New package: FixedRNGs v0.0.1

[JuliaRegistrator]

• Registering package: FixedRNGs
• Repository: https://github.com/cscherrer/FixedRNGs.jl
• Created by: @cscherrer
• Version: v0.0.1
• Commit: 167a785e19492d5799b3f16cfece27dd6fbfd372
• Reviewed by: @cscherrer
• Reference: cscherrer/ConstantRNGs.jl@167a785#commitcomment-120649593
• Description: A simple deterministic RNG, for cases when you need high efficiency and don't care about randomness

[github-actions]

Your new package pull request met all of the guidelines for auto-merging and is scheduled to be merged when the mandatory waiting period (3 days) has elapsed.

Since you are registering a new package, please make sure that you have read the package naming guidelines: https://julialang.github.io/Pkg.jl/dev/creating-packages/#Package-naming-guidelines-1

---

If you want to prevent this pull request from being auto-merged, simply leave a comment. If you want to post a comment without blocking auto-merging, you must include the text [noblock] in your comment. You can edit blocking comments, adding [noblock] to them in order to unblock auto-merging.

[gdalle]

[noblock]

Hi @cscherrer! Judging by the name I was expecting some Quasi Monte-Carlo scheme, but looking at the code it's just a fixed number every time, so I was a bit surprised 🤣
Can I suggest a renaming to FixedNumberGenerators? The R really doesn't belong here IMHO

[cscherrer]

[noblock]

Thanks @gdalle . I agree the name can be improved, but I think the name should either indicate that it's <:AbstractRNG, or something about the main use case. FixedNumberGenerators doens't meet either of these IMO. How about either

• TrivialRNGs or ConstantRNGs, or
• TestValues. Our main use case is the testvalue function, providing an arbitrary point in the support of a measure. The name is taken from PyMC, so it's maybe a little more familiar. In this case we'd bring the testvalue function here (currently it's in MeasureBase).

@oschulz any thoughts on this?

[gdalle]

[noblock]

How about FakeRNGs?

[cscherrer]

How about FakeRNGs?

...maybe? Let's wait for input from @oschulz. There's no rush on this, just a small component of MeausureBase we realized could be independent, so we've talked about factoring it out for wider use.

[noblock]

[oschulz]

Hmmm there are fixed, but also not random, but they are (and have to be) subtypes of AbstractRNG. Maybe something like "DummyRNGS"?

[noblock]

[cscherrer]

DummyRNGs works for me. What do you think @gdalle ?

[noblock]

[gdalle]

[noblock]

I agree that it fits, but I must say I don't like the connotations.

pandas-dev/pandas#35724

Perhaps ConstantRNGs.jl?

[cscherrer]

[noblock]

Oof, good point. Ok, let's go with ConstantRNGs, I think that captures everything we need. Thanks for the discussion @gdalle

[cscherrer]

@gdalle I re-registered as ConstantRNGs (#86926). Does that mean we should kill this one? Or keep this and kill the new one?
[noblock]

[oschulz]

I like ConstantRNGs too.

[noblock]