Why do you speak of OSGi vs JCP?? Wouldn't either JSR numbers by more appropriate, or for that matter IBM/OSGi vs Sun ?
On 10/16/06, BOTTARO Andre RD-MAPS-GRE
<[EMAIL PROTECTED]> wrote:
Not sure what you mean by 'exhaustive comparison'. The JSR-291 is based on 7 years of hard-earned experience, use-cases and evolution in a highly complex problem domain. There are several people on this list who knows that spec inside out.
On the other hand, the JSR-277 is mostly mind or whiteboard construct, possibly based on 'known practices' (such as OSGi and Maven). Noone (at least outside Sun) has any experience whatsoever with working with this spec, and most of us will have to wait a few more years.
I quite don't "get" the distinction between a class and a 'resource'. Somehow, the EG has decided that classes are more involved in the 'production process' to give it a separate status. Personally, I don't see a difference between that and, let's say, a PNG generated from SVG or a _javascript_ generated by an Ajax framework. Perhaps it will be clearer to me in the future how brilliant the idea is.
No. I doubt that OSGi will require Java 7 until much, much later, if ever. Possibly when the majority of mobile/embedded devices can run Java 7, maybe 10 years from now ;o)
What Sun seems to have ignored is that, OSGi provides a solution to most (and more) of the problem addressed by JSR-277, but does not require several years waiting time. It is nothing else but a NIH syndrome and "show who is in control". It is not the first time Sun has done this, and it won't be the last either.
I wonder if someone made an exhaustive comparison between the two models.
Not sure what you mean by 'exhaustive comparison'. The JSR-291 is based on 7 years of hard-earned experience, use-cases and evolution in a highly complex problem domain. There are several people on this list who knows that spec inside out.
On the other hand, the JSR-277 is mostly mind or whiteboard construct, possibly based on 'known practices' (such as OSGi and Maven). Noone (at least outside Sun) has any experience whatsoever with working with this spec, and most of us will have to wait a few more years.
They look pretty much the same, don't they ? However, there are new features like the import/export of resources in JCP's one for example.
I quite don't "get" the distinction between a class and a 'resource'. Somehow, the EG has decided that classes are more involved in the 'production process' to give it a separate status. Personally, I don't see a difference between that and, let's say, a PNG generated from SVG or a _javascript_ generated by an Ajax framework. Perhaps it will be clearer to me in the future how brilliant the idea is.
Natural questions are now:
- Will the OSGi Alliance align his modularity model on JCP's one before the arrival of Java 7 ?
No. I doubt that OSGi will require Java 7 until much, much later, if ever. Possibly when the majority of mobile/embedded devices can run Java 7, maybe 10 years from now ;o)
- How the JCP will handle the discrepancies between the technical answer to JSR 232, 291 and the one to JSR 277, 294 ? Do the JCP expect the modification of OSGi spec ?
Cheers
Niclas
_______________________________________________ OSGi Developer Mail List osgi-dev@bundles.osgi.org http://bundles.osgi.org/mailman/listinfo/osgi-dev
