Suggestion to update label property type

[marcus-j-davies]

Very often I see config files that are in support of various devices, with nothing more than a difference in there label.
Eaton being just 1 (have just raised a PR for one)

I assume, the labels may change between regions.

Currently, it seems this is managed by simply adding the various labels as one string: "Label 1 / Label 2".
Would it be more "structured" to allow this to come in the form of an array also?

"label": "1 Label"
"label": ["Label 1", "Label 2"]

This already happens with the zwaveAllianceId key

The driver (when seeing this as an array) will just concatenate using / as the separator - as to not break front ends, or the indexing format.

I think this makes it align with the zwaveAllianceId property?

@blhoward2 - not sure if you have a view on this?

[AlCalzone]

I think this is basically #1675 (Move manufacturer labels into devices array), but for device labels. I guess it could make sense to be able to specify the device and/or manufacturer in the devices array instead of the root level. That way we could assign device labels to individual IDs without splitting the files.

[marcus-j-davies]

I'm probably not qualified enough, to offer my opinion down to that level.
But the likes of Jasco, Eaton, Yale + many more, have a lot of labels as one string.

Formatting (or at least, making them) a collection of some sort, I think will offer a cleaner interface, for label management.

[AlCalzone]

I also don't think we should also assume it's always the case they should go into one file.

I don't think this should be an incentive to do that, but if they are already, this could help.
https://github.com/zwave-js/node-zwave-js/pull/2689/files

like where Yale sells several models that all fingerprint the same.

In this case all we can do is structure the labels as we do right now.

Formatting (or at least, making them) a collection of some sort, I think will offer a cleaner interface, for label management.

I don't like making it an array just because we can, this just means we have to concatenate the labels again after reading the file. If we can say (like in your linked PR) that device ID 1 is label 1 and device ID 2 is label 2, with the device being identical, that's okay.

[marcus-j-davies]

Here is an interesting approach.

If a label appears for a set of ID's - it takes priority, if not, falls back to the root label?
This will work, for the current format, and the one depicted here #1675

or am I jumping ahead? - I have tendency to do that.

[blhoward2]

This seems overly complicated. If the devices are different enough they need different labels, and can be distinguished, they like have different manuals, etc. I’d rather them just be different files in that case. If they can’t be distinguished by the fingerprint then we’re just back to concatating them back together again.

[marcus-j-davies]

i.e - jumping ahead 😄