Effective W approximation

[kilian]

Implement the EWA (already exists as dummy).

[Juergen Reuter]

Structure functions now are non-zero and give reasonable values, but proceeding further is frozen because of lacking communication.

[Juergen Reuter]

OK, finally I stop this until we set up a more properly suited communication.

[Juergen Reuter]

Finally, I have the idea to handle W and Z structure functions differently.

[Juergen Reuter]

Ok, sometimes you get kicked in the ass by a different problem: the cross sections have been zero because it was not caught (as in W1) that the kinematical function lambda (of which the sqrt is taken) could become negative if sqrts_hat is smaller than the sum of the two incoming masses. Now EWA gives a non-zero cross section of the correct order of magnitude, but there is a 20 per cent difference to W1. Let's go debug.

[Juergen Reuter]

In r2158, EWA is technically working in the unpolarized case. However, there is no agreement with WHIZARD 1 even when I set the structure function identical to unity. This seems to mean that there is a problem in the correct assignment or normalization of the state matrix or a problem/difference in the assignment of helicities. The partonic cross sections agree. Postponing this to 2.0.1 (cum core grave).

[Juergen Reuter]

Maybe WK got enough experience with the polarized structure functions from CIRCE2 that we could put this ahead of 2.1.0 to 2.0.5? Please comment.

[Juergen Reuter]

This has been lingering around for a VERY long time now....

[kilian]

Conversely, I'd like to ask: Do we really need this? There has been no real demand, and the technicalities are somewhat nontrivial. Neither was it supported in W1.

It would make for a nice self-contained (bachelor) thesis, however.

[Juergen Reuter]

Well, it was supported in W1, and thats the reason. ALso, there might be some need for this. I know that it is a task where there is not much motivation to do it, but sometimes it hits you....

[Juergen Reuter]

TV is making considerable progress, plots are looking good. We can hope to close this rather soon :)

[Juergen Reuter]

The factor of (3/2) or (3/2)^2 is understood: it is a mismatch between the number of spin states given in the unit matrix element when using beams/structure functions.What is unresolved is the spread in values for the functional dependence on the structure function. TV testet it with a constant function, it is present even there. Maybe a bad mapping when using structure functions? Ideas?

[Juergen Reuter]

There are many interesting results from the work of TV as a summer student. We should discuss this at the next phone meeting...

[Juergen Reuter]

Milestone renamed

[Juergen Reuter]

Milestone renamed

[Juergen Reuter]

Transferred to Gitlab issue 329.