[FIX] clarify MEG empty-room recording naming conventions

[sappelhoff]

this section was unclear, see mne-tools/mne-bids#421 (comment)

This PR aims at improving / clarifying the used language and examples.

---

I am wondering, why noise as a task name and date as a session label have only "SHOULD", and not "MUST" definitions 🤔 anyone got a clue?

[hoechenberger]

I am wondering, why noise as a task name and date as a session label have only "SHOULD"

Maybe to allow omitting them entirely? And the intention was not to allow arbitrary values?

[jasmainak]

Maybe to allow omitting them entirely? And the intention was not to allow arbitrary values?

But they cannot be ommitted. Task label is a required label.

[jasmainak]

why noise as a task name and date as a session label have only "SHOULD", and not "MUST" definitions 🤔 anyone got a clue?

If you ask me, SHOULD is pretty much OPTIONAL because validator does not check for it :) 30% of the spec is checked in the validator.

[hoechenberger]

But they cannot be ommitted. Task label is a required label.

Maybe to allow omitting them entirely? And the intention was not to allow arbitrary values?

But they cannot be ommitted. Task label is a required label.

That's true. But session is optional…

[hoechenberger]

I've made a few suggestions, feel free to apply or ignore them :)

[hoechenberger]

I made some suggestions regarding phrasing and markup, feel free to apply to ones you like and leave out the others.

[sappelhoff]

thanks @hoechenberger :)

[sappelhoff]

Thanks for your reviews @jasmainak @robertoostenveld and @hoechenberger

before I merge I just want to make sure you all saw the final version and approve of it.

[hoechenberger]

LGTM!

[robertoostenveld]

looks good, thanks!

[jasmainak]

sorry for the last minute comments! 🙈

[sappelhoff]

@robertoostenveld @jasmainak we got an answer on MRI phantoms: #480 (comment)

Also, I voiced my opinion here: #480 (comment)

would be cool if you could take a minute to respond, so that we can move forward with this PR.

[robertoostenveld]

I am happy with the modifications as proposed in this PR, as they align with what I recall from the in-person discussions we had about this in the preparation of BIDS-MEG. But I would also be fine to leave out the 2nd strategy (task-noise), as long as the 1st strategy (coding it as sub-emptyroom) does not become REQUIRED and remains RECOMMENDED. I don't think it is required that the 2nd strategy is actually explicitly documented; it is something that researchers can do anyway given the current specification. As such the original text that is now still online here is also fine with me.

Furthermore, I think that clarifications to the specification are possibly not always the best discussed as changes to the specification text. We have other channels (google group, neurostars, BIDS starterkit, brainhack mattermost) that I think are more appropriate for clarification questions, for documenting recommendations and use-cases, and for discussion things like those happening now here.

[sappelhoff]

Okay, I'll then strip down the proposed changes heavily so that the way emptyroom is currently described in the spec will essentially remain as is.

I'll take the other things we talked about (re: emptyroom) to a page in the starter kit

Thanks for your reviews all, I'll ping once more when I finished what I just said I'll do.

[sappelhoff]

I rewrote the changes and I think they are now mergeable.

I am not introducing any new notions (anymore), but merely:

• RECOMMEND the strategy, which so far was "just there" (not recommended or anything)
• reorder the text to make more sense (folder name, task, session, ... in order)
• capitalize should -> SHOULD
• adding explanatory sentences
• completing the example to also include the metadata, which was previously missing

[jasmainak]

this looks good to me. So now you are essentially just adding clarifying language to directly address @hoechenberger 's issue without mixing it with other issues. If there are new use cases (= data or code) that suggest another strategy is needed to store empty room files (as explained by @robertoostenveld ), I'd be happy to discuss this in a separate issue!

[sappelhoff]

Thanks for the discussion and reviews everybody. The section is a tiny bit better (more understandable and consistent) now :-)