Steven Devijver

supporting Java code for writing build scripts come with a lot of technical hurdles

What are those hurdles? I don't see any technical problem.

Before running the Java code it needs to be compiled. And after compiling it it needs to be loaded by a ClassLoader. Only then can the Java code actually be executed. This brings up all kinds of questions, like which class name and package name to use. Also, this Java code won't be compiled before the build system runs which is not what Java developers are used to.

This needs to be looked in to. It probably works technically - at least on paper - but if it becomes too weird and cumbersome for developers to write Java build code it may not be worth the while.

supporting Java code for writing build scripts come with a lot of technical hurdles

What are those hurdles? I don't see any technical problem.

Before running the Java code it needs to be compiled. And after compiling it it needs to be loaded by a ClassLoader. Only then can the Java code actually be executed. This brings up all kinds of questions, like which class name and package name to use. Also, this Java code won't be compiled before the build system runs which is not what Java developers are used to.

This needs to be looked in to. It probably works technically - at least on paper - but if it becomes too weird and cumbersome for developers to write Java build code it may not be worth the while.

supporting Java code for writing build scripts come with a lot of technical hurdles

What are those hurdles? I don't see any technical problem.

Before running the Java code it needs to be compiled. And after compiling it it needs to be loaded by a ClassLoader. Only then can the Java code actually be executed. This brings up all kinds of questions, like which class name and package name to use. Also, this Java code won't be compiled before the build system runs which is not what Java developers are used to.

This needs to be looked in to. It probably works technically - at least on paper - but if it becomes too weird and cumbersome for developers to write Java build code it may not be worth the while.

supporting Java code for writing build scripts come with a lot of technical hurdles

What are those hurdles? I don't see any technical problem.

Before running the Java code it needs to be compiled. And after compiling it it needs to be loaded by a ClassLoader. Only then can the Java code actually be executed. This brings up all kinds of questions, like which class name and package name to use. Also, this Java code won't be compiled before the build system runs which is not what Java developers are used to.

This needs to be looked in to. It probably works technically - at least on paper - but if it becomes too weird and cumbersome for developers to write Java build code it may not be worth the while.

The main problem I see with build systems is that actually you need one. People were very happy with Visual Basic and Deplhi because they let programmers draw a screen, press a button and see it working. Why can't we do that with a web application? Because we want to write operations in Java, therefore recompile, then redeploy, so that instad of "a = b + c;" now it says "a = b - c;". That is a waste of time. The UI, the business rules and the database should be configured, not programmed. This way, as in Smalltalk, the application would be always running, you would just change the behavior on the fly by changing the UI, the business rules and the database, but not a single line of code should change.

There's a lot of room for innovation.

The main problem I see with build systems is that actually you need one. People were very happy with Visual Basic and Deplhi because they let programmers draw a screen, press a button and see it working. Why can't we do that with a web application? Because we want to write operations in Java, therefore recompile, then redeploy, so that instad of "a = b + c;" now it says "a = b - c;". That is a waste of time. The UI, the business rules and the database should be configured, not programmed. This way, as in Smalltalk, the application would be always running, you would just change the behavior on the fly by changing the UI, the business rules and the database, but not a single line of code should change.

There's a lot of room for innovation.

The main problem I see with build systems is that actually you need one. People were very happy with Visual Basic and Deplhi because they let programmers draw a screen, press a button and see it working. Why can't we do that with a web application? Because we want to write operations in Java, therefore recompile, then redeploy, so that instad of "a = b + c;" now it says "a = b - c;". That is a waste of time. The UI, the business rules and the database should be configured, not programmed. This way, as in Smalltalk, the application would be always running, you would just change the behavior on the fly by changing the UI, the business rules and the database, but not a single line of code should change.

There's a lot of room for innovation.

Please consider using Java for your future build systems. As others have mentioned you can use a declarative syntax in any language, including Java. By letting developers specify a build file using normal Java code you leverage the existing Java tools and lower the learning barrier to entry. The key is to spend a lot of time in the design phase, to ensure the API is very clean and intuitive.

That's certainly something we're considering, although supporting Java code for writing build scripts come with a lot of technical hurdles. I think it can be done technical. Remains to be seen if writing this Java code will be easy and straightforward enough.

Please consider using Java for your future build systems. As others have mentioned you can use a declarative syntax in any language, including Java. By letting developers specify a build file using normal Java code you leverage the existing Java tools and lower the learning barrier to entry. The key is to spend a lot of time in the design phase, to ensure the API is very clean and intuitive.

That's certainly something we're considering, although supporting Java code for writing build scripts come with a lot of technical hurdles. I think it can be done technical. Remains to be seen if writing this Java code will be easy and straightforward enough.

Please consider using Java for your future build systems. As others have mentioned you can use a declarative syntax in any language, including Java. By letting developers specify a build file using normal Java code you leverage the existing Java tools and lower the learning barrier to entry. The key is to spend a lot of time in the design phase, to ensure the API is very clean and intuitive.

That's certainly something we're considering, although supporting Java code for writing build scripts come with a lot of technical hurdles. I think it can be done technical. Remains to be seen if writing this Java code will be easy and straightforward enough.

Such a build system will become available in 2008 and this list is just a teaser. This build system will offer features most developers probably have never thought about.

Now all the Maven bashing in previous posts finally makes sense.

LOL

Hopefully this new build system makes sense. If Maven 2 works for you then we're happy about that. We just want to offer an alternative for people that are interested.

About the Maven bashing, I'm guilty as charged. However, I've been bashing Maven 2 long before I got involved in this new build system. Actually, I got involved because I think build systems can and should be better than Maven 2.

Such a build system will become available in 2008 and this list is just a teaser. This build system will offer features most developers probably have never thought about.

Now all the Maven bashing in previous posts finally makes sense.

LOL

Hopefully this new build system makes sense. If Maven 2 works for you then we're happy about that. We just want to offer an alternative for people that are interested.

About the Maven bashing, I'm guilty as charged. However, I've been bashing Maven 2 long before I got involved in this new build system. Actually, I got involved because I think build systems can and should be better than Maven 2.

Such a build system will become available in 2008 and this list is just a teaser. This build system will offer features most developers probably have never thought about.

Now all the Maven bashing in previous posts finally makes sense.

LOL

Hopefully this new build system makes sense. If Maven 2 works for you then we're happy about that. We just want to offer an alternative for people that are interested.

About the Maven bashing, I'm guilty as charged. However, I've been bashing Maven 2 long before I got involved in this new build system. Actually, I got involved because I think build systems can and should be better than Maven 2.

Such a build system will become available in 2008 and this list is just a teaser. This build system will offer features most developers probably have never thought about.

Now all the Maven bashing in previous posts finally makes sense.

LOL

Hopefully this new build system makes sense. If Maven 2 works for you then we're happy about that. We just want to offer an alternative for people that are interested.

About the Maven bashing, I'm guilty as charged. However, I've been bashing Maven 2 long before I got involved in this new build system. Actually, I got involved because I think build systems can and should be better than Maven 2.

Maven is an excellent tool to force common conventions regarding development environment, build & release management across team/projects.

These conventions are valuable. There is however no value in enforcing them.

Will it slow down developers, of course it will until they learn the new way as with any adoption of new technology.

Criticism on Maven 2 is founded. The core Maven 2 build system and its core plugins is poorly implemented. Bug fix releases are too few. Maven 2 and its developer team just don't seem to have what it takes to take developers down the new road.

Blaming developers is too easy. One interesting thing to notice though is that as far as I know there is today no other build system that tries to mimic Maven 2's conventions. That's about to change.

Maven is an excellent tool to force common conventions regarding development environment, build & release management across team/projects.

These conventions are valuable. There is however no value in enforcing them.

Will it slow down developers, of course it will until they learn the new way as with any adoption of new technology.

Criticism on Maven 2 is founded. The core Maven 2 build system and its core plugins is poorly implemented. Bug fix releases are too few. Maven 2 and its developer team just don't seem to have what it takes to take developers down the new road.

Blaming developers is too easy. One interesting thing to notice though is that as far as I know there is today no other build system that tries to mimic Maven 2's conventions. That's about to change.

Maven is an excellent tool to force common conventions regarding development environment, build & release management across team/projects.

These conventions are valuable. There is however no value in enforcing them.

Will it slow down developers, of course it will until they learn the new way as with any adoption of new technology.

Criticism on Maven 2 is founded. The core Maven 2 build system and its core plugins is poorly implemented. Bug fix releases are too few. Maven 2 and its developer team just don't seem to have what it takes to take developers down the new road.

Blaming developers is too easy. One interesting thing to notice though is that as far as I know there is today no other build system that tries to mimic Maven 2's conventions. That's about to change.

Maven is an excellent tool to force common conventions regarding development environment, build & release management across team/projects.

These conventions are valuable. There is however no value in enforcing them.

Will it slow down developers, of course it will until they learn the new way as with any adoption of new technology.

Criticism on Maven 2 is founded. The core Maven 2 build system and its core plugins is poorly implemented. Bug fix releases are too few. Maven 2 and its developer team just don't seem to have what it takes to take developers down the new road.

Blaming developers is too easy. One interesting thing to notice though is that as far as I know there is today no other build system that tries to mimic Maven 2's conventions. That's about to change.

But sadly, I am still forced to write a lot of stuff into my project descriptors that I shouldn't have to... like for generating an assembly or an RPM - these should just work out of the box. I just want to label my project as "webapp", "library", "user application" or whatever, tell it my dependencies, and have it all just work.

We want the same thing ;-)

But sadly, I am still forced to write a lot of stuff into my project descriptors that I shouldn't have to... like for generating an assembly or an RPM - these should just work out of the box. I just want to label my project as "webapp", "library", "user application" or whatever, tell it my dependencies, and have it all just work.

We want the same thing ;-)

But sadly, I am still forced to write a lot of stuff into my project descriptors that I shouldn't have to... like for generating an assembly or an RPM - these should just work out of the box. I just want to label my project as "webapp", "library", "user application" or whatever, tell it my dependencies, and have it all just work.

We want the same thing ;-)

But sadly, I am still forced to write a lot of stuff into my project descriptors that I shouldn't have to... like for generating an assembly or an RPM - these should just work out of the box. I just want to label my project as "webapp", "library", "user application" or whatever, tell it my dependencies, and have it all just work.

We want the same thing ;-)

It is fine that you don't like Groovy, no problem its a free world. However, this is the first article I have seen (and as I said I disagree with it and it is not the view of the Groovy community) that has stated Groovy is going to replace Java, most of that noise has come from the Ruby/Scala/Erlang/x language camps.

Graeme, I feel I have to respond before this gets out of hand. Does my article say Groovy will replace Java?

I can't believe I have to quote the article on top, but here goes:

Groovy offers joint Java compilation, meaning that .java and .groovy files get compiled at the same time. Java classes can thus extend Groovy classes without a glitch.

As Groovy IDE support improves - which is already decent for IntelliJ - more and more frameworks for Java will be written - at least in part - in Groovy. This is already happening.

As Groovy starts to play an increasing role in software development an increasing number of developers will be touched. Think Grails, think easyb, think of some novel new use of Groovy yourself.

(Grails and easyb built on top of the fact that Groovy integrates so well with Java)

Groovy will become very, very popular is what I'm saying. As such its market share will become so important that it will gradually close in on the Java language. Groovy's perfect integration with Java will probably be the main reason for its popularity.

Steven

It is fine that you don't like Groovy, no problem its a free world. However, this is the first article I have seen (and as I said I disagree with it and it is not the view of the Groovy community) that has stated Groovy is going to replace Java, most of that noise has come from the Ruby/Scala/Erlang/x language camps.

Graeme, I feel I have to respond before this gets out of hand. Does my article say Groovy will replace Java?

I can't believe I have to quote the article on top, but here goes:

Groovy offers joint Java compilation, meaning that .java and .groovy files get compiled at the same time. Java classes can thus extend Groovy classes without a glitch.

As Groovy IDE support improves - which is already decent for IntelliJ - more and more frameworks for Java will be written - at least in part - in Groovy. This is already happening.

As Groovy starts to play an increasing role in software development an increasing number of developers will be touched. Think Grails, think easyb, think of some novel new use of Groovy yourself.

(Grails and easyb built on top of the fact that Groovy integrates so well with Java)

Groovy will become very, very popular is what I'm saying. As such its market share will become so important that it will gradually close in on the Java language. Groovy's perfect integration with Java will probably be the main reason for its popularity.

Steven

It is fine that you don't like Groovy, no problem its a free world. However, this is the first article I have seen (and as I said I disagree with it and it is not the view of the Groovy community) that has stated Groovy is going to replace Java, most of that noise has come from the Ruby/Scala/Erlang/x language camps.

Graeme, I feel I have to respond before this gets out of hand. Does my article say Groovy will replace Java?

I can't believe I have to quote the article on top, but here goes:

Groovy offers joint Java compilation, meaning that .java and .groovy files get compiled at the same time. Java classes can thus extend Groovy classes without a glitch.

As Groovy IDE support improves - which is already decent for IntelliJ - more and more frameworks for Java will be written - at least in part - in Groovy. This is already happening.

As Groovy starts to play an increasing role in software development an increasing number of developers will be touched. Think Grails, think easyb, think of some novel new use of Groovy yourself.

(Grails and easyb built on top of the fact that Groovy integrates so well with Java)

Groovy will become very, very popular is what I'm saying. As such its market share will become so important that it will gradually close in on the Java language. Groovy's perfect integration with Java will probably be the main reason for its popularity.

Steven

It is fine that you don't like Groovy, no problem its a free world. However, this is the first article I have seen (and as I said I disagree with it and it is not the view of the Groovy community) that has stated Groovy is going to replace Java, most of that noise has come from the Ruby/Scala/Erlang/x language camps.

Graeme, I feel I have to respond before this gets out of hand. Does my article say Groovy will replace Java?

I can't believe I have to quote the article on top, but here goes:

Groovy offers joint Java compilation, meaning that .java and .groovy files get compiled at the same time. Java classes can thus extend Groovy classes without a glitch.

As Groovy IDE support improves - which is already decent for IntelliJ - more and more frameworks for Java will be written - at least in part - in Groovy. This is already happening.

As Groovy starts to play an increasing role in software development an increasing number of developers will be touched. Think Grails, think easyb, think of some novel new use of Groovy yourself.

(Grails and easyb built on top of the fact that Groovy integrates so well with Java)

Groovy will become very, very popular is what I'm saying. As such its market share will become so important that it will gradually close in on the Java language. Groovy's perfect integration with Java will probably be the main reason for its popularity.

Steven

It is fine that you don't like Groovy, no problem its a free world. However, this is the first article I have seen (and as I said I disagree with it and it is not the view of the Groovy community) that has stated Groovy is going to replace Java, most of that noise has come from the Ruby/Scala/Erlang/x language camps.

Graeme, I feel I have to respond before this gets out of hand. Does my article say Groovy will replace Java?

I can't believe I have to quote the article on top, but here goes:

Groovy offers joint Java compilation, meaning that .java and .groovy files get compiled at the same time. Java classes can thus extend Groovy classes without a glitch.

As Groovy IDE support improves - which is already decent for IntelliJ - more and more frameworks for Java will be written - at least in part - in Groovy. This is already happening.

As Groovy starts to play an increasing role in software development an increasing number of developers will be touched. Think Grails, think easyb, think of some novel new use of Groovy yourself.

(Grails and easyb built on top of the fact that Groovy integrates so well with Java)

Groovy will become very, very popular is what I'm saying. As such its market share will become so important that it will gradually close in on the Java language. Groovy's perfect integration with Java will probably be the main reason for its popularity.

Steven

It is fine that you don't like Groovy, no problem its a free world. However, this is the first article I have seen (and as I said I disagree with it and it is not the view of the Groovy community) that has stated Groovy is going to replace Java, most of that noise has come from the Ruby/Scala/Erlang/x language camps.

Graeme, I feel I have to respond before this gets out of hand. Does my article say Groovy will replace Java?

I can't believe I have to quote the article on top, but here goes:

Groovy offers joint Java compilation, meaning that .java and .groovy files get compiled at the same time. Java classes can thus extend Groovy classes without a glitch.

As Groovy IDE support improves - which is already decent for IntelliJ - more and more frameworks for Java will be written - at least in part - in Groovy. This is already happening.

As Groovy starts to play an increasing role in software development an increasing number of developers will be touched. Think Grails, think easyb, think of some novel new use of Groovy yourself.

(Grails and easyb built on top of the fact that Groovy integrates so well with Java)

Groovy will become very, very popular is what I'm saying. As such its market share will become so important that it will gradually close in on the Java language. Groovy's perfect integration with Java will probably be the main reason for its popularity.

Steven

It is fine that you don't like Groovy, no problem its a free world. However, this is the first article I have seen (and as I said I disagree with it and it is not the view of the Groovy community) that has stated Groovy is going to replace Java, most of that noise has come from the Ruby/Scala/Erlang/x language camps.

Graeme, I feel I have to respond before this gets out of hand. Does my article say Groovy will replace Java?

I can't believe I have to quote the article on top, but here goes:

Groovy offers joint Java compilation, meaning that .java and .groovy files get compiled at the same time. Java classes can thus extend Groovy classes without a glitch.

As Groovy IDE support improves - which is already decent for IntelliJ - more and more frameworks for Java will be written - at least in part - in Groovy. This is already happening.

As Groovy starts to play an increasing role in software development an increasing number of developers will be touched. Think Grails, think easyb, think of some novel new use of Groovy yourself.

(Grails and easyb built on top of the fact that Groovy integrates so well with Java)

Groovy will become very, very popular is what I'm saying. As such its market share will become so important that it will gradually close in on the Java language. Groovy's perfect integration with Java will probably be the main reason for its popularity.

Steven

It is fine that you don't like Groovy, no problem its a free world. However, this is the first article I have seen (and as I said I disagree with it and it is not the view of the Groovy community) that has stated Groovy is going to replace Java, most of that noise has come from the Ruby/Scala/Erlang/x language camps.

Graeme, I feel I have to respond before this gets out of hand. Does my article say Groovy will replace Java?

I can't believe I have to quote the article on top, but here goes:

Groovy offers joint Java compilation, meaning that .java and .groovy files get compiled at the same time. Java classes can thus extend Groovy classes without a glitch.

As Groovy IDE support improves - which is already decent for IntelliJ - more and more frameworks for Java will be written - at least in part - in Groovy. This is already happening.

As Groovy starts to play an increasing role in software development an increasing number of developers will be touched. Think Grails, think easyb, think of some novel new use of Groovy yourself.

(Grails and easyb built on top of the fact that Groovy integrates so well with Java)

Groovy will become very, very popular is what I'm saying. As such its market share will become so important that it will gradually close in on the Java language. Groovy's perfect integration with Java will probably be the main reason for its popularity.

Steven

It is fine that you don't like Groovy, no problem its a free world. However, this is the first article I have seen (and as I said I disagree with it and it is not the view of the Groovy community) that has stated Groovy is going to replace Java, most of that noise has come from the Ruby/Scala/Erlang/x language camps.

Graeme, I feel I have to respond before this gets out of hand. Does my article say Groovy will replace Java?

I can't believe I have to quote the article on top, but here goes:

Groovy offers joint Java compilation, meaning that .java and .groovy files get compiled at the same time. Java classes can thus extend Groovy classes without a glitch.

As Groovy IDE support improves - which is already decent for IntelliJ - more and more frameworks for Java will be written - at least in part - in Groovy. This is already happening.

As Groovy starts to play an increasing role in software development an increasing number of developers will be touched. Think Grails, think easyb, think of some novel new use of Groovy yourself.

(Grails and easyb built on top of the fact that Groovy integrates so well with Java)

Groovy will become very, very popular is what I'm saying. As such its market share will become so important that it will gradually close in on the Java language. Groovy's perfect integration with Java will probably be the main reason for its popularity.

Steven

It is fine that you don't like Groovy, no problem its a free world. However, this is the first article I have seen (and as I said I disagree with it and it is not the view of the Groovy community) that has stated Groovy is going to replace Java, most of that noise has come from the Ruby/Scala/Erlang/x language camps.

Graeme, I feel I have to respond before this gets out of hand. Does my article say Groovy will replace Java?

I can't believe I have to quote the article on top, but here goes:

Groovy offers joint Java compilation, meaning that .java and .groovy files get compiled at the same time. Java classes can thus extend Groovy classes without a glitch.

As Groovy IDE support improves - which is already decent for IntelliJ - more and more frameworks for Java will be written - at least in part - in Groovy. This is already happening.

As Groovy starts to play an increasing role in software development an increasing number of developers will be touched. Think Grails, think easyb, think of some novel new use of Groovy yourself.

(Grails and easyb built on top of the fact that Groovy integrates so well with Java)

Groovy will become very, very popular is what I'm saying. As such its market share will become so important that it will gradually close in on the Java language. Groovy's perfect integration with Java will probably be the main reason for its popularity.

Steven

It is fine that you don't like Groovy, no problem its a free world. However, this is the first article I have seen (and as I said I disagree with it and it is not the view of the Groovy community) that has stated Groovy is going to replace Java, most of that noise has come from the Ruby/Scala/Erlang/x language camps.

Graeme, I feel I have to respond before this gets out of hand. Does my article say Groovy will replace Java?

I can't believe I have to quote the article on top, but here goes:

Groovy offers joint Java compilation, meaning that .java and .groovy files get compiled at the same time. Java classes can thus extend Groovy classes without a glitch.

As Groovy IDE support improves - which is already decent for IntelliJ - more and more frameworks for Java will be written - at least in part - in Groovy. This is already happening.

As Groovy starts to play an increasing role in software development an increasing number of developers will be touched. Think Grails, think easyb, think of some novel new use of Groovy yourself.

(Grails and easyb built on top of the fact that Groovy integrates so well with Java)

Groovy will become very, very popular is what I'm saying. As such its market share will become so important that it will gradually close in on the Java language. Groovy's perfect integration with Java will probably be the main reason for its popularity.

Steven

It is fine that you don't like Groovy, no problem its a free world. However, this is the first article I have seen (and as I said I disagree with it and it is not the view of the Groovy community) that has stated Groovy is going to replace Java, most of that noise has come from the Ruby/Scala/Erlang/x language camps.

Graeme, I feel I have to respond before this gets out of hand. Does my article say Groovy will replace Java?

I can't believe I have to quote the article on top, but here goes:

Groovy offers joint Java compilation, meaning that .java and .groovy files get compiled at the same time. Java classes can thus extend Groovy classes without a glitch.

As Groovy IDE support improves - which is already decent for IntelliJ - more and more frameworks for Java will be written - at least in part - in Groovy. This is already happening.

As Groovy starts to play an increasing role in software development an increasing number of developers will be touched. Think Grails, think easyb, think of some novel new use of Groovy yourself.

(Grails and easyb built on top of the fact that Groovy integrates so well with Java)

Groovy will become very, very popular is what I'm saying. As such its market share will become so important that it will gradually close in on the Java language. Groovy's perfect integration with Java will probably be the main reason for its popularity.

Steven

It is fine that you don't like Groovy, no problem its a free world. However, this is the first article I have seen (and as I said I disagree with it and it is not the view of the Groovy community) that has stated Groovy is going to replace Java, most of that noise has come from the Ruby/Scala/Erlang/x language camps.

Graeme, I feel I have to respond before this gets out of hand. Does my article say Groovy will replace Java?

I can't believe I have to quote the article on top, but here goes:

Groovy offers joint Java compilation, meaning that .java and .groovy files get compiled at the same time. Java classes can thus extend Groovy classes without a glitch.

As Groovy IDE support improves - which is already decent for IntelliJ - more and more frameworks for Java will be written - at least in part - in Groovy. This is already happening.

As Groovy starts to play an increasing role in software development an increasing number of developers will be touched. Think Grails, think easyb, think of some novel new use of Groovy yourself.

(Grails and easyb built on top of the fact that Groovy integrates so well with Java)

Groovy will become very, very popular is what I'm saying. As such its market share will become so important that it will gradually close in on the Java language. Groovy's perfect integration with Java will probably be the main reason for its popularity.

Steven

It is fine that you don't like Groovy, no problem its a free world. However, this is the first article I have seen (and as I said I disagree with it and it is not the view of the Groovy community) that has stated Groovy is going to replace Java, most of that noise has come from the Ruby/Scala/Erlang/x language camps.

Graeme, I feel I have to respond before this gets out of hand. Does my article say Groovy will replace Java?

I can't believe I have to quote the article on top, but here goes:

Groovy offers joint Java compilation, meaning that .java and .groovy files get compiled at the same time. Java classes can thus extend Groovy classes without a glitch.

As Groovy IDE support improves - which is already decent for IntelliJ - more and more frameworks for Java will be written - at least in part - in Groovy. This is already happening.

As Groovy starts to play an increasing role in software development an increasing number of developers will be touched. Think Grails, think easyb, think of some novel new use of Groovy yourself.

(Grails and easyb built on top of the fact that Groovy integrates so well with Java)

Groovy will become very, very popular is what I'm saying. As such its market share will become so important that it will gradually close in on the Java language. Groovy's perfect integration with Java will probably be the main reason for its popularity.

Steven

Hey Graeme,

It depends how you count. If Groovy is used together with Java it's a 0.5 for Groovy and 0.5 for Java.

However, applications written with Grails are likely to have no Java code at all, so that's 1 for Groovy and 0 for Java. Given five more years of Groovy and Java evolution - both share the same ecosystem - things could turn out in surprising ways.

Steven

Hey Graeme,

It depends how you count. If Groovy is used together with Java it's a 0.5 for Groovy and 0.5 for Java.

However, applications written with Grails are likely to have no Java code at all, so that's 1 for Groovy and 0 for Java. Given five more years of Groovy and Java evolution - both share the same ecosystem - things could turn out in surprising ways.

Steven

Hey Graeme,

It depends how you count. If Groovy is used together with Java it's a 0.5 for Groovy and 0.5 for Java.

However, applications written with Grails are likely to have no Java code at all, so that's 1 for Groovy and 0 for Java. Given five more years of Groovy and Java evolution - both share the same ecosystem - things could turn out in surprising ways.

Steven

Hey Graeme,

It depends how you count. If Groovy is used together with Java it's a 0.5 for Groovy and 0.5 for Java.

However, applications written with Grails are likely to have no Java code at all, so that's 1 for Groovy and 0 for Java. Given five more years of Groovy and Java evolution - both share the same ecosystem - things could turn out in surprising ways.

Steven

"Building a better Maven with Ant"? Do you guys know the history of Maven? Maven was once a layer on top of Ant (say: "a better Ant" ;-) ). After the developers discovered, that Ant is not flexible enough, they created their own stack.

If this is true they made a huge mistake. One word: AntBuilder.

Maven is not just for dependency management. Its strengths lie in a comprehensive project structure and plugins, that make project management easy. With maven, you only set up a project descriptor. With Ant you have to create all the steps for yourself. With a working project descriptor, Maven can execute many reports and tools, such as building a website, creating documentation, running tests, publishing files etc. With Ant you have to do it all by yourself. Want some more? With a project descriptor, you can create IDE specific project files on the fly. You work with Netbeans, but your co-worker works with Eclipse? No problem.

Great ideas, unfortunately not very well implemented. These ideas are not unique to Maven. A Maven killer could easily provide them while at the same time make users happy, like Maven should do.

"Building a better Maven with Ant"? Do you guys know the history of Maven? Maven was once a layer on top of Ant (say: "a better Ant" ;-) ). After the developers discovered, that Ant is not flexible enough, they created their own stack.

If this is true they made a huge mistake. One word: AntBuilder.

Maven is not just for dependency management. Its strengths lie in a comprehensive project structure and plugins, that make project management easy. With maven, you only set up a project descriptor. With Ant you have to create all the steps for yourself. With a working project descriptor, Maven can execute many reports and tools, such as building a website, creating documentation, running tests, publishing files etc. With Ant you have to do it all by yourself. Want some more? With a project descriptor, you can create IDE specific project files on the fly. You work with Netbeans, but your co-worker works with Eclipse? No problem.

Great ideas, unfortunately not very well implemented. These ideas are not unique to Maven. A Maven killer could easily provide them while at the same time make users happy, like Maven should do.

"Building a better Maven with Ant"? Do you guys know the history of Maven? Maven was once a layer on top of Ant (say: "a better Ant" ;-) ). After the developers discovered, that Ant is not flexible enough, they created their own stack.

If this is true they made a huge mistake. One word: AntBuilder.

Maven is not just for dependency management. Its strengths lie in a comprehensive project structure and plugins, that make project management easy. With maven, you only set up a project descriptor. With Ant you have to create all the steps for yourself. With a working project descriptor, Maven can execute many reports and tools, such as building a website, creating documentation, running tests, publishing files etc. With Ant you have to do it all by yourself. Want some more? With a project descriptor, you can create IDE specific project files on the fly. You work with Netbeans, but your co-worker works with Eclipse? No problem.

Great ideas, unfortunately not very well implemented. These ideas are not unique to Maven. A Maven killer could easily provide them while at the same time make users happy, like Maven should do.

Hey Matt,

I can assure you that "Maven sucks" exists in the real world. As the Don Brown interview indicated, as I have repeated on numerous occasions and as other people continuously repeat Maven 2 makes its users loose time.

What's hard about big Ant files is the chore of setting up and maintaining them. Understanding the individual Ant tasks is not very hard, most if not all of them have decent documentation.

I've used the kind of common Ant files that Xavier proposes, they included Ivy support. The problems they caused for me is similar to my Maven 2 experience: when some unexpected error occurs, I may have to understand the internals of the black box I'm using in order to figure out the cause of my error.

This is a waste of time when it happens often. The effort required is frustrating because I'm unsure whether or not the complex abstraction is the cause of my error. The effort is not rewarding since often the complex abstraction makes it harder to understand the cause of certain errors.

A common set of Ant files and Maven's plugin system both enforces such a complex abstraction on their users. I believe that build systems need to have these layers of complex abstraction. The only way to avoid frustration and waste of time with your users as much as possible is then to:

1. Design, design, get feedback, re-iterate, design, design, keep on designing, ... until your users are happy. The Maven brotherhood is doing some of this but far too little. Being religious about your design decisions does not help your users at all.
2. Focus on your users, a lot. If your users are complaining, listen to them. Try to understand their complaints and try to help them. Again, the Maven brotherhood comes across as a religious sect that thinks it is right and nagging users are wrong.

In my experience Xavier has a history of listening to his users and helping them so I'm looking forward to study the design of his proposal. The big problem for me is the Maven culture where a few people are deemed to have the right opinions and everybody else is wrong or "doesn't get Maven". Maven sucks because they get some things right yet using Maven hurts and there's no medicine.

Steven

interesting perspective. I never thought of builders as a state machine (which they are, of course). However, the most useful part would be to allow IDEs to provide developer support when using a builder, i.e. mark state machine violations. To that end, a machine definition like yours could help.

The "builder-builder-DSL" could still be improved, though. I wouldn’t go for the ".once()" approach because it is too limiting (it’s the Java-style of doing builders). How about something that follows more the route of groovy MockFor?

Another approach would be to define it in terms of a grammar (think DTD or EBNF).

Speaking in terms of state machines would also suggest to define the structure as a state-transition-matrix.

interesting perspective. I never thought of builders as a state machine (which they are, of course). However, the most useful part would be to allow IDEs to provide developer support when using a builder, i.e. mark state machine violations. To that end, a machine definition like yours could help.

The "builder-builder-DSL" could still be improved, though. I wouldn’t go for the ".once()" approach because it is too limiting (it’s the Java-style of doing builders). How about something that follows more the route of groovy MockFor?

Another approach would be to define it in terms of a grammar (think DTD or EBNF).

Speaking in terms of state machines would also suggest to define the structure as a state-transition-matrix.

interesting perspective. I never thought of builders as a state machine (which they are, of course). However, the most useful part would be to allow IDEs to provide developer support when using a builder, i.e. mark state machine violations. To that end, a machine definition like yours could help.

The "builder-builder-DSL" could still be improved, though. I wouldn’t go for the ".once()" approach because it is too limiting (it’s the Java-style of doing builders). How about something that follows more the route of groovy MockFor?

Another approach would be to define it in terms of a grammar (think DTD or EBNF).

Speaking in terms of state machines would also suggest to define the structure as a state-transition-matrix.

interesting perspective. I never thought of builders as a state machine (which they are, of course). However, the most useful part would be to allow IDEs to provide developer support when using a builder, i.e. mark state machine violations. To that end, a machine definition like yours could help.

The "builder-builder-DSL" could still be improved, though. I wouldn’t go for the ".once()" approach because it is too limiting (it’s the Java-style of doing builders). How about something that follows more the route of groovy MockFor?

Another approach would be to define it in terms of a grammar (think DTD or EBNF).

Speaking in terms of state machines would also suggest to define the structure as a state-transition-matrix.

# 1.) Concept of transitive dependencies works well, but there is lack of well defined POM descriptors in most cases for OSS libraries. Absolut classic is Jakarta Commons Logging with Servlet API dependency. This is not problem of concept Atransitive dependencies, but crap POMs. There is simple solution, use an internal repository and maintain POMs itself. This approach may work for internal project, but doesn`t work for OSS projects. Therefore repo1.maven.org should be maintained in a better way.

# 2.) Use dependency management and collect all you dependencies and plugins to a central project POM.

# 3.) It`s funny because it isn`t problem of Maven. It is as saying that Eclipse is bad since there are poorly implemented plugins.

# 4.) From my point of view is better to depend on something what`ve been tested.

# 1.) Concept of transitive dependencies works well, but there is lack of well defined POM descriptors in most cases for OSS libraries. Absolut classic is Jakarta Commons Logging with Servlet API dependency. This is not problem of concept Atransitive dependencies, but crap POMs. There is simple solution, use an internal repository and maintain POMs itself. This approach may work for internal project, but doesn`t work for OSS projects. Therefore repo1.maven.org should be maintained in a better way.

# 2.) Use dependency management and collect all you dependencies and plugins to a central project POM.

# 3.) It`s funny because it isn`t problem of Maven. It is as saying that Eclipse is bad since there are poorly implemented plugins.

# 4.) From my point of view is better to depend on something what`ve been tested.

# 1.) Concept of transitive dependencies works well, but there is lack of well defined POM descriptors in most cases for OSS libraries. Absolut classic is Jakarta Commons Logging with Servlet API dependency. This is not problem of concept Atransitive dependencies, but crap POMs. There is simple solution, use an internal repository and maintain POMs itself. This approach may work for internal project, but doesn`t work for OSS projects. Therefore repo1.maven.org should be maintained in a better way.

# 2.) Use dependency management and collect all you dependencies and plugins to a central project POM.

# 3.) It`s funny because it isn`t problem of Maven. It is as saying that Eclipse is bad since there are poorly implemented plugins.

# 4.) From my point of view is better to depend on something what`ve been tested.

# 1.) Concept of transitive dependencies works well, but there is lack of well defined POM descriptors in most cases for OSS libraries. Absolut classic is Jakarta Commons Logging with Servlet API dependency. This is not problem of concept Atransitive dependencies, but crap POMs. There is simple solution, use an internal repository and maintain POMs itself. This approach may work for internal project, but doesn`t work for OSS projects. Therefore repo1.maven.org should be maintained in a better way.

# 2.) Use dependency management and collect all you dependencies and plugins to a central project POM.

# 3.) It`s funny because it isn`t problem of Maven. It is as saying that Eclipse is bad since there are poorly implemented plugins.

# 4.) From my point of view is better to depend on something what`ve been tested.

Troggan, your comment doesn't add any value or insight and is passive aggressive.

I clearly said that the "stupid dog" works for our small team and that it's not proven for remote teams.

And since I assume you haven't tried "the stupid dog" you're in no position to make any judgment on how it compares to Teamcity.

Frankly, your comment is lame and I'd appreciate it if you defer from ever making such comments on DZone again.

Troggan, your comment doesn't add any value or insight and is passive aggressive.

I clearly said that the "stupid dog" works for our small team and that it's not proven for remote teams.

And since I assume you haven't tried "the stupid dog" you're in no position to make any judgment on how it compares to Teamcity.

Frankly, your comment is lame and I'd appreciate it if you defer from ever making such comments on DZone again.

Troggan, your comment doesn't add any value or insight and is passive aggressive.

I clearly said that the "stupid dog" works for our small team and that it's not proven for remote teams.

And since I assume you haven't tried "the stupid dog" you're in no position to make any judgment on how it compares to Teamcity.

Frankly, your comment is lame and I'd appreciate it if you defer from ever making such comments on DZone again.

From a technical perspective Grails is clearly the winner compared to Rails, with a big margin. It's easier to deploy, easier to cluster, easier to integrate with Java applications, ... .

For many people these arguments do not matter and for many other they do. Saying that any one of both Rails or Grails is better in general is according to me baseless. It all depends on what you need.

And like with previous battles of the frameworks many people will continue to choose frameworks for the wrong reasons.

Every application is different, and one testing strategy may be a better fit than an other.

We test in the container a lot and it has helped us to increase the quality of our software dramatically, something that we could not achieve before when we focused on unit testing. Running tests takes a lot of time, we gladly accept that and have found ways to work around that problem.

[...] I do hope that dzone doesn't clone InfoQ [...]

[...] I do hope that dzone doesn't clone InfoQ [...]

[...] I do hope that dzone doesn't clone InfoQ [...]

Naturally, if you didn't specify exact version of plugin you are using in your pom.xml Maven will look for updates and will alsways use the latest version. That is why it is a good idea to always specify plugin versions and I would call it one of the most critical maven best practices.

That would indeed work if it were not that the old version is called 1.0-SNAPSHOT and the new version is called .... 1.0-SNAPSHOT.

Please, just release plugins together with the build system. Why be so persistent in doing the wrong thing?

Naturally, if you didn't specify exact version of plugin you are using in your pom.xml Maven will look for updates and will alsways use the latest version. That is why it is a good idea to always specify plugin versions and I would call it one of the most critical maven best practices.

That would indeed work if it were not that the old version is called 1.0-SNAPSHOT and the new version is called .... 1.0-SNAPSHOT.

Please, just release plugins together with the build system. Why be so persistent in doing the wrong thing?

Naturally, if you didn't specify exact version of plugin you are using in your pom.xml Maven will look for updates and will alsways use the latest version. That is why it is a good idea to always specify plugin versions and I would call it one of the most critical maven best practices.

That would indeed work if it were not that the old version is called 1.0-SNAPSHOT and the new version is called .... 1.0-SNAPSHOT.

Please, just release plugins together with the build system. Why be so persistent in doing the wrong thing?

As for JavaRebel, those are some extraordinary claims. Since there are no technical details and its not open source there's no way to know what is and what is not.

Ehmh... This really isn't some cat in a bag. Just download and try. We have a lot of happy users, which will testify for that as well.

As for JavaRebel, those are some extraordinary claims. Since there are no technical details and its not open source there's no way to know what is and what is not.

Ehmh... This really isn't some cat in a bag. Just download and try. We have a lot of happy users, which will testify for that as well.

As for JavaRebel, those are some extraordinary claims. Since there are no technical details and its not open source there's no way to know what is and what is not.

Ehmh... This really isn't some cat in a bag. Just download and try. We have a lot of happy users, which will testify for that as well.

So this expensive product actually can not reload applications when things get hairy. On top of that, it does nothing more than what Grails does in terms of reloading or what any other web framework could implement.

It reloads any changes to classes and works for all containers and all frameworks. Whereas Grails only works for Grails. And you will never get anything near the same level of support in web frameworks (at least not the Tap5 way). It works for pretty special cases, when the whole state of the instances is reconstructible, since without dropping all existing instances the classes will not reload. And even at that there are cases when it will fail due to classloader conflicting. JavaRebel uses a completely different approach.

So this expensive product actually can not reload applications when things get hairy. On top of that, it does nothing more than what Grails does in terms of reloading or what any other web framework could implement.

It reloads any changes to classes and works for all containers and all frameworks. Whereas Grails only works for Grails. And you will never get anything near the same level of support in web frameworks (at least not the Tap5 way). It works for pretty special cases, when the whole state of the instances is reconstructible, since without dropping all existing instances the classes will not reload. And even at that there are cases when it will fail due to classloader conflicting. JavaRebel uses a completely different approach.

So this expensive product actually can not reload applications when things get hairy. On top of that, it does nothing more than what Grails does in terms of reloading or what any other web framework could implement.

It reloads any changes to classes and works for all containers and all frameworks. Whereas Grails only works for Grails. And you will never get anything near the same level of support in web frameworks (at least not the Tap5 way). It works for pretty special cases, when the whole state of the instances is reconstructible, since without dropping all existing instances the classes will not reload. And even at that there are cases when it will fail due to classloader conflicting. JavaRebel uses a completely different approach.

If this expensive product can keep a Hibernate SessionFactory object working after a reload of domain classes, I'd be impressed.

Domain classes will update without a glitch, but of course you need to rebuild the session factory at the moment. It's really not that much of a problem, as database schema does not update too often. Of more concern are Dependency Injection frameworks like Spring, since business layer is updated much more often. We do have some plans for those.

If this expensive product can keep a Hibernate SessionFactory object working after a reload of domain classes, I'd be impressed.

Domain classes will update without a glitch, but of course you need to rebuild the session factory at the moment. It's really not that much of a problem, as database schema does not update too often. Of more concern are Dependency Injection frameworks like Spring, since business layer is updated much more often. We do have some plans for those.

If this expensive product can keep a Hibernate SessionFactory object working after a reload of domain classes, I'd be impressed.

Domain classes will update without a glitch, but of course you need to rebuild the session factory at the moment. It's really not that much of a problem, as database schema does not update too often. Of more concern are Dependency Injection frameworks like Spring, since business layer is updated much more often. We do have some plans for those.

One of the bigger issues in Maven's design is that plugins are not released with the build system.

Automatically updating plugins is non-sense. For example, just yesterday the Emma plugin was updated, out of the blue. We had to upgrade to Maven 2.0.8 on all our developer machines and CI machines.

Maven is supposed to be about gaining time and not developing the build system, remember?

The hard part is reloading domain classes ...

And thats where "expensive" products come in. JavaRebel really does reload everything and always.

The hard part is reloading domain classes ...

And thats where "expensive" products come in. JavaRebel really does reload everything and always.

The hard part is reloading domain classes ...

And thats where "expensive" products come in. JavaRebel really does reload everything and always.

Actually, reloading a controller/action class is pretty easy to implement. Grails and Tapestry 5 have it. Spring supports reloading of Groovy classes.

It's important to understand that reloading classes involves playing with the Java ClassLoader so there's no difference between Java and Groovy. You also don't need expensive products :-)

The hard part is reloading domain classes ...

Yeah, it's my MO :-)

I wrote a Builder once with BuilderSupport and I found it not intuitive at all. BuilderSupport is a good solution if you want to write a Builder in Java, but in Groovy EMC and delegates is in my opinion more interesting.

You can still call upon propertyMissing() and methodMissing() if you want to. Actually, it would be nice if

you could overload these methods, like:

void methodMissing(String name, Closure cl) { }

or

void methodMissing(String name, Map params = [:], String[] names, Closure cl) {}

That would make DSL support code even more concise.

Yeah, it's my MO :-)

I wrote a Builder once with BuilderSupport and I found it not intuitive at all. BuilderSupport is a good solution if you want to write a Builder in Java, but in Groovy EMC and delegates is in my opinion more interesting.

You can still call upon propertyMissing() and methodMissing() if you want to. Actually, it would be nice if

you could overload these methods, like:

void methodMissing(String name, Closure cl) { }

or

void methodMissing(String name, Map params = [:], String[] names, Closure cl) {}

That would make DSL support code even more concise.