70 likes | 186 Views
Workflow Related Issues in MBP Development. Chengzhi Liang ICIS Workshop, Peth March 3, 2010. Key Issues. Search Simple keyword search without domain knowledge required Refined search, eg , ID/accession search Advanced search, eg , mart functions Browse
E N D
Workflow Related Issues in MBP Development Chengzhi Liang ICIS Workshop, Peth March 3, 2010
Key Issues • Search • Simple keyword search without domain knowledge required • Refined search, eg, ID/accession search • Advanced search, eg, mart functions • Browse • All datasets in the database, to give user summary info for each dataset • Integration of software programs for various purpose, eg, • Data input tools, curation tools, QC tools • Germplasm list management • Breeding prediction/simulation tools • Visualization tools, eg, for segregation events, graphical reporting • Configurable • Each program is pluggable, ie, can be replaced by another program of same functionality seamlessly • Well-defined data exchange format, via file or db records • Pipeline set up • Software program repository • Register all programs (allow different versions) • Automatic updateable – to relieve the burden of users
IRIS Web • ‘Ad hoc’ development for rice data • Serve as pioneer work for MBP • Will also be an advertisement site for MBP • Project management • JIRA • All new feature requests will go there • Replace CropWiki for issue tracking • Two projects • IRIS • MBP • Most tasks will be shared
Some Technical-side Issues • Top down approach (vs. Bottom up) • GUI/workflow/pipeline (vs. Domain model) • User demand-driven (vs. Technology-driven) • User expectation (vs. available low-level coding support) • Code refactoring (vs. Coding w/ pre-design) • Ad hoc (data-usage centric) + abstraction (vs. pre-designed objects) • Useful for high level of functionalities • Solid foundation for the latter above; However, • Interface is not a simple assembly of underlying parts • Extra layer of complexity can’t be resolved by reduction • Data storage vs. utilization • Time to use the other direction • Eg, simple search, pipeline (one-click) for dummies • Separate data from interface • User interface is diverse – refactor common ones • Maximizing re-usage of non-interface codes • Reducing time to convert interface from one to another
ICIS Workbook for Data Input & QA • Most breeders use Excel-based input tools • Do we need to keep a simplest Excel-based workbook for data input and QA? • Not possible to store all genomic data on a laptop • Only subset of MBP features available to non-internet users • The workbook will be integrated into the work flow • Used independently – load data local only • Used in pipeline – combine w/ other features in MBP • Issue: • Not all programs will be java-based