Should we parse the challenges in Python at scrape time?

[benoit74]

Currently, challenges are parsed in the UI, with JS.

I tend to consider this is a bad thing because:

• it consumes resources (parsing Markdown ...) client-side, while we know some clients are not really high-end machines
• it means that should parsing fail, it will be discovered very late (we could even create completely broken ZIMs without noticing if the format changes on freeCodeCamp side)
• it avoids from making informed decision on what to include inside the ZIM (e.g. remove unsupported challenge types, fail the scrape if there are too many, ...)

What I don't know yet is why freeCodeCamp chose this markdown format. Is it just more convenient / less verbose to write for course developers and an heritage of Gatsby framework they are using? Or is there any painful consequence when transforming it into JSON?

To be investigated