Results for approval voting for Stan Autodiff/Math Library Name

[Allen B. Riddell]

Tie between StanADE and The Stan Math Library.

pdf attached, or you can see the results here:

http://doodle.com/intqfgg4kskkn6ainzm2hkww/admin#table

Best,

Allen

p.s. I'm not sure if Bob understood the poll used approval voting -- if
not, StanADE won.

[Daniel Lee]

I completely forgot about it -- I'd vote for Stan Math Library.

Daniel

--
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.

[Rob Trangucci]

Ditto, though I would have voted for StanADE

Rob

[Bob Carpenter]

What's approval voting?

- Bob

[Allen B. Riddell]

You vote for the options you support---as many as you like. It's easier
to do than ranked balloting and appears to satisfy voters empirically. I
think the IEEE and the MAA use it.

As you can see, it has some problems breaking ties.

Best,

-a

[Bob Carpenter]

OK --- I just voted for all the ones I could live with.

No big deal to me one way or the other.

- Bob

[Avraham Adler]

If you submit ranks, we could use Condorcet voting[1]. Implementing Condorcet + Schulz[2] isn't that hard; I have some R/Rcpp code that does it[3].

[1] <https://en.wikipedia.org/wiki/Condorcet_method>
[2] <https://en.wikipedia.org/wiki/Schulze_method>
[3] <http://gallery.rcpp.org/articles/condorcet-voting-with-rcpp/>

[Andrew Gelman]

No, we should give numerical ratings and then analyze the results in Stan using IRT.

[Bob Carpenter]

Let's kick this one to the Nominator to decide.
So what do you think, Andrew?

* StanAD
* Stan AD
* Stan Automatic Differentiation Library

* StanMath
* Stan Math
* Stan Math Library

I really don't like the "Engine", so would like to rule out
"StanADE", despite having a clean pronunciation.

The "AD" focuses on autodiff, which is why someone would use
it, whereas "Math" leaves it more general and inclusive.

We could have our cake and eat it too, using Stan Math Library as
an umbrella covering:

* autodiff
* special functions
* probability-related functions
* matrices
* diff eqs

- Bob

[Andrew Gelman]

Oooh, Stan Math Library, I like that.