2011 working group Mo technical challenges¶

• finalize schema details
• choose database/web framework
• clarify services and interfaces
  • services: how researchers and automated systems interact w/ database
  • optimize interfaces for visual presentation and performance
• workflow
  • heterogenous raw sources
  • some go straight into confed database
  • data supplemented w/ lat/long in db
  • nomenclature, taxonomy issues
  • expose confed resource so it can talk to other frameworks
• finalizing schema details
  • independent researchers should also be able to contribute
  • how lossy can it be? what is minimal info that data coming into framework must have?
    • e.g. Latin binomial, spatial reference
  • VegBank, CTFS, SALVIAS
  • VegX: is semantic scope what we want?
    • needs revisions or good enough as is?
    • how plays together with DwC?
  • specimens and occurrences, TCS, DwC, TraitNet (plant traits)
    • semantics approach to storing trait information
  • how to reference traits stored in trait db
    • omics community wants to reference traits
• VegBank
  • fully-fledged ER model w/ full impl for storing plots info
  • recent hardware refresh
  • implications of deciding we like VegBank
• choosing a framework
  • VegBank (Java/Postgres), BIEN 2 (PHP/MySQL), VegX (XMLS), Python-Django (PostGIS)
  • locational info about jellyfish: Django on Postgres
  • Python gaining traction in geospatial community: GeoDjango
• link to
  • taxonomic info: TCS, TNRS
  • geospatial GNRS, MoL, SOS
  • traits: TraitNet, ontologies (earth science, biology, phenotypes, life sciences)
  • RDF/OWL, triplestores
  • LAMP stacks
  • take less conventional web app approach?
  • how web services implemented in framework
• services and interfaces
  • VegBank home page (Bob designed): simple/advanced plot search
  • well-documented db: tutorial, instructional info
  • is BIEN 3 intended to be highly accessible framework?
  • simple plot search interface streamlined
  • discuss how interfaces should look
• data acquisition
  • TEAM: alliance of institutions; vegetation protocol
  • MBG: how to integrate taxonomic side of things
  • how to get data into system
  • more sophisticated data acquisition tool
  • Eric, Sandy Andelman?
  • data provider for repos
  • VegBranch, TurboVeg load data directly into db
  • now in scope to make progress on data acq
• objectives
  • decide confed schema
  • decide tech framework
  • brainstorm linkages to other services: GBIF exposes resources through services
  • Brad, Jim Regetz, Aaron involved with technical side
  • borrow from ALA?
  • iPlant/NCEAS: dedicated development workforce
• more issues
  • access control granularity
  • differentiate between atomic occurrence and aggregated view of data
    • % coverage vs determining % form indiv occurrences w/in plot
  • inadequate metadata
  • allow people to choose from methodologies for aggregating data
  • handle specimen and occurrence data better
  • time series/repeated measurements need special flagging and treatment
    • VegBank: observations nested in a plot, multiple data observations nested in the first one
    • potential responses of species to climate change: need old observations
    • CTFS data is occurrence data
  • as develop services (TNRS), something wrong w/ data -> provide feedback to data provider
  • when John using BIEN 2 framework and experiences problem, capture issues
  • iterative next gen of BIEN
    • get pieces accessible to operate on
    • research perspective
• technical challenges
  • see RedMine wiki: changes on day-to-day basis
  • reporting/tracking what's going on
• time dimension is important
• how users will add data
• earlier iterations of BIEN: wanted to simplify schema so eliminated fields
  • now, throw nothing away because some fields critical metadata (DwC elements)
  • specimens: every element of DwC that pertains to plants should be in there
• SALVIAS static, users now using BIEN
• CTFS -> BIEN 3
• what would benefit from becoming part of BIEN
• need samples of data/schema so we can see types of attribute info being collected
  • can VegBank capture everything? VegX?
• each subgroup: responsibility includes writing use case (due We afternoon 1pm)
  • BIEN 3 group will use them to define BIEN 3
  • example use cases