Implement subevent-ary observables

[sschmidt]

Up to now, all observables are unary or binary. Implementing observables that take subevents as their argument would make it possible to use event shape observables such as thrust easily from within SINDARAIN, like

analysis = record thrust_histogram (eval Thrust [@evt])

[Juergen Reuter]

This is a SINDARIN construct that needs to be documented etc. and affects directly the way the user steers WHIZARD. This should be done very quickly....

[kilian]

'm implementing user-defined event-shape functions (cf. #429). Event shapes are not uniquely defined, so I haven't yet implemented them natively in Sindarin. We should discuss further.

[Juergen Reuter]

Maybe there was a misunderstanding during the phone conference but observables likes thrust or MT2 should be incorporated into the WHIZARD core, and not only be available via the external user.f90 file. Please keep that task wish/feature request into your mind when working on that!

[kilian]

Yes, if we agree on the definition (and on an algorithm for computing them). Technically adding them is trivial now.

[kilian]

Please, could you give a list of important observables we could agree on? Thrust I heard already (reference for a good algorithm how to compute that?), more?

[Juergen Reuter]

Ok, DW, SS, JRR will sit together (hopefully on Monday) and make a list. Possible candidates are event shape variables, mass edge and kinematics variables as well as jet algorithm variables.

[dwiesler]

Ok, here's my short list:

E_T M_T M_T2 Sum E_T H_T Slashed H_T Alpha_T and M_eff

The definition and some explanatory words can be found in the attached variables.pdf.

[dwiesler]

variablewishlist

[sschmidt]

here are my suggestions:

• Thrust, Thrust major and Thrust minor and Oblateness, Hemisphere broadenings.

The definitions are given in my paper, an algorithm for maximization for thrust is given in the PYTHIA manual, section 15.2.2. The algorithm can also be used for thrust major. All other observables mentioned can then be calculated.

• Jet resolution scales

Jet resolution scales can be implemented using the KTCLUS package (also used in the matching).

[Juergen Reuter]

This is mandatory for #67.

[dwiesler]

Replying to sschmidt:

• Jet resolution scales

Jet resolution scales can be implemented using the KTCLUS package (also used in the matching).

Imho this should be made independent of the JetAlgo?, i.e. #67 is connected to this.

[Juergen Reuter]

Replying to dwiesler:

Ok, here's my short list:

E_T M_T M_T2 Sum E_T H_T Slashed H_T Alpha_T and M_eff

The definition and some explanatory words can be found in the attached variables.pdf.

Going back to that list, we have an implemetation of MT2 by Thomas James. It just has to be attached to the core after a clean up.

[Juergen Reuter]

THis will be done after #544.

[kilian]

P2 because it depends on #544.

[kilian]

Part of #544. The observables could be hardcoded, but the plan is to also enable genuine user functions coded in Sindarin. Some observables might require external native code, though.