[Phylobase-devl] Phyloinformatics Summer of Code 2008

[Peter Cowan]

On Mar 4, 2008, at 11:19 PM, Steve Kembel wrote:
> Peter said...
>> Areas that come to mind for me would be:
>> * Optimizing basic operations, analyses on multitree objects are
>> likely to be memory intensive and slow.
>> * Non-ape or ade4 based plotting.  Perhaps implemented in the grid
>> system.
>> * Transitioning ape and ade4 code to code that operates directly on
>> phylo4 objects
>
> I'd add finishing the implementation of a data standard for phylo4d
> objects - either completing the basics of implementing trees +
> (meta)data, or linking some specific existing metadata standard to
> trees (do we have a phyloXML yet?).

I like this idea, and I've managed to come up with one more.   
Implementing unit testing using RUnit [^1].  This is somewhat similar  
to the test folder that is run with R CMD check but allows, and  
encourages, smaller tests.

> I think Peter's last suggestion would be particularly useful -
> implementation and optimization of some of the basic tree operations
> natively in phylobase. To prune a tree we currently are extracting the
> phylo4 tree back into the ape format and pruning it using the ape
> drop.tip function and then turning it back into phylo4... I haven't
> tried manipulating any huge trees yet but this cannot be very
> efficient. I'd be happy to mentor someone working on this. If this
> alone wasn't sufficient for a 12-week project, perhaps other
> objectives (plotting? multiTrees? phylo4d?) could be added and several
> of us could be the primary mentors?
>
> Anyone else interested in participating? My assumption is that we
> could do some collaborative mentoring via phylobase-devl. Hilmar, when
> would we need a proposal ready, and how much detail would be required?

As mentioned I'm willing to help out with this.  Perhaps Steve and I  
could draft up a proposal and send it out to the list?

[1] <http://cran.cnr.berkeley.edu/web/packages/RUnit/index.html>

Peter