From 31044d5cfa4f861f510fc0fe130f5a82fae48b6a Mon Sep 17 00:00:00 2001 From: Jeremie Koenig Date: Fri, 8 Apr 2011 17:45:08 +0200 Subject: gsoc2011 (java): conclusion --- user/jkoenig/gsoc2011_proposal.mdwn | 47 +++++++++++++++++++++++++++++++++++++ 1 file changed, 47 insertions(+) (limited to 'user') diff --git a/user/jkoenig/gsoc2011_proposal.mdwn b/user/jkoenig/gsoc2011_proposal.mdwn index 7179ff87..5e5d2735 100644 --- a/user/jkoenig/gsoc2011_proposal.mdwn +++ b/user/jkoenig/gsoc2011_proposal.mdwn @@ -547,3 +547,50 @@ The dates listed are deadlines for the associated tasks. ## Conclusion +This project is arguably ambitious. +However, I have been thinking about it for some time now +and I'm confident I would be able to accomplish most of it. + +In the event multiple language bindings projects +would be accepted, +some work could probably be done in common. +In particular, +[ArneBab](http://www.bddebian.com/~hurd-web/community/weblogs/ArneBab/2011-04-06-application-pyhurd/) +seems to favor a low-level approach for his Python bindings as I do for Java, +and I would be happy to discuss API design and coordinate MIG changes with him. +I would also have an extra month after the end of the GSoC period +before I go back to school, +which I would be able to use to finish the project +if there is some remaining work. +(Last year's rewrite of procfs was done during this period.) + +As for the project's benefits, +I believe that good support for Java +is a must-have for the Hurd. +Java bindings would also further the Hurd's agenda +of user freedom by extending this freedom to more people: +I expect the set of developers +who would be able to write Java code against a well-written libfoofs +is much larger than +those who master the intricacies of low-level systems C programming. +From a more strategic point of view, +this would also help recruit new contributors +by providing an easier path to learning the inner workings of the Hurd. + +Further developments +which would build on the results of this project +include my planned [[experiment with Joe-E|objcap]] +(which I would possibly take on as a university project next year). +Another possibility would be to reimplement some parts +of the Java standard library +directly in terms of the Hurd interfaces +instead of using the POSIX ones through glibc. +This would possibly improve the performance +of some Java applications (though probably not by much), +and would otherwise be a good project +for someone trying to get acquainted with Hurd. + +Overall, I beleive this project would be fun, interesting and useful. +I hope that you will share this sentiment +and give me the opportunity to spend another summer working on Hurd. + -- cgit v1.2.3