[matja]

> At the time, raw format was the only thing that preserved all necessary information to make scientific observations.

Which "RAW" format? I think that is the problem - there are many and few are completely documented, which is where DNG comes in (https://www.adobe.com/content/dam/acom/en/products/photoshop...).

Great point about throwing away metadata, and probably worthwhile to safeguard that one's flavor of RAW files (CR2/NEF/etc) can be reliably read in the future when the software necessary to read them inevitably disappears from the cloud.