Metadata and Structure conventions

I am looking for general structure and metadata conventions (I notice the contradiction of “general” and “convention”).

Are you aware of more general conventions that only rely on general concepts like or
Are you aware of alternatives to NeXus?

Thank you, Steffen

This probably doesn’t answer your question(s), but have you looked at Allotrope? G.

Yes, I took a look. Thank you for making me aware of it.
It is similar or more specific than NeXus but they seem to keep all information to themselves as they are a closed consortium. I did not find any example.hdf5 (as opposed to NeXus where there are examples)
Thanks for the information, anyhow. I learned something so that is great.

Hi @s.brinckmann,

I think this is a very interesting topic. I don’t know what your time zone is but I plan to talk about this at the next Call the Doctor session on October 11, early afternoon US Eastern time. I will send you connection info if you are interested to join.


1 Like

Hey Alexsandar,
I would be very interested and am in i suitable timezone. I assume you send me a DM within the forum-system.
Best, Steffen

Hi @s.brinckmann,

Excellent! Here’s the registration page for the tomorrow’s session:

You should receive the meeting details after registering. I would like you to describe your use case.


Hi @s.brinckmann,

Please note the start time for tomorrow’s session moved up to 1:30pm US Eastern time.


Below is a summary of the discussion on this topic held during the Call the Doctor session on October 11.

  • Having a reasonable and minimal metadata convention for HDF5 data would be helpful, although not as impactful as it could have been if done much earlier.
  • There are many best practices and good starting points to consider.
  • Dataset, ESIP Science On Schema.Org, and Google Search Central Dataset are suggested as good sources for general metadata.
  • Any effort to create a widely applicable HDF5 metadata convention must be community-driven.
  • Describing physical units could be first to address given its importance for science and engineering data.
  • There is no time line when to start working on this.