judicious use of custom development in an open source component architecture
TRANSCRIPT
Judicious use of custom development in an open source component architecture
Sky Bristol1
Natalie Latysh1
Dell Long2
Steve Tekell2
Jeff Allen2
U.S. Department of the InteriorU.S. Geological Survey
1. USGS Core Science Analytics, Synthesis and Libraries
2. USGS Fort Collins Science Center
AGU Fall Meeting 2014
Talking Points
• ScienceBase – inspiration and a brief history• Development methodology• Information architecture• Cyberinfrastructure architecture• API first• Lessons learned and motivators for future• Open invitation (and invite us to your party
too)
AGU Fall Meeting 2014
20062007
20092015myUSGS Data
Explorer/Data Uploader
Scientific Data Catalog/Comprehensive Science Catalog
ScienceBase 1.0 and then 2.0
ScienceBase-Constellation
71 versions
Collaborative tools and simple file upload
Metadata Cataloging & Research Item Concept
Digital Repository & Research Item Faceting
Federation of Disparate Data Schemas
“In the research process, we need more than just a big catalog of data. We need all of the other important information connected to our work – published papers, manuscripts, software, and information about people, labs, projects, and others in our field.”
Inspiration and History
AGU Fall Meeting 2014
Manage and Contribute
Register
Find it at the USGS
Find it at the DOIFind it at Data.gov
Find it in a Data.gov community(climate.data.gov)
Find it in other communities(GeothermalData.org)
Open Data and USGIN Methods (“abridged”)
AGU Fall Meeting 2014
Lessons Learned
• We encounter some new type of data just about every week and need to account for a lot more variability.
• It’s pretty expensive to run an institutional repository, but we have it to do.
• Baking security in from the ground up adds quite a few constraints to the ideal architecture.
• We can’t please everyone all the time, but if we keep building great platforms then people come back.
• APIs (and everything else, for that matter) need to be driven by real, specific use cases to be usable.
• R&D, engineering, and operations are all different and need different management methods.
AGU Fall Meeting 2014
Questions
www.sciencebase.gov
[email protected] Latysh ([email protected])
Sky Bristol ([email protected])