Skip to content

Partial Argument Matching in nlme::varFunc

3 messages · biii m@iii@g oii de@@ey@ws, Phillip Alday

Original
biii m@iii@g oii de@@ey@ws
# 
Hi,

 

Recently, I have turned on warnings for partial argument name matching, and
with that, I'm finding many warnings.  One is while summarizing models
within the nlme library.

 

An example that causes this is:

 

```

library(nlme)

fm1 <- gls(follicles ~ sin(2*pi*Time) + cos(2*pi*Time), Ovary,

           correlation = corAR1(form = ~ 1 | Mare),

           weights = varIdent(form=~Mare))

summary(fm1)

```

 

Which gives the warning:

```

Warning message:

In coef.varIdent(x, uncons = FALSE, allCoef = TRUE) :

  partial argument match of 'uncons' to 'unconstrained'

```

 

This appears to be fixed by the diff I put into an R bug report here:
https://bugs.r-project.org/bugzilla/show_bug.cgi?id=17767

 

I assume that this list has the right people to review and apply the patch.


Thanks,

 

Bill
1 day later
Phillip Alday
# 
Hi Bill,

I see that you've really been getting into the nlme source lately ....
as far as I know, nlme is in maintenance mode right now and pretty badly
neglected in terms of interest in actually maintaining it. Have you
considered offering to take on some of that stewardship? I'm sure R Core
would be happy to have someone actively engaged in keeping a core
package alive.

Best,

Phillip
On 25/04/2020 16:48, bill at denney.ws wrote:
biii m@iii@g oii de@@ey@ws
# 
Hi Philip,

:)

While revisions like the one here are straight-forward, I don't think that my knowledge of underlying algorithms is sufficient for me to actively maintain the package.

I try to be a productive bug-reporter, and where I can, I try to provide patches.  But, from my previous questions (about nlme::getVarCov and nlme::varFixed), I don't know what the correct behavior would be.

Thanks,

Bill

-----Original Message-----
From: Phillip Alday <phillip.alday at mpi.nl> 
Sent: Sunday, April 26, 2020 11:16 AM
To: bill at denney.ws; r-sig-mixed-models at r-project.org
Subject: Re: [R-sig-ME] Partial Argument Matching in nlme::varFunc

Hi Bill,

I see that you've really been getting into the nlme source lately ....
as far as I know, nlme is in maintenance mode right now and pretty badly neglected in terms of interest in actually maintaining it. Have you considered offering to take on some of that stewardship? I'm sure R Core would be happy to have someone actively engaged in keeping a core package alive.

Best,

Phillip
On 25/04/2020 16:48, bill at denney.ws wrote: