Hardcode the mega menu

[wpears]

The mega-menu has an entire backend up stood up to support editing its contents in Wagtail. However, this isn't really something that happens often enough to justify the extra complication. Moreover, by removing the mega-menu from Wagtail, we create the opportunity for separately deployed applications to include it without a bunch of code duplication + effort.

[wpears]

@csebianlander I've included two approaches herein:

• simply saving the mega-menu data as a jinja variable
• hardcoding the actual html that gets produced from the python jinja engine

I think it's pretty inarguable that the first approach is easier to edit, that said, the second approach is interesting to me in that it paves the way for a self-contained app, deployed statically and built with any technology that can ingest html, to also use the mega menu. One might imagine this html being distributed via a design-system-like package which both the wagtail website and any other self-contained apps would be consumers of.

Going a step further, one might imagine that this packaging system (say, cfgov-templates) could be authored similiar to approach 1, but then distribute the results of approach 2. That is, instead of the jinja engine running at request time, it runs at build time when producing the package output.

What do you think about this? Obviously this would require a bit more plumbing to set up, but would the authorship/editing experience be worth it? That is, is editing a jinja-like (or possibly json) representation a substantial improvement over editing the raw html? Is editing such a representation on par with the current experience for ease of use?
@anselmbradford curious about your thoughts on the above as well

[csebianlander]

That is, is editing a jinja-like (or possibly json) representation a substantial improvement over editing the raw html? Is editing such a representation on par with the current experience for ease of use?

Personally, I think that yes, the jinja-like/json approach (based on the file you linked) is vastly easier to parse and much less likely to be accidentally broken in the process of editing. It's probably roughly on par with the existing/Wagtail-based approach all things considered.

I hear you re: the advantage of the hardcoded approach but would caution that if we went that way it's far likelier that updating the mega menu in the future requires developer intervention to a degree I'm not comfortable with at this point.