Expose a way for the user to choose between QR/Cholesky with the formula interface?

[andreasnoack]

Generated from #500 to have a separate place for discussing this. My feeling is that we'd have to punt on this issue if we want to release 2.0 anytime soon so I'm tentatively putting this on the "out of scope for 2.0" milestone. Please comment if you disagree and plan to implement this soon.

[nalimilan]

Where was this feature proposed? Isn't it enough to select the method using an argument to glm?

[andreasnoack]

This was an item in the list in https://github.com/JuliaStats/GLM.jl/issues/500#issue-1390221148. Not sure if it has been discussed elsewhere. I just converted it to an issue to have a place to clearly decide on this. @palday since you create the linked issue you might know more. Do you think this is something worth considering? IMO it is not something we need, but even if we wanted this, I'm not sure that we should include it in 2.0 since it would probably take some time to implement.

[mousum-github]

I am not in favour of this request, as we can fit lm/glm models without using a formula as well.

[nalimilan]

I propose closing this unless @palday really wants it. ;-)