Naming resources

General
1

I was thinking we can add “namespaces” to the resource tree similar to how it works in logging

grandparent.parent.child

each resource has to have a unique name in the name space, but individual resources can have their own name

so like carrier.site0.plate.A1 for example

still globally unique since every child of a given parent has a unique name, but two resources can have the same local name (like two wells in two different plates named A1)

1 Like
2

pros:

  • scales better to big workcells
  • shorter names when looking locally
  • potentially a way for regex style queries later

cons:

  • full resource name depends on its current location
  • full name is very long
1 Like
3

I like the idea of just having unique strings as the only constraint, mainly because as you scale you get situations where the grandparent doesn’t really matter.

robot1.plate2.A1 -> robot2.plate2.A1

Doesn’t matter if plate2 is unique within the namespace of robot1, we want it to be unique everywhere in the lab.

2 Likes
4

Q1: Is this proposal…
(A) in addition to the current ‘single-string defines resource name’ approach or
(B) instead of it?

Q2: And is this for…
(A) auto-naming of resources, e.g. wells, or
(B) PLR-user defined resource naming, e.g. carriers, plates, tipracks?

(What would Q2.B look like?)

I agree, every resource is currently only limited by what it is called, this is very nice and enables very dynamic systems in which resources move around a lot.

The current system also aligns with barcoding really well.

I am not sure I understand this proposal:
Do you want to allow having to resources being named the same in the same workcell?

1 Like