[Rivet] Many empty bins in refdata for JADE_OPAL_2000...

[Andy Buckley]

On 22/06/10 22:17, Holger Schulz wrote:
> On 22/06/10 22:23, Frank Siegert wrote:
>> Holger Schulz, Saturday 19 June 2010:
>>    
>>>> Thanks for volunteering, but if there are no serious drawbacks of
>>>> the zeros, I would suggest not to, because
>>>>
>>>> 1) We shouldn't deviate from Hepdata without need.
>>>> 2) (Existing) AIDA files from the existing Rivet versions can't be
>>>> plotted together with new ones then because the binning has
>>>> changed. 3) Some of the bins without reference data might still be
>>>> filled by (different) Monte-Carlo runs and thus might be useful
>>>> for comparison.
>>>>        
>>> Ok, then lets leave it as it is.  :)
>>>      
>> Seems like Andy went ahead and changed this in r2501.
>> (I just learned it the hard way when my existing and new aida files
>> weren't plotting anymore due to binning conflicts)
>> It's a little bit annoying, but I guess even if just for my point 1)
>> above, we can't avoid it.
>>    
> I am in favour for consistency. Maybe the refdata is going to be
> updated in HepData as well. I don't see a good reason for having
> empty bins in the ref-data. In other cases, like the ALPEH_1996,
> there are bin missing which should be zero bins as well, following
> the logic of the curent Hepdata version of the JAD_OPAL refdata.
> Maybe the zero bins are a relict of a conversion tool.

The zero bins were an artefact of an assumption that all axes in a
dataset had the same number of bins... so there was a loop which always
went to the length of the longest axis and wrote out null bins if the
bin at that index could not be found. It's now fixed, so the data file
with the removed null bins is the correct one. (HepData also now gets
the name right, i.e. JADE_OPAL rather than just OPAL)

Andy

-- 
Dr Andy Buckley
SUPA Advanced Research Fellow
Particle Physics Experiment Group, University of Edinburgh