Exhibit 3 F2F – A mid-course review of development and project status

Mountain View, 7/12/2011 & 7/13/2011

Attending in person:

Google (hosts): David Huynh, Stefano Mazzocchi

Library of Congress:  Chris Thatcher

MIT Libraries: Richard Rodgers, MacKenzie Smith

Zepheira: Eric Miller, Ryan Lee, Mark Baker, Uche Ogbuji

Attending via teleconference:

MIT CSAIL: David Karger

*Meeting Objectives*

   - Increase transparency among the PI and project team with respect to
   focus and direction
   - Evaluate progress of design and coding
   - Accelerate the refactoring of the code, and where best to put in place
   abstraction points based on lessons learned to-date
   - Clarify collectively the project metrics of success
   - Develop a road-map for creating a sustainable community
   - Continue to explore models on the most effective means for the core
   teams to interact

*Meeting Summary*

The meeting validated crucial project assumptions and produced a narrower,
clearer path of development for the remainder of the project. It also
identified several valuable reference datasets, together with related
performance metrics, and outlined more concrete plans for engaging the
community as progressively functional releases become available. Major
findings and accomplishments:

1.     The SIMILE Backstage code was established as a viable basis for the
server end of the staged configuration in terms of scale and performance,
based on testing with datasets whose characteristics matched deliverable
requirements. Backstage is an Exhibit data server POC, built on technology
stack similar to Longwell (Sesame triple-store, etc). The server API,
however, should not necessitate any specific stack, just a set of
scale-related performance characteristics.

2.     Therefore, a number of specific modifications were recommended to the
Backstage codebase, including newer Sesame version, refactor for
triple-store aggregation functions, etc.

3.     Although laden with interesting possibilities, an end-to-end staged
javascript Exhibit3 design using Node.js was deemed too ambitious given
current project resources at this time.

4.     We affirmed the definition of simple Exhibit language in HTML (as
opposed to javascript) to preserve ease of use/existing Exhibit model.

5.     Several reference/sample datasets were created and staged on the
order of 100k records, from LC and Zepheira sources.

6.     We determined that simple ‘remoting’ of database.js methods (i.e. to
http requests) would not immediately be feasible - refactoring would be
required to alter assumptions of locality.

7.     We resolved that Exhibit3 deployments would manage their own external
service dependencies (e.g. Babel), so that MIT & simile-widgets.org would
wind down operation as public services.

8.     Therefore, a concrete plan and schedule was devised for
decommissioning the public infrastructure (Babel, Painter, etc).  Briefly,
MIT SIMILE services black out in Sept 2011, and  simile-widgets in January
2012.

9.     Documentation requirements were extensively clarified and a roadmap
was established for building them out. The extensive internal code-level
documentation will form the basis of end-user documentation.

10.  We resolved that initial focus of Exhibit3 would concentrate on a very
small number of views: possibly only one beyond a list view – viz Timeline.
 The view API will include a ‘capacity’ declaration to allow views to limit
visual density, and a pagination function.

11.  We outlined a community engagement strategy, with specific provisions
for public demos using appropriate datasets, notifications of packaged
releases, GitHub monitoring, etc

-- 
You received this message because you are subscribed to the Google Groups 
"SIMILE Widgets" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/simile-widgets?hl=en.

Reply via email to