feature: allow complex external parameters

[Gauthier]

Dear MG team,

Wouldn't it be nice to allow for complex external parameters?

I would personally find it useful and believe it only requires minimal changes in MG code.

The few changes listed in the enclosed diff may provide a basis for the implementation of this feature. I don't claim this to be fully complete but checked already:
- MG doesn't crash any longer at runtime when a UFO parameter has nature='external' and type='complex'.
- the param_card can then possibly include complex values (for parameters declared as such) of format '(1D-02,2D-2)'
- the 'set' feature of the ME interface accepts a '1e-2+2e-2j' syntax (without any space)

What is your opinion about this?
Could it lead to conflicts I haven't thought about?

Cheers,
Gauthier

[Olivier Mattelaer (olivier-mattelaer)]

Hi Gauthier,

The main reason why I never implement this is that the LesHouches convention agrees that all external parameter should be real.
Which I agree is slightly ridiculous for a bunch of parameter which are therefore split in two entry in that input_file (Real/Imaginary).

So a model with nature=‘external’ and type=‘complex’ is intrinsically a model which didn’t fit the HEP standard.
And therefore, I have no idea how much support this kind of model will have outside MG.
(especially this can make the shower program to crash)

Now, I’m ok to include those changes but I’m not sure that really a good move due to this compatibility issue.
Fabio what do you think?

Olivier

On Feb 18, 2014, at 4:00 PM, G.Durieux <email address hidden> wrote:

> ** Patch removed: "basic set of changes needed to allow for complex external parameters"
> https://bugs.launchpad.net/mg5amcnlo/+bug/1281596/+attachment/3985330/+files/complex_external_parameters.diff
>
> ** Patch added: "basic set of changes needed to allow for complex external parameters"
> https://bugs.launchpad.net/mg5amcnlo/+bug/1281596/+attachment/3985425/+files/complex_external_parameters.diff
>
> --
> You received this bug notification because you are subscribed to
> MadGraph5_aMC@NLO.
> https://bugs.launchpad.net/bugs/1281596
>
> Title:
> feature: allow complex external parameters
>
> Status in MadGraph5_aMC@NLO Generator:
> New
>
> Bug description:
> Dear MG team,
>
> Wouldn't it be nice to allow for complex external parameters?
>
> I would personally find it useful and believe it only requires minimal
> changes in MG code.
>
> The few changes listed in the enclosed diff may provide a basis for the implementation of this feature. I don't claim this to be fully complete but checked already:
> - MG doesn't crash any longer at runtime when a UFO parameter has nature='external' and type='complex'.
> - the param_card can then possibly include complex values (for parameters declared as such) of format '(1D-02,2D-2)'
> - the 'set' feature of the ME interface accepts a '1e-2+2e-2j' syntax (without any space)
>
> What is your opinion about this?
> Could it lead to conflicts I haven't thought about?
>
> Cheers,
> Gauthier
>
> To manage notifications about this bug go to:
> https://bugs.launchpad.net/mg5amcnlo/+bug/1281596/+subscriptions