Day 2 Access working group

Search R & D
-----------
SE Framework optimized for collections up to 100 million URLs - components document processor, query processor, search and matching function, ranking capability

wrappers for at least 6 engines - hyperestraier, javalucene, clucene, kinosearch, xapian, nutch

amazon cloud provides common storage for multiple machine indexing/search - hanzo has built an ingest pipeline to cluster indexing process across multiple machines

nect gen nutchwax
----------------
- integrate nutchwax into nutch as a plug-in
- improve and simplfy installation, indexing and service deployment - esp for 500 mill plus
- provide formal documentation including exhaustive doco of nutch distributed search server

weakness - nutch doesn't have formal documentation

smart playback
-------------

3 phase

- collect and prioritise reqs for revised arch (APIs and reference implementation)
- repackage simplify and document OS wayback
- extend core APIs and ref implementation


top requests:
- w3c accessibility
- simplified video replay mechanisms
- snapshot/thumbnail generator
- JSON interface for accessing file/collection metadata
- install config via scripted workflows
- formal developer and operator documentation