On Mar 23, 2011, at 8:23 PM, Vincent Massol wrote:
Hi devs,
It's time for the roadmap planning again. After a first round of offline discussions
with several committers here's a proposal for XE 3.1.
Please correct or add other things if I've missed anything.
Must have
========
- Extension Manager: Thomas and UI person (Jerome) (Cati needed for additional mockups).
Thomas to define more precisely the list of features he'd like to implement for XE
3.1
- Performance Improvements: Marius, maybe Caleb if available. Marius needs to prepare an
action plan ASAP since "perf improvements" in general is too vague. Caleb, you
mentioned you'd be ok to implement support for marge number of XObjects if you have
the time, right?
- New backend implementation of Rights: Sergiu. Requires proposals/discussions since
it's large and complex beast.
- "Application Within Minutes": continue investigation (Ludovic), committers to
review the Design page and since it's a large feature we need to break it down into
individual tasks and start implementing some of them in 3.1. Sergiu volunteered for
implementing some of them. Ludovic volunteered to do the breakdown.
- Search improvements and especially scoring, French support and wildcard issues: Sergiu
- Friends and Status polishing: Sergiu, but not clear what needs to be done except
"Privacy" which is needed
- Dashboard/Gadgets polishing: Anca (Anca could you detail what you'd like to
implement?)
- Stabilization of the new Filesystem storage implementation for attachments (bugfixing,
performance, etc): Caleb
Nice to have
==========
- Statistics polishing: ?
- XE/XEM Home Page investigation and mockups: Ludovic & Caty
- XEM Workspaces investigation/review: ?
- More modules moved out of xwiki-core. Candidates: xwiki-url, xwiki-action,
xwiki-localization, new xwiki-resources: Vincent to work on this
Proposed Dates
=============
- XE 3.1M1: 25 april
- XE 3.1M2: 16 May
- XE 3.1RC1: 30 May
- XE 3.1 final: 13 June
That's about 2.5 months. Note that at this rate we won't have time to reach XE
3.7 by the end of the year (end of the cycle). So we'll need to decide at some point
if:
a) we consider that a cycle is less major releases (say 5 only if we keep 2.5
months/release)
b) we consider that a cycle spans a year (time-based) and it doesn't matter if
we're at 3.5 or 3.11 at the end of the year, we consider the last release of the year
to be the end of the cycle
c) we consider that 2.5 months is too long for release cycles and we reduce it. To have 7
releases we need to reduce to 1.5 months per release roughly (that could be 2 weeks for
M1, 2 weeks for M2, 1 week for RC1, 1 week for final).
Actually forget all this we already have a strategy I had momentarily forgotten:
d) a cycle is 6 releases: N.0 till N.5 and is not bound to a time frame.... Our current
global planning says 3.5 will be around Jan
2012:http://enterprise.xwiki.org/xwiki/bin/view/Main/Roadmap
Thanks
-Vincent
My preference goes to b) right now and when we'll have moved more stuff out of the
old xwiki-core into new modules, it'll be easier to have tests committed when code is
committed and thus we'll be able to move to c) IMO.
WDYT?
Thanks
-Vincent