Functions library update and questions

[Jonah Gabry]

Before I go ahead and create the repository for the user-defined functions library, I want to find a decent name for the repository. Any preferences from the ones below (I'm not sure about them) or other suggestions?

user-function-library
function-library
stan-function-library
stan-functions
stan-language-functions

[Jonah Gabry]

Also, my current thinking and initial plan is to divide the functions into four high-level categories:

lp: Functions that include statements incrementing the log probability
dist: Functions that compute (primarily the logarithms of) PDFs, PMFs, CDFs, etc.
rng: Functions that generate (pseudo) random numbers
other: Other functions (e.g. for working with Stan's various data types)

If/when we move to using _lpdf, _lcdf, etc. it would also be natural to subdivide the "dist" directory into lpdf, lcdf, lccdf, etc.

I'm also thinking about structuring the "other" directory in some way, but I'm not sure if it should be by return type (e.g. other/matrix, other/vector) or topic (e.g. other/subsetting).

Please feel free to tell me these are bad ideas (preferably also suggesting better ones)!

[Aki Vehtari]

I would guess that if someone would contribute functions for, e.g., generalized Pareto distribution, it would be better to keep gpareto_lpdf, gpareto_cdf, gpareto_rng, etc. together for potential code sharing between those functions and for easier copying to a new model. So I would have all distribution related functions in one place.

Aki

[Bob Carpenter]

I like "stan-functions" or "stan-function-lib", but am not that
picky among these options. I don't think it needs to say
"user" or "language".

- Bob

> --
> You received this message because you are subscribed to the Google Groups "stan development mailing list" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to stan-dev+u...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

[Daniel Lee]

+1 for stan-functions

[Jonah Sol Gabry]

stan-functions works for me. Unless I hear from anyone else with a different preference in the next day or two stan-functions it shall be. 

You received this message because you are subscribed to a topic in the Google Groups "stan development mailing list" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/stan-dev/z7AmVBqLchU/unsubscribe.
To unsubscribe from this group and all its topics, send an email to stan-dev+u...@googlegroups.com.

[Jonah Sol Gabry]

Aki, yes that makes sense. Thanks.