search

nimble-dev / nimble

The base NIMBLE package for R
http://R-nimble.org
BSD 3-Clause "New" or "Revised" License
160 stars 24 forks source link

naming for AGHQ/AGHQuad #1465

Open paciorek opened 5 months ago

paciorek commented 5 months ago

@perrydv @paul-vdb @weizhangstats We have buildAGHQuad but runAGHQ and summaryAGHQ.

I'm somewhat inclined to change to buildAGHQ.

I think we want to decide on a single naming convention. If we decide today or early tomorrow, we could change for the release. Or we could wait for nimbleQuad.

paciorek commented 5 months ago

I should also note "AGHQuad_" is used internally.

perrydv commented 5 months ago

Sorry that's partly my fault on runAGHQ as a late addition. I like "AGHQ" without the "uad", but I could be convinced the other way. I agree consistency will be easier in the long run.

Some names like nQuad_ are used so that that can be an internal (member) variable name and thus leave nQuad available as an argument name. I don't see AGHQuad_ but if it's there I think that kind of reason would be why.

paciorek commented 5 months ago

Most of our internal naming of functions/methods and objects is of the form AGHQuad*. I think it would be ok to leave that and have only user facing be AGHQ, namely buildAGHQ and runAGHQ.

perrydv commented 5 months ago

Agreed, the topic now is user-facing names.

weizhangstats commented 5 months ago

buildQuad? AGHQ is of a particular type and the more generic name Quad for quadrature will allow us to build in other types if of interest?

paciorek commented 5 months ago

I think if we added other types, we could then create a buildQuad as a wrapper. I'm inclined not to do that at this point.

paul-vdb commented 5 months ago

I think it's splitting hairs and agree with buildQuad big picture but buildLaplace and buildAGHQ as a wrapper in the future works just fine. I vote for buildAGHQ for this release. Remove the uad :)