Talk:ProActive

ProActive notability
Hi thumperward,

You recently added the and  tags to article ProActive, my mistake was certainly that I didn't put references according to Wikipedia policy (I did put links to the INRIA website instead of a global paper publication website). I did some modifications and added third-party references.

Could you please check and tell me if it's now more correct ?

I could add plenty of references and papers about the subject, I don't see the point in putting too many of them.

Thanks

—Preceding unsigned comment added by Fabviale (talk • contribs) 17:31, 31 March 2009


 * Most of the references still appear to have been authored by the same people; this makes me a little wary of untagging just yet. However, it's good that you've added more. It's certainly very hard to over-reference an article, so feel free to keep adding them; I'll try to monitor the page's progress. Thanks. Chris Cunningham (not at work) - talk 17:36, 31 March 2009 (UTC)


 * The original 3 references are indeed publications done by member's of the ProActive team, but they make sense here as they are reference documents useful for understanding the theory behind ProActive. The other 4 publications are not at all made by members of the team, they cite ProActive, say for example they used it, compare it to another middleware, etc... Is there a general guildeline for separating references which directly talk of the subject and others which indirectly talk about it ?

I removed a sentence that sounded very promotional indeed (tribute of a blind copy/paste move), I don't see anything else here of the same taste. Please tell me if you see more. Fabien Viale (talk) 13:07, 1 April 2009 (UTC)

Questions...
The article doesn't work at the conceptual or detailed levels. Conceptually—assuming I know the basics of INRIA's distributed computing model from their work in the early 1980s—what is the driving force for ProActive?. In detail, how does ProActive relate to "ProActive Parallel Suite"? Is ProActive a set of abstract principles of which the ProActive Parallel Suite is an implementation in Java? How does this fit with Windows HPC Server? Was there an overarching set of OW2 Consortium principles/objectives that drove development of ProActive? Given that Groupe Bull and France Telecom are likely to have commercial goals in particpating with INRIA, what are their goals, with whom are they competing and how well does ProActive help them achieve their goals? - Pointillist (talk) 22:37, 31 March 2009 (UTC)
 * I understand your point of view and I agree with you that the article is not clear enough and needs to be extended/rewritten. I need to work on it whenever I find the time. What is mainly described here is the set of concrete principles upon which ProActive is based. ProActive (renamed recently ProActive Parallel Suite) represents both the set of principles and the implementation in Java. The most abstract aspects of ProActive (async calls, single request queue, single thread of execution) have been formalized into a model called ASP (Asynchronous sequential processes), and proofs have been written on the model's determinism. The three driving forces of ProActive are 1) this abstract model, it allows to design a parallel application without thinking on how many machines it will run, just by defining actors and roles (active objects) 2) the deployment framework, it acquires resources from many different architectures, p2p, interacts with existing cluster schedulers, even Windows HPC server, abstracts all these resources as a homogenous entities called Nodes, and allows to deploy Active Objects on them seamlessly. 3) Finally, the programmers Active Objects will send/receive messages to each other seamlessy across machines thanks to ProActive's communication protocol. Concerning your other questions, I'm sorry to say that wikipedia is not the best place to talk about commercial strategies. ProActive has nothing to do, in terms of commercial strategy, with France Telecom or Bull. Even if there have been partneship for joint projects (I'm thinking for example of France Telecom's component model called Fractal, on which ProActive component model is based). I don't know which INRIA's distributed computing model of the early 1980s you refer, INRIA is a big organisation, composed of numerous teams working in many different subjects. Today there are several teams working on distributed computing projects, at different levels. OASIS produced ProActive and works at the software level of distributed computing, other works on virtualization e.g. XtreemOS, other works at network level. So there hasn't definitely been just a single distributed computing model ever produced at INRIA. Fabien Viale (talk) 13:57, 1 April 2009 (UTC)