Breakout 1.2
-----------------
What is your vision for cyberinfrastructure for hardware, software, standards, and data for biodiversity sciences?
General orientation discussion
------------------------------------------
Kelling: produce list of products to be available in 10 years for this community
Is ontology infrastructure? Consensus is yes.
1) Federation of data
Just as important: repositories for the large data sets (e.g., GBIF, NASA AMES, ORNL, GAP products)
-- those repositories need more than just data availability -- need tools to allow researchers to interact
2) Hardware to work on the data
3) Tools/software that can work on federated data for e.g., modeling
Use case: Bird distribution predictions at 3km scale for all of US for daily time scales
-- run into all 3 of those problems
-- ontological -- too much time to rectify data
-- hardware -- needed much more hardware to run the simulations (got Teragrid access, via a chaperone to expedite access)
Guala: can create ontologies, but do they really end up applicable to the science questions being asked?
Kennedy: wants a system that can go to look for data (repository), then ask what tools you want to run on them (by non-scientist?)
-- or alternately find a tool (e.g., visualization tools), and then find the data that works with it
-- discovery tools (but really visualization) of data products that come from analysis/models and used for exploration
-- a lot of what we're doing gets lost
Guala: some activity is tedious (e.g., NBII metadata clearinghouse) and therefore isn't done regularly, but then the data are not available when needed afterwards
Guala: also concerned with being able to support anything that might come out of DoB -- so what is the essence of DoB in terms of data
-- only direction of DoB is that we're 'combining' very heterogeneous data in unpredictable ways
Brainstorming activity
---------------------------
-- what is your vision for CI for 10 years?
1. Develop quality standards for data
2. Provide a web-based platform and tools for people to manage data as they collect it (not after the fact) with a side effect that it becomes part of a federated repository system
3. Suite of tools that scientists can use to acquire, preserve, discover, analyze, and visualize data and that they be easy to use
-- tools that semi-automatically ease the process of metadata generation
-- taking the scientist's perspective -- tools aren't user-friendly enough
-- e.g., easily annotate and validate tabular columns as e.g., lat/lon, units, etc, ability to add semantic information to data
-- capture metadata/annotations from tools at data acquisition time rather than after-the fact by understanding semantics implicit in tools
4. Have an agreed upon (potentially virtual) "place" that is actually used by a broad research and user community as a data repository and users can trace that data back to its source
-- agrees that (3) is a good approach to get people to use it
5. Training for a users within and outside of US in software use for biodiversity research
6. Need to be able to integrate from federated repositories
-- includes either standard data models or data exchange standards, and tools that can operate on these
-- only so much standardization you can do with CI -- divergence of scientific methodology is also an issue
-- data discovery tools -- how do we find the data that you want (search is far too imprecise now)
7. Good CI will promote a world at large that will understand what biodiversity is and how it is important to their lives
8. Lower the bar for users in terms of using all these tools we've been talking about
-- Address the usability issues
9. Reproducible chain from original data collection through analysis and modeling through end-data products for any published biodiversity paper
-- also be able to repurpose components
10. An ontology of everything: assertions about the relationships between pieces of data
-- need to be able to find the data -- need to decide on a global identifier approach
-- data that is "pre-qualified" for fitness of use (as opposed to raw measurements) with metadata that describes those potential years
11. Focus on making the data useful that users are already familiar with (Excel, R); repositories of middleware (e.g., R packages)
12. Single virtual web portal/interoperable set of repositories (like DataONE) that can serve as one-stop-shopping for data, tools, and best-practices [similar to (4)]
13. Coalescing a community of practice (i.e., AGU-equivalent) for informatics earth science discipline (e.g., ESIP)
14. Everything that has to do with biodiversity research (software, data management, analytical tools, GIS, etc.) will be open source
15. Incentive for providing metadata
-- could be some benefits here from (3)
-- review and curation process needs clear incentives
16. Cloud-based, federated system for storing and accessing people's data
17. A common semantic model of scientific observations that supports semi-automated, ad-hoc data integration that spans the range of biodiversity data, plus the tools for those integration services
18. Cyberinfrastructure should allow for feedback/annotation/reviews/ratings of data/models/etc.
19. Linked-data web based on IDs
Synthesized topics
-------------------------
A. Federated, interoperable data system for global access with Web-2.0 UI that will promote understaning of what biodiversity is and how it is important to their lives {7, 4, 12, 18}
B. Training and capacity building globally {5, 8} and improved usability of data lifecycle tools
C. Suite of tools for the whole data lifecycle: acquire, preserve, discover, analyze, and visualize data; ease of use; cross-domain tools {3, 11, 16}
D. General Ontology
-- A common semantic model of scientific observations that supports semi-automated, ad-hoc data integration that spans the range of biodiversity data, plus the tools for those
E. Ability to replicate analysis from source data fully through to publication {9}