[Phylobase-devl] labels() and other questions

Brian O'Meara bcomeara at nescent.org
Fri Feb 29 21:12:17 CET 2008


On Feb 29, 2008, at 2:55 PM, Ben Bolker wrote:

> Brian O'Meara wrote:
>
>> I'm more than happy to change (or have changed) the IO function  
>> names  to have them consistent with general R/phylobase style. Ape  
>> conflicts  may be a problem, there, too (such as read.nexus).  
>> There's also the  issue of having one input format (nexus) and  
>> multiple output  structures (tree, data, or tree+data object).  
>> Currently, this is  handled by NexusToPhylo4, NexusToDataFrame,  
>> and NexusToPhylo4D. I  don't know the best style: does one add a  
>> suffix to indicate the  target ("read.nexus.phylo4d"), add the  
>> target type as an argument to  a single read.nexus function, or  
>> some other approach?
>
>   I would vote for a single argument; that way if the function
> wanted to do some guessing (on the basis of what was in the Nexus
> file) it could ...
>   Don't know what to do about the name conflict, though.
>
>    Ben
>
>

> I would add the target as an argument. Having a combinatorial  
> number of functions usually sucks to maintain after a while (and is  
> harder to parameterize from a calling function).
>
> 	-hilmar
>


Okay. I can make the change to use one function with an argument once  
we figure out the new name (read.nexus seems logical) and, if it  
exists in ape or elsewhere, how to deal with the conflict. We could  
use "read.nxs" or "read.nex" which don't conflict with ape, but both  
abbreviations are used frequently as nexus extensions, so some users  
might not find one or the other. "phylobase.read.nexus"?

Brian


More information about the Phylobase-devl mailing list