Friday, 25 January 2008

The Case for haXe on the JVM

I believe if Nicolas wants haXe to grow beyond a language filling a niche, he needs to target the JVM.

Critical Mass

ES4/JS2 is going to eat haXe's lunch very soon (in terms of new users), strong typing, fast vm
and good platform coverage. New users looking for a platform will go with the known, _heavily_ promoted technology.

We need a critical mass of users, to make haXe self sustaining before ES4 is deployed.

For some time now the Java community has been discussing new language features for Java 7. It's likely that closures will make it into the language, but there is a large swell of opinion that prefers the language to remain closer to it's roots. At the same time there are newer dynamic JVM languages available; Groovy, scala
that were designed from the ground up to have closures, first class functions etc.

To sum up there are many Java programmers who want features of dynamic languages on the JVM and are looking for the correct vehicle on which to do so. This is where haXe can find it's new users.

Total Platform Coverage

haXe supports the premier client technology, Flash, but it does not support the premier server side technology, the JVM.

I think neko is great, i love it, but I can't sell it to my customers as a server side solution. I've tried! The JVM is a no brainer, I don't have to justify it's usage. So, right now, I simply can't do all my development in haXe; in this sense, I can say haXe does not provide server side development.

Further, haXe on the JVM could provide the killer application; Android/Java ME integration libraries.

Conclusion

If haXe supported the JVM it's my belief that it would be the top candidate for all jobs, it's end to end capabilities would simply be unrivalled, taking as it does the best from all worlds.

4 comments:

|3 |2 I /-\ |\| said...

Strongly concur.

If haxe could generate Java code, like it generates Javascript or AS3, I could start using it in production now.

A lot of shops have huge Java infrastructures already in place and are extremely resistant to change.

Alex Miller said...

Actually, it's been stated that closures will NOT make it into Java 7.

Dan Howard said...

I haven't looked much at haxe but since Java 6 there were a lot of new features added to aid in building a scripting language into Java. I would look at haxe if it targeted the JVM.

Anonymous said...

haxe just a toy, very poor unicode support