protux-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Protux-devel] Possible Migration to Java


From: Luciano Domenico Giordana
Subject: [Protux-devel] Possible Migration to Java
Date: Wed, 31 Aug 2005 13:04:41 -0300

hello All !

This is just to let you know that I have been considering port protux/mustux to Java.
This is NOT a decision.This is just a intention "report". I´d like to hear from the community and specially from the team.
My reasons are quite founded on several "perceptions". Some of them :

- Java is a de-facto technology. QT isnt, yet.
- Productity in Java is hugely better than C++/QT, specially due Eclipse
- Developing using CDT on eclipse/Kdevelop , when compared to development java on eclipse, just suck.
- Java has a lot of ready-to-use classes that save us a lot of time. That includes powerfull collections, Wrappers, Serialization stuff, reflexion/instrospection, Garbage collecting, free/openm source components, and so on
- Exception handling in Java is far better than C++´s. We can even apply Responsability Oriented Programming concepts.
- Real clean coding is much easier in Java. No .h´s, prototypes, makefiles... stuff like that..
- building is ridiculously simple in eclipse. actually you dont even see it !
- Java performance has been compared to c++ performance. in some cases, it is faster due very smart modern VM´s (such Java Tiger) strategies, like native hot-compilation, hot syncing and binary caching.
- ALSA can be (as far as it seems) used in Java using JNI
- XML parsing are a transparent , ready-to-use and robust technology in Java
- So serialization is
- Documentation is also a strong point in java. Actually may doc engines in c++/qt was inspired from javadoc.
- There a lot of java programmers out there, much more than QT programmers. So the communities are stronger.
- GUI stuff can be made using visual editors, just like QT-designer. actually. it is easier, cause no moc/meta-classes are needed.
- fast painting device for audio editing can be done with hardware accelerated Java3D. other options can be considered.
- JMB seems to be perfectly implementable in Java.
- Deploy/distribution is very simple. We can even put a protux running in the site as an applet for previous-testing. More users can be attracted by that.
- Learnign curve for current developers is small. Actually, 1 month is enough for a c++ programmer to learn java.

there are more "perceptions". My intention is to show that we could speed up the project by taking this radical move. Please, RFC as much as you can.

Thanks




--
Luciano Domenico Giordana
Software Engineer / Java/C++ Senior Developer
IBM do Brasil - 55 19 2132-7423 - address@hidden - address@hidden
Project Protux : http://www.nongnu.org/protux

reply via email to

[Prev in Thread] Current Thread [Next in Thread]