[Rivet] Rivet 2 numerical validation survey complete... next steps to release

[Andy Buckley]

Hi all, particularly Hendrik and David,

I've just blasted through the remaining items flagged up as numerical
mismatch issues between Rivet2 with YODA and Rivet1 with AIDA... noting
that it's not only the histogramming that's changed.

In several cases the problem really looks to be on the AIDA side, e.g.
bins in output histos that don't exist in the AIDA reference. In quite a
few cases there were Rivet2/YODA issues to fix (which I've hopefully
done), mostly to do with our change to the bookScatter2D behaviour (it
doesn't create a set of zeroed points with x values+errs based on the
ref data anymore, after much discussion.) And there are a bunch where I
really don't know where the errors can have come from: typically cases
where there are *zero* bins in the output, and the code *really* looks
like there have to be some.

So we'll need to do another run of the Rivet2+YODA validation jobs based
on the current trunks of both Rivet and YODA: David, can you set those
running? And please have a look in the rivet/private/yodadiff hg repo to
see my comments on the datasets I checked.

It'd be *really* nice to get our new releases of Rivet and YODA out in
the next couple of weeks. Always very hard to keep things moving when
we're not in the same place and able to work *together*, but a bit of a
push now to *finally* get this release out will be a big relief (for me
at least) and free us up to work on more creative / enhancement
developments :-)

Cheers,
Andy

-- 
Dr Andy Buckley, Royal Society University Research Fellow
Particle Physics Expt Group, University of Edinburgh / PH Dept, CERN