Re: [Treebase-devel] ABI proposal for phyloinformatics

[Westneat, M. <mwe...@fi...>]

Hi all, Picking this up again today, I agree with Bill that feasibility and
scope are key for this grant, so if it is focused on the TreeBASE
re-engineering and first steps of interop to ToLWeb, that seems like a
feasible project.  I didn't get that impression, however, from reading the
proposal draft (seemed like retooling of both projects), so it may need some
honing to make that clear.  So, for our ToLWeb group, in this grant we
should plan on open sourcing, and helping make the tie-ins with TreeBASE,
and perhaps use this exercise as a way to plan for future more complex
modifications to ToLWeb.
Trying to think this through, here are some questions, ideas, devil's
advocate positions, etc.
1. How many phylogenetic trees have been published?  The totality of all
phylogenetic knowledge is probably not very big.  Maybe 10,000 studies
proposing topologies?  For maybe a half million taxa?  Just a guess.  We
should map out the path to the capture of all phylogenetic knowledge, past
and future, and declare how far this grant will get us toward that goal.
2. What should the ultimate set of tools for phylogeny storage, retrieval,
viewing, and use look like?  Are we getting there by re-engineering
TreeBASE?  I think we are, but we better make that really clear in the
proposal that this time, a new TreeBASE is really going to be awesome and
give user joy to the community.  Reviewers will demand this, and I see the
various criticisms of TB2 as a potential risk of having this proposal be too
TreeBASE centric.
3. Will TreeBASE and ToLWeb remain independent projects during and after
this grant?  Or should they be somehow fundamentally joined to produce a
resource for storage, retrieval, and public display of evolutionary trees?
I can see advantages of either path, but we are going to have to make a
decision on this, defend it in the grant, and show reviewers that this is
headed where the community wants it to.
4. The idea of integrating social networking into TreeBASE (or ToLWeb) is
risky- I would advise against it.  There are plenty of ways to connect with
colleagues without trying to duplicate FaceBook tools.  IMO we would be
better off building a way to view and browse the trees.  Or, I like Hilmar's
suggestion of building a platform, flexible enough to incorporate social
tools and viz tools, etc.
5. The pushback on integrating tree viewers in this project is surprising to
me- I recommend against a rework of TreeBASE plumbing and infrastructure
without starting from what the users want, which is to be able to easily see
the trees.  Not sure if you have done any user needs survey or audience
assessment, but the lack of a friendly UI and tree viewer are the two most
common things I hear about TB2.
6. I like the idea of the links between TB and TW.  Thinking user
functionality, what will this provide? While browsing ToLWeb, an icon or
link will automatically take me to a TreeBASE tree that has any of the same
taxa.  Or while searching TreeBASE for primate trees, the ToLWeb site for
primates is offered as an option.  This will require taxonomic resolution of
clade names and tips between the two resources.  Is this name resolution
functionality part of the grant- has it been tried?  Shouldn't be too hard
to see what the matches and misses currently are, but an auto taxon mapping
tool may be more challenging.
Those are my thoughts for now- probably not helpful for actually writing the
grant, I'm afraid, but perhaps useful for deciding what kinds of things to
include.  Cheers, Mark


On Wed, Jun 15, 2011 at 6:25 PM, Hilmar Lapp <hl...@ne...> wrote:

> Great discussion. To throw in my two cents: I think it's important to keep
> this vision of creating a platform in mind. A platform doesn't solve all
> problems and doesn't serve all needs, but it provides unified and easily
> programmable access to useful content (it's the data, stupid!) such that
> people with the skills and drive can innovate. The content being useful in
> this case to me implies that it is much more comprehensive than it is now
> (most published phylogenetic trees are not deposited in TreeBASE, and the
> ToL is far more incomplete than current state of published phylogenetic
> knowledge), and that it is much better annotated and curated than it is now,
> which I find hard to accomplish without a much improved connection between
> ToLWeb and TreeBASE, and without well-thought through approach to overcoming
> the social barriers to user engagement.
>
> Having said that, small proof-of-concept viz (and other) apps can serve
> well to guide and validate the "platform" and API goals. But they shouldn't
> take a scope that would start to distract from the main focus. Perhaps a
> good instrument for supporting these proof-of-concept efforts are
> developer-engagement events (challenges, hack-days, competitions, ?).
>
> -hilmar
>
> Sent with a tap.




-- 
Mark W. Westneat
Curator of Zoology
Robert A. Pritzker Director, Biodiversity Synthesis Center of the
Encyclopedia of Life
Field Museum of Natural History
1400 S Lake Shore Dr.,  Chicago, IL  60605-2496
(312) 665-7734

My website: http://synthesis.eol.org/users/mwestneat<http://biosync.fieldmuseum.org/users/mwestneat>
BioSynC website: http://synthesis.eol.org