Apache is being forced into a Java Fork
zdnet.com
zdnet.com
* Harmony is already a fork of java. But without the patent grant there is an unknown.
* ASF wanted to pass the TCK for the patent grant and wanted Oracle to get rid of the field of use related to the tck.
* Dalvik is using some of Harmony's code.
In order for it to be a fork, something has to get forked from the parent ... like the source code, or the standard.
Given the disappointment over closures and collection literals being delayed until java 8 there must be a temptation within the harmony camp to push ahead with these features and produce the version of java that many are clamoring for.
This is why I don't see a fork sustainable. Hostilities have already been mounted. I think Oracle taints the entire concept of using the JVM for anything as far as Apache is concerned. The only think I see now is the transition to separate and a completely independent technology. I wonder how difficult it would be to port the language and classes over to the LLVM, if their would even be legal ramifications in doing that and in the end if it is worth it, given that languages have progressed beyond Java now. Still maybe Apache would get something together that would allow it's libraries to operate on the LLVM while embracing a new language for future development. It's food for thought given that their is an imminent need to move beyond Oracle's reach.
GCJ would be what you're looking for here, I think. The critical question would be how badly, if at all, GCJ would be affected by an Oracle militantly pursuing perceived patent violators.
> languages have progressed beyond Java now
It's not about the language. It's about the collosal amount of libraries in the ecosystem.
My point as far as languages go was not that the libraries are not worth it but if you have to go so far as moving all of them over too a wholly separate run-time, is it even worth it to continue Java the language or would it be better to just adopt Scala or some other language that incorporates modern features and make it compatible with the ported library set. I was really pondering if you go all that way, is it worth it to just go that last 1/10 of a mile, instead of trying to bring Java the language into the modern age.
Isn't that basically what Google did with Android/Dalvik and are now being sued by Oracle for?
I am not 100% sure of the case, but it seems like that was what the articles where saying when the Google suit was hot in the news. I will go back and see if I can find info on it.
EDIT: http://tirania.org/blog/archive/2010/Aug-13.html
This would explain why Google went through the trouble of making the Dalvik virtual machine explicitly incompatible with the existing Java virtual machine instructions.
yes it seems that the VM that Google used is as Microsoft's was incompatible with vanilla JVM instructions.
I wrote a text about this earlier today if someone is interested. Not published here since it is kinda rantish.
http://labs.teppefall.com/2010/12/calling_oracles_sco_bluff....
For me personally, if I where asked to throw my lot in with Apache or Oracle, it would be Apache, I use their technology every day, I do a project with Oracle technology (excluding the JVM now) maybe once a year. I think Oracle thinks the entrenchment is to great to overcome, for our sake I hope they are wrong.
Fortunately for myself, I was an assembly developer before I was a C developer before I was a Java developer. So my exit, should it become untenable is easy enough given that I am a freelancer. I don't have a significant technology infrastructure investment like product people have. I truly feel for you guys that do, I would be really worried if I where in your shoes.