Confused about the state of the Geometry module

[DJDuque]

A lot of effort has gone into the ACTS documentation (this is greatly appreciated), but it is also a lot of information and I am a bit overwhelmed on "How to get started from scratch".

From what I understand, the very first step would be to define my geometry (right?). This would include the description/placement of both sensitive and passive elements (right?). Is this a reasonable (standalone) first step that I can focus on before looking at e.g. NagivatorState, Delegate, and all the tracking/vertexing?

I see mentions of Gen1, Gen2 (layer-less?), and Gen3 Geometry. I would be happy to play around with the latest/experimental ways of defining a geometry. Are these the things I should be looking at i.e. Acts::Surface, Acts::Experimental::Portal, Acts::Experimental::DetectorVolume, and Acts::Experimental::Detector? It is not clear to me if these are Gen2, and then there is a separate Gen3 way not documented yet?

Also, in terms of putting all the geometry together, I see "2 ways of doing it":

1. In one approach, I should look at the Acts::ProtoDetector and Acts::KDTreeTrackingGeometryBuilder. Is this Gen1/Gen2?
2. In the other approach I should look at the Acts::Experimental::Blueprint. Is this Gen3?

Basically, to summarize: I want to play around with the latest/experimental (Gen3?) ways to define my geometry.

1. Is defining the detector geometry a reasonable standalone first step?
2. Are the classes here what I should be focusing on/understanding?
3. Is the Acts::Experimental::Blueprint what I should be focusing on/understanding to put all this together?

Thanks

[paulgessinger]

Hey @DJDuque,

in the end it's sort of an in-progress evolution from the original (Gen 1) geometry model based on the ATLAS one, to an experimental first layer-less (Gen 2) model, to a unified (hopefully) final Gen 3 model. At this point, for new developments, I would encourage you to look at the Gen 3 version.

I have a relatively comprehensive recent set of slides about the geometry model here.

If you want, we can get on a call and I walk you through what we have? Shoot me an email if you want to do that!

[DJDuque]

Hi @paulgessinger

Thanks for the link to the set of slides, they are very helpful.
There is also a bunch of useful presentations/materials in that indico event. I'll be slowly going through that and I'll reach out when I have more questions.

Thanks again.

[asalzburger]

To make it less confusing, I will remove all Gen2 parts that are not related to the detray conversion, which is the main reason to keep the Experimental branch alive.