Re: [rt-pmc] Updated charter

[Joakim Erdfelt <joakim@xxxxxxxxxxx>]

Speaking on the Jetty side ...

The scope section seems ill worded, and too focused on a happy world with everything under the Eclipse umbrella using Equinox + OSGi.

While the jetty project make efforts to expose their artifacts in OSGi friendly ways (manifest + p2 repos), we do not use OSGi internally, even for our own module system.

We even have code for a set of jetty-osgi artifacts, but there is no one left on the Jetty side that uses it.  There has been a slow, steady, and increasing drumbeat from a number of committers for archiving those artifacts, as no person, group, or project seem to be using it.

We also tend to run far ahead of the rest of Eclipse when it comes to technology, for example, Jetty 9.1 (latest stable release) supports Servlet 3.1, which enforces a Java 1.7+ minimum requirement.

We know of no Eclipse projects willing to upgrade their OSGi projects to have a Java 1.7 minimum.

It's likely that the other Eclipse projects will eventually catch up, and start to use Jetty 9.1, but by then Jetty 11.x (on JDK 9?) will be out and Jetty 9.x will have no more active development.

This often means our efforts within OSGi are largely untested and unused by us.

This lack of synchronization between active Jetty development and other Eclipse project use means that the OSGi efforts in Jetty is often ignored or avoided as there seems to be little demand or pressure for it.

Maintaining knowledge/expertise across 3+ major versions of Jetty is a difficult task in of itself.

Note: Right now we essentially maintain 2 major versions, as Jetty 7 and Jetty 8 are essentially the same codebase, with minor changes for Servlet Spec differences.

Jetty 9 is a big change, the leap from Servlet 3.0 to 3.1 was significant.

So ultimately, the Scope of the RT project should be more relaxed to compensate for the realities present in the RT projects.  Avoid an urge to sneak in "must" or "should" style language into the Scope as related to OSGi.  Especially since the Eclipse Foundations is trending away from strictly OSGi, and instead is trying to appeal to a broader audience that is more focused on intellectual property concerns.

--

Joakim Erdfelt <joakim@xxxxxxxxxxx>

webtide.com - intalio.com/jetty

Expert advice, services and support from from the Jetty & CometD experts

eclipse.org/jetty - cometd.org

On Wed, Nov 20, 2013 at 7:24 AM, Thomas Watson <tjwatson@xxxxxxxxxx> wrote:

I think the scope should still have a few bullets that outlines the OSGi technology as it applies to runtimes.  But to be honest many of these bullets are specific things that fall under the charters for Equinox and others (e.g. ECF, gemini etc).

Here are a few bullets to get us started:

Overview:

Concretely, RT projects supply technologies that are used to build modular runtimes or provide a complete runtime themselves.  Many RT projects use a modular architecture based on OSGi technologies, but an RT project is not required to use or live in an OSGi Framework.   The nature of this work is scoped as follows:


 - The implementation of generally applicable runtime standards (e.g. OASIS, JCP, OSGi).
 - Implement various OSGi specifications that contribute to the modular architecture used by various runtime technologies at Eclipse.
 - Investigate and research future innovations in runtime technologies.  This includes influencing various expert groups that are developing new versions of runtime related specifications used by Eclipse.  For example, the OSGi Alliance.
 - Implement key components used to build the various platforms at Eclipse.

Others?

Tom



Ian Bull ---11/19/2013 10:34:33 PM---In between EclipseCon submission reviews (thank-you everyone who submitted a talk, BTW), I took a fi

From: Ian Bull <irbull@xxxxxxxxxxxxxxxxx>
To: Runtime Project PMC mailing list <rt-pmc@xxxxxxxxxxx>,
Date: 11/19/2013 10:34 PM
Subject: [rt-pmc] Updated charter
Sent by: rt-pmc-bounces@xxxxxxxxxxx

---

In between EclipseCon submission reviews (thank-you everyone who submitted a talk, BTW), I took a first stab at updating the RT Charter[1]. I've pushed it to GitHub so we can collaborate on it [2].

[1] https://github.com/irbull/rt-charter/blob/master/rt-charter.md
[2] https://github.com/irbull/rt-charter

I removed the OSGi specific parts, which went well for the first few sections. However, the section on 'scope' is entirely about OSGi. This raises the question, what is the scope of an RT Project now? Historically we were all about providing OSGi related technologies (bundles, reference implementations, etc...). Do we have anything defining now? Obviously 'Runtime' technologies, but that can be categorized as everything that's not tooling.

Thoughts?

Ian

--
R. Ian Bull | EclipseSource Victoria | +1 250 477 7484

http://eclipsesource.com | http://twitter.com/eclipsesource _______________________________________________
rt-pmc mailing list
rt-pmc@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/rt-pmc

_______________________________________________
rt-pmc mailing list
rt-pmc@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/rt-pmc