Literature cannot be forked the way we think about it in code. My variation of the potpie isn’t simply a fork of the ingredients and editing lines. It’s “likely” a different style altogether and in most cases required the new author to write something ground up. We tried fixing this but the edge cases became too complex to handle.
There's a few recipes I've used for which this is the case. Same for my mum - I've got magazine clipped recipes of hers that say something along the lines of "add cherries at the end, use sultanas instead of currants; I add a dash of brandy" or others that say "you need to cook for about 45 mins, not 30".
I think variations allow for a sort of evolution of a recipe towards your own personal preference.
But of course that doesn't take account of the actual implementation of such a feature being possible to do in a way that works well and fits with the aims of this site.
I'd be interested to know what sort of edge cases make it complex. Surely it's just a pre-filled version of a page, that one edits and optionally has a link saying "based on $recipe". I could see it getting abused for people wanting to link their recipes to popular ones (like youtube response vids that aren't really responses) but beyond that sort of thing it seems pretty basic. Were you trying to digest the multiple variations to offer options within a recipe or something?
We have a wish list in our mind where we can add structured annotations (tips/alt ingredient/photo) to the original literature and have a repository of such stuff residing on top of the original recipe itself.
This way it can become a far more richer data, set in the original context and offer mutual learning to all.
Thoughts?