Comments on: Qi4J the next Java? Forget Scala

By: Mert Nuhoglu

Mert Nuhoglu — Mon, 06 Jun 2011 11:00:16 +0000

Hi Stephan,

Inspiring article. Thanks for sharing with us. You told that you would write about other goodies in Qi4J. We are waiting for it :)

By: rod

rod — Tue, 14 Dec 2010 10:58:44 +0000

Could you elaborate on your assertion that scala is rubish ?
I think scala provides many basic concepts that should highly ease the implementation of a DDD framework. There would probably be many benefits to write Qi4J in scala.

By: Qi4j a new way of software development « Dotsystems

Qi4j a new way of software development « Dotsystems — Wed, 16 Dec 2009 15:31:24 +0000

[...] including evolved concepts from AOP, DI and DDD“. The man behind the project is Rickard Öberg, well known for development of JBoss, Webwork and XDoclet. It is 100% pure Java implementation of [...]

By: Vick Hobster

Vick Hobster — Thu, 19 Feb 2009 04:28:27 +0000

Good stuff, looking forward to sticking around.

By: Jake Hierson

Jake Hierson — Thu, 19 Feb 2009 04:27:35 +0000

amazing stuff thanx :)

By: video hentai mama

video hentai mama — Thu, 04 Sep 2008 22:11:57 +0000

mama keroro hentai mama hentai kazama

By: Yuen-Chi Lian

Yuen-Chi Lian — Sat, 22 Mar 2008 03:28:18 +0000

> I spoke to several of the Qi4j-members at the announcement at ÖreDev, and unfortunately I think they’re
> trying to implement the idea in the wrong language.

Similar argument we had in a JUG meeting with the presence of Rickard. Some people think that Qi4J is merely a hack and why Java at the first place?

Qi4J is just an implementation of COP and you can implement it with any language. Java is a stronger platform for enterprise development in comparison with other languages and it therefore worth investing. I’m sure a lot of the enterprise folks will welcome this.

- yc

By: stephan

stephan — Fri, 11 Jan 2008 17:44:39 +0000

Regarding “They implemented Qi4J in the wrong language”. Qi4J is an idea and can be implemented in any language. Qi4R? Qi4P? Qi4S? My only complaint is, they should try to implement this on top of Guice instead of doing their own DI. Sure that would mean compromises but increase adoption. Evolution instead of revolution will win the heart of the enterprise.

By: Viktor

Viktor — Fri, 11 Jan 2008 10:32:50 +0000

Lots of opinions, less facts.
I spoke to several of the Qi4j-members at the announcement at ÖreDev, and unfortunately I think they’re trying to implement the idea in the wrong language.

By: stephan

stephan — Thu, 10 Jan 2008 22:24:33 +0000

@Doug: The brain you know, compared to say Marc. Before it went down. Well intellectually at least. The money went up. But isn’t that always the case? Brains out, money in? Or is it the other way round?

By: Doug

Doug — Thu, 10 Jan 2008 18:44:08 +0000

You had me until “JBoss”

By: stephan

stephan — Thu, 10 Jan 2008 16:21:07 +0000

I think they go into a similar direction but differ form Qi4j.

By: Robert

Robert — Thu, 10 Jan 2008 12:53:35 +0000

Just for completeness there is:

Naked Objects
JMaki

Those are both in the same category as Qi4j.

By: stephan

stephan — Wed, 09 Jan 2008 14:36:57 +0000

I wrote about a @Getter annotation

http://stephan.reposita.org/archives/2007/11/24/getter-annotation/

before. get/set should no longer be needed for Java. Wasn’t it just so that GUI builders could detect attributes by looking at the method name? Annotations should solve that.

What I especially like about the Qi4J examples is the @NoEmpyString annotation in the interface instead of the class definition. It’s more like a contract that way.

By: stephan

stephan — Wed, 09 Jan 2008 14:34:25 +0000

Thought about the property when reading on your blog and the examples.

Setting the property does look like this then?

person.name().set(“Stephan”)?

Not sure if that’s a good idea.

person.setName(“Stephan”) seems shorter, so I’d like to have that in the interface but my PersonBean could have a Property name() which is automatically mapped from the interface by Qi4j.

person.name!(“Stephan”) could be done too.

Attributes in interfaces are final, but could they be declared?

public interface Name {
public String name;
}

and Qi4j maps a person.name=”Stephan” call to a setName() in the bean?

Good look with solving the decade old Java property problem.

By: Rickard

Rickard — Wed, 09 Jan 2008 14:13:32 +0000

Thanks for the kind words! The only thing I have to add really is that we are currently looking into explicit support for properties, which means that declaring the Name interface will look something like this instead:
public interface Name
{
Property name();
}

As you wrote, the implementation will be provided automatically, so this is all you have to do. This will allow you to work with properties much better, get meta-information about them easily (example: “person.name().getPropertyInfo(DisplayInfo.class).getLabel()”), do UI binding, etc.

It’s all work in progress, but I agree that I think this has good potential for solving a bunch of the problems with current software development in Java. Time will tell!