Community
Participate
Working Groups
Mylyn has had incubation quality components prior to the introduction of Incubator projects at Eclipse. The umbrella for these components has been the Mylyn Sandbox: http://wiki.eclipse.org/Mylyn_Contributor_Reference#Sandbox There is now a mechanism that supports this kind of split, the twinned incubation projects used by WTP and Platform, described here: http://eclipse-projects.blogspot.com/2007/09/getting-new-people-started-in-your.html We have been talking about the benefits of this approach for some time now and should explore doing this sooner rather than later. In addition to the benefits listed in Bjorn's blog post, some key benefits are: * Our custom Sandbox infrastructure will properly follow the Eclipse project model * We will have access to the parallel IP process * We will be able to make the Incubator project a place that encourages experimentation and innovation by having a lower quality/ui/tests bar for contributions
+1 that sounds like a great idea to help us pursue some of the ideas that are not in the main focus of the project but could provide a great benefit such as Wiki integration.
Steffen: would you be interested in taking on this project proposal? It's pretty straightforward and I can guide you through it.
Need to get moving on this since the summer of code is coming.
Email sent to Tools PMC. Steffen: if we don't hear back in a few days we may want to go ahead with submitting a proposal that's derived from the Equinox one.
David: we havnen't heard back from the PMC on this yet so we will be creating a standard project proposal.
thanks for CC'ing me on this one -- I look forward to getting started on the Mylyn incubator as discussed. Please let me know if I can help move this one along.
Here's a pass at the project structure. Let's iterate on tomorrow's call. We still need to write up the exact criteria for each category, but I put some notes in square brackets. It's unfortunate that we still need an Extras update site and we can consider ways of getting rid of that. All sites should be obviously linked from a single download page in order to encourage installation. EPP/Ganymede distribution [mature components, supported by project committers with P1s fixed by next release] * Features * Task List * Focused UI * Integration * Mylyn Bridge: Eclipse IDE * Mylyn Bridge: Java Development * Mylyn Bridge: Plug-in Development * Mylyn Bridge: Team Support * Mylyn Connector: Bugzilla Mylyn Extras [mature components, community supported with no support guarantees made by the project] * Features * Mylyn UI Usage Reporting * Integration * Mylyn Connector: Trac [see bug 235224] * Mylyn Connector: JIRA [might move] Mylyn Incubator [supported by Incubator component leads] * Features * Mylyn UI Experiments (Incubation) * Mylyn WikiText (Incubation) * Mylyn Developer Tools (Incubation) [not on update site] * Integration * Mylyn Connector: Web Templates (Incubation) * Mylyn Connector: XPlanner (Incubation) * Mylyn Trac Connector WikiSupport (Incubation) [not on update site]
Created attachment 103215 [details] revised proposal Steffen: here is a revision of your proposal.
(In reply to comment #7) > EPP/Ganymede distribution [mature components, supported by project committers > with P1s fixed by next release] > ... > Mylyn Extras [mature components, community supported with no support guarantees > made by the project] > ... > Mylyn Incubator [supported by Incubator component leads] > * Features > * Mylyn Connector: Web Templates (Incubation) I am confused. If I am readin this correctly, web connector is not supported by project committers, immature and don't have community?
The statement in the angle brackets was not intended to be read negatively and that's definitely not what I was implying for either of XPlanner or the Web Connector. I think that we agreed in conversation about the Web Connector going to the incubator project. Did you have any additional thoughts about that? I'm wondering since this is our opportunity to structure the incubator project in a way that meets your and that connectors' communities' needs. I think that the other components (XPlanner, the upcoming WikiText contribution, bug 234210) will have similar needs (e.g., visibility of the download page). I have some concern that we are making several connectors harder to find: * Trac moves from EPP to Extras * XPlanner moves from Extras to Incubation * Web Connector moves from Extras to Incubation So in addition to structuring it right, we need to make sure that the download page and all other relevant pages and site links do a good job informing users where to get the connectors they need.
(In reply to comment #10) > I think that we agreed in conversation about the Web Connector going to the > incubator project. Did you have any additional thoughts about that? We also agreed that before moving anything to incubator we should have clearly defined criteria for moving things out. If that you put in those brackets is such criteria, then it doesn't seem like web connector qualify for incubation.
(In reply to comment #10) > The statement in the angle brackets was not intended to be read negatively To clarify, I wasn't reading it negatively, but it is quite confusing.
(In reply to comment #12) > (In reply to comment #10) > > The statement in the angle brackets was not intended to be read negatively > > To clarify, I wasn't reading it negatively, but it is quite confusing. I didn't mean negatively in the English sense, but in the mathematical one (i.e., the fact that one site's description did not include text of another's didn't mean that it was an attribute exclusive of that site). But I was not very clear. Your point about having clearly-defined and documented criteria is a very important one and I will post those for review asap.
Notes from today's call: * Need to update project proposal with components * Need to decide on CVS structure, whether to group by components * Create an Incubator component in Bugzilla
I would be interested in working on these incubator projects: - the developer tools, as I provided most property sources and the repository spy - the generic SQL connector, which has somewhat stalled waiting for 3.0 release and the incubator project Can the SQL connector be added to this? pending Wim's approval as well of course
Wim agrees! (In reply to comment #15) > Can the SQL connector be added to this? pending Wim's approval as well of course Wim agrees, can you add this to integration? * Mylyn Connector: Generic SQL (Incubation)
Created attachment 106272 [details] updated to include the Generic SQL Connector
Steffen: I pinged the Tools PMC about this (CC'ing you) to see what the next step is. I'll take this back for now.
Created attachment 106507 [details] updated with mentors listed
Steffen: please submit.
After reading through the Eclipse process documents in more details I have a slight concern that a separate incubator project would add significant process overhead and would be time consuming to setup due to the required reviews and approvals. Each Eclipse project has it's own infrastructure such as mailing lists, website, download area etc. that would have to be maintained and may duplicate existing facilities of the Mylyn project. The Eclipse Tools Project charter mentions that projects can be divided further into components with a different set of committers and commit privileges: http://www.eclipse.org/tools/eclipsetools-charter.php "The component lead is designated as a committer for the Project and represents the component in discussions and votes pertaining to the Project as a whole. Component committers do not participate in votes at the level of the Project as a whole, unless they are also the component lead." Based on the incubator project proposal existing code could be moved into these components and additional components could be considered (e.g. Generic SQL connector): - XPlanner (lead by Helen Bershadskaya) - WikiText (lead by David Green) - Sandbox (lead by a Mylyn committer) The components would benefit from the Mylyn community and get visiblity through aligned releases. The differences in maturity or the experimental character of components could be communicated through the website as well as by the naming of the update site and features. Another advantage would be that additional committers on the Mylyn project would broaden the diversity of the project. The advantage of a parallel IP process would be limited to the cases that comply with the policy changes recently made for mature projects by the Eclipse board. This may slow merging of contributions but on the other hand we have had few large contributions in the past that required a formal review compared to the amount of smaller contributions. My sense it that dividing the Mylyn projects into PMC established components would provide most of the benefits of a separate incubator project with significant less overhead.
As per our discussion on the call, this makes sense to me too. Let's hold the incubator proposal for now to see if we can make things work sufficiently with components (or sub-projects, if the pending changes to the Eclipse Development Process go through).
Stale. We're continuing with the sandbox approach for now. Can consider an incubator sub-project later.
