[Rivet] ATLAS analysis to be included into Rivet

[Roman Lysak]

  Hi Andy,


On 08/16/2013 03:08 AM, Andy Buckley wrote:
> Hi Roman (and Kiran),
>
> Just to let you know, we are finally *nearly* ready to make our
> long-awaited Rivet 2.0.0 release: just a few more analyses need code
> tweaks and then we will be very happy to get that version out! We'll
> issue a final 1.x release shortly afterward, and a 2.0.1, to include the
> most recent analyses you've sent us. Thanks for those, and apologies for
> not replying earlier.
thanks for letting us know.
Up to now, I've been always preparing  1.X  release compatible analyses 
(i.e. aida compatible).
I guess we will probably still prepare a few analyses using 1.X release  
before moving to 2.X (for the validation, I think we will need to 
include 2.X rivet release into some release of ATLAS software, you know 
this better than me :) ). Hope this will not be a big problem for you.


>
> Since we're thinking about these updates now, I wondered if you had had
> a chance to look at updating this FB correlation analysis to use event
> weights?
yes, I updated the analysis to use the event weights and sent it to 
rivet mailing list in one of my emails on Aug 5.
So, please use the updated version and then I think it should be marked 
validated.

Cheers,
   Roman




> Otherwise I guess I will add it to the new releases as
> "unvalidated".

>
> Thanks again,
> Andy
>
>
> On 23/04/13 14:34, Roman Lysak wrote:
>>   Hi Hendrik,
>>
>>
>> On 04/18/2013 04:07 PM, Hendrik Hoeth wrote:
>>> Hi Roman, hi Kiran,
>>>
>>> Thus spake Roman Lysak (lysak at fzu.cz):
>>>
>>>>> I'm having a brief look at the code right now. Have you tested it with
>>>>> weighted events? You do fill a weight vector _vecWeight, but never use
>>>>> it. The regression seems to assume equal weights for all events. Is
>>>>> that
>>>>> correct?
>>>> That is correct.
>>>>
>>>> The azimuthal part of the analysis includes the weights.
>>>> The fwd-bkwd part of the analysis does not include the weights.
>>> We've discussed this a bit, but we are very hesitant to put the analysis
>>> into Rivet as it is. Weighted events get more and more common, and there
>>> are plenty use cases for them (e.g. tuning an NLO generator). So at very
>>> least we would put a mechanism in place such that weighted events would
>>> throw an exception -- which of course would render that part of the
>>> analysis useless for weighted runs.
>> fine, we will try to update this so it properly account for weighted events
>>
>>> Therefore we would very much prefer a solution where the analysis has
>>> proper support for weights. How much effort do you think that would be?
>> not sure, I suppose we will not get help from the original authors of
>> the analysis, so probably I will have to do it by myself. I will let you
>> know.
>>
>> Cheers,
>>    Roman
>>
>>> Cheers,
>>>
>>>      Hendrik
>>>
>> _______________________________________________
>> Rivet mailing list
>> Rivet at projects.hepforge.org
>> http://www.hepforge.org/lists/listinfo/rivet
>>
>