Establish schema for metadata useful for aligning to training courses etc

[yarikoptic]

• audience-level: {novice, intermediate, expert}
• speaker-level: ... well -- we do not want to "judge" or upset anyone. But it could be along "student", "researcher", "professor" scale?
• repro-steps: [12345] or verbalized ;)
• category: some ontology to group them up
• software: List(str) - to list relevant software the talk touched upon etc. Might be auto-populated from Do OCR across all videos #2 and Fetch/update per video metadata #3 ;)
• datasets: List(str) - similar to software but accent on data sources

If such metadata is (also or just there) stored at the level of git-annex, with git annex view we could very quickly establish alternative or custom hierarchies, e.g. git annex view audience-level=novice repro-steps/title to populate a hierarchy with top level being the reprosteps and then videos for the "novice" audience.

For this metadata we should establish an easy process to receive "updates".