SC_20100921

Owned by gendlerk
Last updated: Sept 30, 2010Version comment
3 min read

iPG2P Steering Committee
September 21, 2010; 2 to 3pm EDT

Present: Matt Vaughn, Tom Brutnell, Steve Goff, Ruth Grene, Jeff White, Chris Myers, Steve Welch, Eric Lyons, Tina Lee (notes), Claire Stuebe (EMEC)

Notes/Agenda:

  • Overview of API - Matt Vaughn
    • Grad students know about tools before you do
    • Point is to get young developers, postdocs, etc. to use iPlant, develop modular functionality, making it so they don’t have to worry about hosting and distribution, get early adopters on board, so get API and platform built and see that they are supported well
    • Brutnell agrees but thinks top down is important too as awardees are the ones writing grants and talking to NSF
    • Concern that SC doesn’t have a sense of what the developers are doing.
    • Many nice tools are in the works, what iPlant is doing just needs to be communicated to a broader audience
      • Suggestion: Get WGs working again for a start
      • iPlant/G2P needs to communicate directly with community websites, mailing lists, etc to publicize tools, species supported, etc.
      • talk with other PIs doing similar things, Alice Barkin, NGS-Maize GDB, so get more visible by getting button on the browser. Is there an equivalent tool for iPlant?
    • VA prototype needs more airing in Grene's opinion but Vaughn says not yet ready to do that
      • Maybe do an online presentation to put in its best light, e.g., iPlant’s ASPB effort
    • What are some venues to showcase the tools?
      • iPlant will be doing something at PAG
      • Meeting with IBP group
        • propose to integrate people from IRRI, ICRISAT, and CGIAR to work at iPlant to work with iPlant platform.
    • As soon as API stabilizes, we can show toolkit.
      • API and toolkit should be relatively concurrent, show by end of the month.
    • Mid-October for authentication
    • VA prototype and ToL functionality using API being modified
    • Specs are under construction right now, a draft will show everything done to date, by end of this month
      • There is a doc Matt could send, it’s already out of date (Rion and Matt have been traveling), but meeting next week to hammer out the description and find common ground.
    • The iPlant API is intended to be a rational way of dealing with all of the different types of things you can do
      • authentication, submitting jobs, it’s a gateway for accessing services
      • Specs for subroutines, data structures, types of software objects, or specific communication protocols.
    • An API by itself is abstract
    • SDK is an additional set of tools, API, all its documentation, files, headers, libraries, debugging tools, other utilities to help build data structures automatically, sample code, formal documentation, etc
    • iPlant API will support use cases (5 major examples). Doesn’t cover GUI applications however.
    • Languages and Protocols – java, perl, python, json, xml, REST
    • Architecture – picture of what infrastructure looks like.
    • Roadmap – schedule. SDK lags by some time after API.
    • Event: better description once Rion and Matt discuss but essentially it’s event notification. Like writing code and being notified when a mouseover event occurred, similar to call-backs.
    • Mashup is first prototype for beta testing
      • January mashup, that’s loose, not even defined when it’s going to be yet
      • Testers will be DNALC, they are mashup pros
      • Rion and Matt are script pros
      • Matt will have doc describing draft of API and will circulate to the SC, he won’t not vouch for its clarity to the lay person, and it’ll change.
    • On I/O point, will people be able to put data thru pipeline?
      • Still must deal with large file upload, esp. for desktop data, but yes, should be the entry point
      • Could do additional genome support in quick and dirty way but not sustainable method, so Doreen wants a more sustainable strategy for supporting additional plant genomes
        • Slots 1 and 2 are Arabidopsis and maize, others would follow shortly.
    • Awardees weren’t feeling iPlant was for them because they work in other species like tomato.

Action Items:

  • Matt to send Jeff White and Steve Welch a draft/prototype of the API prior to their attendance at the iPlant-IBP meeting next week (Sept. 29). Others can wait until next week to see the draft of the current API.
  • Matt to give same presentation on API to the NSG WG.

Next meeting time: October 5, 2010.

  • Next meeting Matt will bring in Chunlao Tang, who's taking over day-to-day oversight for NGS.

Adjournment: 11:55 AM MST