Closed TomMD closed 4 years ago
If the goal is to have a machine-parseable output format then TOML is a more favourable choice, since we already depend on the tomland
library. Bringing new dependencies to the project will make the installation process slower for all Stan users, even if they don't need machine-parseable output.
OK, thanks for the direction. Making a full TOML is a bit more work both because tomland affords less automation and encourages a bidirectional codecs, while the envisioned JSON would be lossy.
I hacked up a JSON that will suffice until we get the toml and I also get a good toml/json translation external to stan. Will keep the toml task on my radar.
Stan lacks a machine-parsable output format, such as JSON. We should add one!