cornquest/doc/retro

51 lines
1.7 KiB
Plaintext

2023-10-23
i think the story is "done"
i had grand designs for a very complicated narrative structure
but i ended up scrapping 99.9% of that in order to just finish it.
using rec as a database i think is fine
and i think i would do it again.
but somewhere around 50/60 nodes,
i started to become unable to keep the whole story in my head
all at once.
that is,
i would look at the graph
and couldn't immediatly map it to a part of the story
in my mind.
i think i could fix that by having the graph
include both the id and the name?
but also i suspect that at that size,
the graph simply isn't a great tool any longer
for assessing the shape and flow of the story.
the recfile itself started to become kind of had to navigate at that size too.
i created a tags file to help navigate,
and added some syntax highlighting,
and make some snippets,
and that helped a little bit with the cognative load.
but also i broke the story at least twice
by accidentally creating a node with a non-unique name.
so i think that if i were to do this again,
i made make a 'recins' wrapper
in order to enforce data integrity
instead of just creating data by hand in the recfile.
i also think that i might create a new schema
to include 'links' or 'branches'
because i think elevating that information to its own field/value
could make it easier to export the data into, say, twine
or some other kind of text driven engine.
there was at least one point
at which i wish i had the ability to create a 'data node'
and then query that for a value.
e.g. "if has_visited? then goto(id) else goto(other) end"
this might be possible in recutils 1.10
if jem is successful in adding scripting to recfmt.
in summary,
it worked well enough!
but started to become a little unwiedly toward the end.