do we store the actual data in the schema

I would think this is reasonable (e.g., it is a bit inconsistent to store BP/MP.. but not the NMR), but i think there are also reasons to keep it in the JCAMP

if we keep data in the JCAMP, do we describe what the columns mean in the schema? Latter is kind of weird as this is not really general and artificially constrains us. Still we will always need to allow for some attachments, e.g., for images/videos

Start by reviewing the schema and JCAMP data representations, including how BP/MP and NMR data are currently handled. Compare the trade-offs of storing actual data in the schema versus retaining it in JCAMP, and define how column meanings and image or video attachments should be represented; done means the storage approach is decided and documented.

Written by the indexing model from the issue text.