`particleName/`: Clarify it is arbitrary

[DavidSagan]

It may be desired to group particles in other ways besides particle type. For example, by what bunch they are in when particles are bunched. So particle groups should be allowed to use names other than the particle species type.

@ax3l: Please put this in project 2.0

[ax3l]

Thank you for the question!

We currently do not impose that a particle name needs to describe the species type. It can be anything and it not interpreted in the base standard.

In our applications, we also have multiple particle groups for various selections of the same physical particle species, e.g. filtered by origin/generation (e.g. which shell of an atom emitted an e- in a plasma), energy, bunch-number, pointing, etc. We currently just encode this by unique names.

The idea we discussed recently in our VC for an "physical particle" extension is to add an attribute to the particle group with reference to fundamental physical particle species types defined in the "physical particle" extension. I think this would cover this issue as well and since we did not open a new issue for it yet, let us use this one! :)

The idea of this extension would be to define physical

• fundamental particles
• ion species
• maybe quasi-particles

just in a list and as soon as the extension is advertised, a new attribute in each particle group is required referring to one of the defined physical particle species names of the extension. The name can with that still be arbitrary.

Side-effect: related to #42 in case one dumps some species multiple times in varying names.

[DavidSagan]

@ax3l

We currently do not impose that a particle name needs to describe the species type.

I found this in the standard:

Each particle species shall be represented as a group particleName/ that contains all its records. Particles records are generally represented in one-dimensional contiguous records, where the n-th entry in particleName/recordNameA and the n-th entry in particleName/recordNameB belong to the same particle.

This is confusing since it seems to imply that particleName be the species name. If the aim is for the standard to not describe the species type then the above should be modified to make this clear. How about:

Particles are divided up into sets. Each particle set shall be represented as a group particleSetName/. For example: bunch1/. Particle sets are to be homogeneous in terms of the particle species. That is, a given set only describes particles that are all of the same species. records are generally represented in one-dimensional contiguous records, where the n-th entry in particleSetName/recordNameA and the n-th entry in particleSetName/recordNameB belong to the same particle.

[ax3l]

After having some time to think about it, we could make this extension even more flexible.

The extension will basically define a name of lists and an optional attribute that can be used. I would not not only allow the attribute on particle records but on generally all records. This allows us to e.g. also use it on meshed quantities, such as phase spaces, densities, etc. :)

[ax3l]

Example: let's allow a general group string-attribute witch all records, including meshes and particles.

The group can be user-given and just needs to match to define a group between arbitrary records. If we define again ; as a special character, we can allow to add a N:M relation of records in user-defined groups with this attribute. Automated reading can just build a map of groups to lookups of record names.

[DavidSagan]

@ax3l:

There seems to be two topics here. One is my proposal to simply clarify the standard (this is, simply change the wording in the standard) and the other is your thoughts about adding a group attribute. Would it not be better to split these into two separate issues?

[ax3l]

I might be a little confused on this one, we can discuss it in the VC in a few minutes :)