Hi all,
here's the result of today's Communities call between Andreas and I. We
split our demos into the following subprojects:
Demo 1: Experiment Design Wizard
============================
Subproject 1. Rarefaction OTUs / Proportion of Community, Suggested
sequencing depth by technology (Andreas point)
Subproject 2. Error calc / QA (Andreas point)
Demo 2: Bioprospecting
===================
Subproject 3. OTU / species abundance and gene family / fxnal category
abundance (Andreas point)
Subproject 4. Link OTUs to species (Dylan point)
Subproject 5. Functional specialization by gene tree (Dylan point)
These will be the projects subject to the design reviews and the focus of
our hack-a-thon. We may be asking for your help putting together the
workflow / dependency diagrams and mock-ups for these.
The plan for the hack-a-thon is to build some inter-core glue and methods
to achieve the goals of the demos. We expect that the CDM and persistent
store revealed during the pre-release meeting on Jan 18-19th will not meet
the needs of the Communities domain, and it will be incumbent upon us to
identify what's missing and needs to be changed. We'll probably use the
hack-a-thon to write prototypes for those methods that are missing (perhaps
all of them!). I want to stress that these will be prototypes, and while
we will do our best to conform to the architecture design and libraries
built by infrastructure, we expect they will ultimately have to
re-implement any methods we build during the hack-a-thon to match their
paradigm. Therefore we agreed we don't want to spend all our time during
the hack-a-thon perfecting them to match infrastructure's and instead
should bang out as much ugly code that works as we can.
One thing that remains undetermined is how much and what Communities data
will live in the CDM, and how much will be accessed dynamically (sliced
into the persistant store / cache) from the cores. We'll want to get that
clarified soon and cemented at the meeting on the 18-19th.
Thanks,
- Dylan
