- HTML takes 3x the space of Markdown. Even though Markdown would be rendered client-side, it doesn’t quite mean 3x advantage in transmitted payload because gzip would close the gap.
- HTML will also require
textContentreturned separately to be indexed for full-text-search, whereas Markdown only uses non-alphanumeric characters which would be omitted by the tokenizer. This means there would be two columns of the body in the database: one with the full HTML for returning, and another with all the tags stripped out for indexing.
- IMO markdown is actually quite frustrating to use. A lot of it has to do with it trying to “help” and getting in the way instead. For example, sometimes I see a giant wall of text on Reddit, but viewing the source reveals that OP actually made paragraphs with single line breaks.
- There is no true WYSIWYG editor for Markdown. I’ve only found editors with a live-preview that call themselves “WYSIWYG-like”, which IMO is too cumbersome to use – exhibit 1, people often edit a post on StackExchange just to fix their Markdown.
- Related: https://ux.stackexchange.com/questions/15919/is-markdown-friendly-enough-for-non-technical-users
I don’t know if it’d be feasible to convert from one to another later on, so I would like to make the correct decision at the start.