FInalize structure function implementation

[Juergen Reuter]

Here are the currently open tasks, to complete 2.1.1 support: [ 1. Circe1 with ISR for resonance/on-shell production ] [ 2. Circe1 generator mode ] [ 3. Circe1 photons ] [ 4. EPA (this should work alone and in combination with CIRCE1) ] [ 5. All e-p collision setups ] [ 6. Reenable (but don't delete or disable) Circe2 (and endow with warnings), JRR insists on that. Should work the same way as in 2.1.1. ] [ 7. Beam events ] [ 8. energy scans ] Not really in 2.1.1

1. EWA (also w/ Circe, ISR, PDFs) (should work the same way as in 2.1.1.)

[

1. Check the status of non-collinear radiation (this last point has gone to a different ticket, #631).

]

[Juergen Reuter]

QED ISR is mostly done in r4576. Still, it has to be connected to the rest of the program for proper steering, but that should be more or less straightforward. No numerical comparison with W2.1 yet.

Next, I will go to EPA.

[Juergen Reuter]

EPA and EWA implemented (incl. unit tests). But all (incl. mappings) has to properly connected with the rest of the program.

[Juergen Reuter]

Next, I will do beam events.

[Juergen Reuter]

No, at the moment I do energy scan, then beam events, the CIRCE1 and CIRCE2.

[Juergen Reuter]

As discussed with WK, I will move on to this task, continuing with the beam events.

[Juergen Reuter]

Spectra seem not to work as intended. Or I don't understand them properly. Refraining from this ticket. :((( Also I don't know how to properly translate this affects_beam variable. Lost in translation.

[kilian]

I don't change the severity, but to be done after current bug fixes.

[kilian]

Next: try to re-enable CIRCE, along the lines as discussed in WÜ.

[Juergen Reuter]

Point 5: just stumbled over it in the HERA DIS example.

[kilian]

r5470: ISR for e+e- is up and running.

Actually, this is partly a new feature since it works and is efficient also if there is an s-channel resonance, or a single on-shell particle. This didn't work, or performed very poorly, in previous Whizards.

Next: CIRCE+ISR

[kilian]

r5489: Enabled Circe+ISR (mapping mode), including resonance and on-shell.

The performance is at least as good as 2.1, sometimes better. Next: Circe generator mode (which should become the default again).

[kilian]

Generator mode (w/ and w/o ISR) works again in r5511.

[kilian]

... caveat: resonance+on-shell versions to follow soon.

[Juergen Reuter]

The last big open point together with the MLM matching issue.

[kilian]

Status as of r5533: items 3 to 5 done.

[kilian]

Beam events work again in r5540.

The beam-events file is searched for first in the working directory, then in the system directory.

[kilian]

Energy scan is enabled again (r5540)

[kilian]

Circe2 re-enabled in r5550.

As discussed on the phone, it is restricted to gamma-gamma collisions and uses generator mode. Which implies that on-shell 2->1 processes are excluded.

If there is an actual request, it should be straightforward to extend this minimal setup with suitable mappings, in analogy with Circe1. However, there is no photon collider spectrum for any ILC design.

[kilian]

The remaining tasks are still valid for 2.2.0, but of lower priority.

[Juergen Reuter]

Most of the EWA is done in r5685. What is not yet resolved are the mappings of multiple structure functions (like ISR+EPA/EWA and CIRCE1+ISR+EPA/EWA), and the concatenation of "flavor-changing" SF like e.g. PDF + EWA.

[Juergen Reuter]

As said above, the only open point here is the concatenation of "flavor-changing" SF like PDF+EWA (which one would like to have). So, two things: 1) flavors should be set by dispatch_sf_config and dispatch_sf_data, 2) there should be a possibility to have an array of beam particles (at least for structure functions like EPA, EWA, maybe ISR). Directing the baton to WK again.

[kilian]

status r5766: the machinery of flavor(-sum) tracking works, but is probably not yet useful since all quantum numbers in the flavor sum must be uniform. I'll extend the interface for EWA etc. as to support flavor-dependent matrix elements in the structure-function interaction. When this is complete, the ticket can be closed.

[kilian]

With r5772, multiple flavors with non-uniform charges are allowed for EPA and EWA. In the EWA case, the EWA_mass value is now always used, since otherwise one could not sum over b (-> tW) in the PDF array.

Non-uniform charges are still excluded for ISR, since the charge enters the kinematics mapping there. The combination PDF -> ISR is nonsense anyway, physically.

This should settle the issue. Further possible refinements are in #631 and #562.