>> > On Thu, May 20, 2010 at 2:14 PM, Adrian Cole <[email protected]> wrote: >> >I suggest joining a community like jclouds >> Again, the goal is a single API abstraction, not multiple.
I definitely see the goal here, and support it. Adrian, do you disagree? On Thu, May 20, 2010 at 12:39 PM, Adrian Cole <[email protected]> wrote: > Is there anyone who isn't from IBM that would like to add a java or C > rewrite of libcloud to the project? I would not mind seeing libcloud ported to other languages. Another options is working with other projects (such as jclouds) to agree on interfaces. However, I feel like we have discussed doing the same interfaces and decided it was not the best approach. There seems to be interest from others too (although not much progress): http://rubyforge.org/projects/libcloud/ (and this java thread) Again, my concern is that there is a strong community behind it. Sounds like Eric and IBM are committed to building it... are you committed to maintaining it, adding drivers, and helping build the java libcloud community? -Alex > On Thu, May 20, 2010 at 12:24 PM, Peter Haggar <[email protected]> wrote: > >> I think a Java version makes sense and would like to contribute to it. >> >> >too early to create fragmentation in the cloud communities. >> That's exactly what I am trying to avoid and why I think Eric's suggestion >> is sensible. There's too much fragmentation now and his idea is a way to >> coalesce some of that under one umbrella. >> >> >Libcloud is undeniably python focused, and Java doesn't make sense here. >> No argument that it is python focused today, but the question is does it >> make sense to expand to other languages? I think the answer is Yes because >> then there is a single API abstraction that supports a wide variety of IaaS >> vendors in multiple languages. Can you explain why that doesn't make >> sense? >> Moving this to >> jclouds would create fragmentation and would require users to use the >> libCloud API for an IaaS abstraction if they are coding in Python, but a >> different IaaS API if they use Java. Presumably if the community wanted to >> add other languages you'd suggest yet another API? That makes no sense. >> I'd prefer one API. >> >> >...and find it a bad fit, submit your apache incubator proposal. >> I would like to avoid that as that creates fragmentation. >> >> Peter >> > >> >> > My thoughts are its too early to create fragmentation in the cloud >> > communities. Libcloud is undeniably python focused, and Java doesn't >> make >> > sense here. I suggest joining a community like jclouds or dasein where >> you >> > don't have to change it into something different. Im not sure what you >> are >> > so worried about, as IBM cloud is an easy api and on our todolist. If you >> > actually try an existing java community and find it a bad fit, submit >> your >> > apache incubator proposal. >> > >> > My 2p >> > Adrian >> > >> > On May 20, 2010 10:36 AM, "Eric Woods" <[email protected]> wrote: >> > >> > Hey guys, >> > >> > I've been following this discussion about libcloud's future direction >> very >> > closely, and thank you all for the great input. I'm trying to make sense >> > of >> > all this and how it relates to the Java development. I'm still very >> > interested in developing a Java code base for libcloud, but I'm not >> getting >> > a clear sense of the community's support. I, of course, don't want to >> > spend >> > the development time and effort writing a Java version of libcloud if the >> > community doesn't support it. >> > >> > What are your thoughts on Java development as a sub-project of Apache >> > libcloud? >> > >> > Thanks, >> > - Eric >> > >> > -- co-founder, cloudkick.com twitter.com/cloudkick echo V2UgYXJlIGhpcmluZyEK | openssl base64 -d
