Storm using Java to write games (the Republika semi-boardgame and some other title at least).

Apart from mobile phone and "parlor" games, I seem to recall only Red Storm using Java to write games (the Republika semi-boardgame and some other title at least).

There is a lot of game programming done in Java, but it is only applicable for certain types of games. For example, several of the online MMPG engines are Java. (The reason that I know this is that we are working on releasing a customer story about a certain MMPG engine done in Java and clustered with Coherence.)

Peace,

Cameron Purdy
Tangosol Coherence: Clustered Shared Memory for Java

This article disappointed me actually. I am in a middle of structuring My Company, Egyscape. In fact, we are doing Business Websites targeting small-medium companies in the Middle-east. The Funny thing is, I am a JAVA developer, Coldfusion Developer and I was thinking to base our company on PHP. So I was moving on with PHP as my website scripting language. So this Article made me confused me big time. And I wonder if anyone can help out. Well, I can't choose JSP because it takes a lot of time to develop from my own experience. Coldfusion is the best for the web but expensive too. So, I am left with PHP cheap, powerful and almost available on every Hosting. So my clients won't bother about the cost of hosting. I am trying to think as a business oriented here rather than an Object oriented :).

And you know what, my JAVA knowledge makes my scripting language so TASTY, because simply I can switch any web-application to a powerful Desktop application with my sweet heart JAVA. :) I am anti MS technology by the way.

So if you guys can give me your opinions about moving forward with PHP.
Is it the right decision or ..?.

Mo