On Fri, Nov 28, 2008 at 12:57 PM, Wojtek Janiszewski < [EMAIL PROTECTED]> wrote:
> I'd like to start working on integration with Erlang soon. Where should I > do my commits? This extension could be experimental at the beginning so is > it better to use sandbox instead of sca/modules in trunk? > > Thanks, > Wojtek > > > Wojtek Janiszewski wrote: > >> I did a little research on Erlang integration and decided to publish >> results on wiki [1] (it's much more readable than in e-mail). >> >> What do you think? Is there anyone with some knowledge on Erlang? I'll >> appreciate any comments. >> >> Thanks, >> Wojtek >> >> [1] - >> http://cwiki.apache.org/confluence/display/TUSCANYWIKI/Integration+with+Erlang >> >> Raymond Feng wrote: >> >>> I have to admit that I have little or very limited knowledge about these >>> technologies. With some reading of the materials on the internet, I think >>> they are all interesting. >>> >>> Map/Reduce is still a TODO. The GSoC project didn't achieve much for this >>> area. >>> >>> For Erlang, my understanding from [3] is that it's more like a >>> combination of binding and implementation, similar with implementation.ebj >>> and binding.ejb. >>> >>> Thanks, >>> Raymond >>> >>> -------------------------------------------------- >>> From: "Wojtek Janiszewski" <[EMAIL PROTECTED]> >>> Sent: Tuesday, November 11, 2008 7:53 AM >>> To: <[EMAIL PROTECTED]> >>> Subject: [PROPOSAL] Support for Styx, Erlang, Map/Reduce >>> >>> Hi, >>>> as some of you know I'm (still) in front of choosing subject of my >>>> master >>>> thesis. I decided to connect it with Tuscany and I've picked three >>>> topics >>>> which I'm interested in: >>>> >>>> >>>> 1. Support for Styx protocol, which is used in operating systems like >>>> Plan9 or Inferno. >>>> >>>> This extension could be realized as binding extension which could >>>> provide >>>> access to Styx resources (reference bindings). Tuscany components could >>>> be >>>> also served as Styx resources (service bindings). I thought I could use >>>> JStyx which is Java Styx implementation [1]. Little challenge here could >>>> be inventing method of mapping Java interface to structure of Styx >>>> resource (tree). >>>> >>>> 2. Support for Erlang language (inspired by GSoC 2008 proposal). >>>> >>>> This one sounds interesting, but after reading [2] and [3] I'm still not >>>> sure how this could work as implementation type in Tuscany. Can we >>>> assume >>>> that input implementation file contains list of erl shell commands which >>>> would be translated to JInterface calls? >>>> >>>> 3. Support for Map/Reduce - integration with Apache Hadoop (GSoC 2008 >>>> proposal). >>>> >>>> It looks like this GSoC project wasn't finished successfully. Is it >>>> true? >>>> If so then this project is also worth of more research. >>>> >>>> >>>> I'll appreciate any comments. What do you think about usefulness of each >>>> proposition? >>>> >>>> Thanks, >>>> Wojtek >>>> >>>> [1] - http://www.resc.rdg.ac.uk/jstyx/index.html >>>> [2] - http://wiki.apache.org/general/SummerOfCode2008#tuscany-erlang >>>> [3] - >>>> >>>> http://www.theserverside.com/tt/articles/article.tss?l=IntegratingJavaandErlang >>>> >>>> >>> >> >> >> > Hi Wojtek The erlang thing looks interesting but have to admit that I haven't looked at the details. We have recently moved the trunk over to start developing toward a 2.x release so there is some cleanup going on in the core there. We have a /contrib directory there for new modules. However the changing nature of trunk at the moment may make it a little uncomfortable to develop new bindings there. Having said that the 1x code branch is still current so you could start off there and either develop in modules or your sandbox as you see fit and then move to 2.x when it's settled down a bit. Looking at erlang and JInterface they are released under something called the EPL (The Erlang one not the Eclipse one) which is derived from the MPL. Would have to take this past legal@ before committing any code that references the licensed material. Not necessairly a problem but have to check first. Regards Simon [1] http://www.apache.org/legal/3party.html
