[Phylobase-devl] node retrieval / recapitulation

[Thibaut Jombart]

Peter Cowan wrote:
>> - "nodes": equals 'internal nodes' (has my vote) /vs/ all nodes  
>> (tips +
>> internal)
>>     
>
> I'm torn here, I'll be fine with either decision.  I do think that we  
> need not feel bound by the decisions of ape.  One of the motivations  
> for phylobase is that ability to make changes to the ape approach.
>   
I agree, but I since phylobase users are going to be using ape as well, 
I just want to avoid unnecessary confusion.
>> - "NexusToPhylo4" (and related IO functions): there seems to be a
>> consensus for changing the name, and using an argument to specify the
>> output; "read.nexus" would conflict with ape, but this may be solved  
>> by
>> using a namespace (is this overkill?); Hilmar suggested naming the
>> function from the content read (phylogeny / comparative data) rather
>> than from the format (newick, nexus, ...). I think content is more
>> generic, but I find it more visible to have a read.[format].
>>     
>
> I don't think a namespace is overkill, but that said, I haven't  
> wrapped my head around them.  I would prefer names such as  
> read.fileformat(), like read.csv(), unless we could reliably write a  
> function that would detect the format and dispatch the proper method -  
> which is probably more work than it is worth.
>
>   
I don't think so. But both options can be used: have several 
read.[format] methods and a dispatcher. Once the appropriate IO methods 
work correctly, it is actually straightforward to detect the extension 
of the file and dispatch the appropriate method.
>> - naming functions: there seems to be a consensus about the  
>> "camelCase"
>> (i.e. "lowerCamelCase" rather than "UpperCamelCase").
>>     
>
> I would also favor singleCamel, rather than doubleCamelCase.
>
>   
>> Appart from this, one colleague here is presently using/crash-testing
>> phylobase - as an enduser -, and I begin to receive good feedback  
>> (among
>> bunches of "why is this not working?!?"). A good point for us all, I  
>> guess.
>>     
>
> That is both exciting and scary.  Your colleague should feel free to  
> join the mailing list and complain to the lot of us, if s/he is willing.
>
> Peter
I'll ask him. But he really has a great experience in complaining to me 
about things 'not working in R', so I am not sure he's going to change 
his habits ;-)

Best,

Thibaut.

-- 
######################################
Thibaut JOMBART
CNRS UMR 5558 - Laboratoire de Biométrie et Biologie Evolutive
Universite Lyon 1
43 bd du 11 novembre 1918
69622 Villeurbanne Cedex
Tél. : 04.72.43.29.35
Fax : 04.72.43.13.88
jombart at biomserv.univ-lyon1.fr
http://lbbe.univ-lyon1.fr/-Jombart-Thibaut-.html?lang=en
http://adegenet.r-forge.r-project.org/