[Rivet] Rivet sprint report, release plans, etc.

[Andy Buckley]

Hi Riveteers,

This week we've had a mini sprint session for Rivet, with a total of six
of us gathering in sunny (today!) Glasgow to argue, hack, and generally
make the world a better place. We decided that we should have these
meetings every 6 months, and the sketch plan is for the next to be in
Dresden in December, if Frank's schedule permits: to be confirmed.

We've done some very useful things, like going through the list of
unvalidated analyses to switch as many as possible to VALIDATED status,
or to OBSOLETE status or deleted. There are _14_ new experimental
analyses to be integrated into the next release, and Holger & Chris are
working through them with a mix of disbelief and impressive enthusiasm
;-) I've just made a bunch of improvements to the FastJets interface,
and David is improving the WFinder, whose shortcomings we debated a lot.

These big changes mean that we're going to go straight from release
2.2.1 to 2.3.0, rejecting the temptations of a 2.2.2 release code. We're
planning to make that release before the end of July. And we also
decided that we should be stricter in our (my!) use of the release
branches, to really always make new features or interface changes on the
default branch and reserve release-x-y-z branches for new analyses and
urgent bugfixes -- this should make the releases a bit less painful, and
we're going to try to keep to a schedule of at least one release every
three months. Be prepared to be asked to do your share of analysis
integration!

The next major release will be 3.0.0, and this will include the
longstanding multi-weights and event groups feature that currently lives
on a completely separate private development branch. This currently
works for 1D histograms, and David & Chris have been (and will continue
to!) working on extending the coverage to other histogram types. Frank
is going to help to extend this to include the "re-entrant finalize"
that we've discussed equally at length, but we think staging the work is
the way to get it out, so 3.0.0 will be a halfway step to the ultimate
plan for general multi-weight events and run merging. We're planning on
a 3.0.0 beta release in time for BOOST in August.

I think that's enough text from me. I'll be on holiday for the next 10
days, so hope to come back and find everything finished :-P

Thanks for an excellent few days, and here's to all these releases :-)

Andy

-- 
Dr Andy Buckley, Lecturer / Royal Society University Research Fellow
Particle Physics Expt Group, University of Glasgow