gistfile1.txt

Extending NewGRF APIs - impedance mismatch

Term borrowed from https://en.wikipedia.org/wiki/Object%E2%80%93relational_impedance_mismatch

In no particular order

* not many OpenTTD devs familiar with grf spec
    * so designing and reviewing can be hard
* some of the requests from grf authors are misguided
    * interacting with newgrf authors can involve blah blah politics
    * but sometimes multiple misguided requests point the way to one solid spec extension
    * nobody writes nfo any more, so very few grf authors actually understand the action0/1/2/3 spec, even less 6/7/9/D
* getting test grfs is a PITA (see 'nobody writes nfo any more')
* nmlc is easy to patch, but the inertia of testing is high
    * requires patched openttd
    * requires patched nmlc
    * requires test newgrf (see also 'docs')
* testing some grf features requires savegames
    * savegames are prone to breaking when testing both OpenTTD patches and grf changes  
* docs
    * we rely on maintaining grf spec and docs in a wiki
    * we haven't found a convention to include docs with OpenTTD or nml PRs
    * this means we often ship without docs done (maybe fine?)
    * this means that it is hard for newgrf authors to write test grfs as there limited or no docs