CloudServices/Location/Meetings/2013-08-15

« previous week | index | next week »

Agenda

Stumbler

• cpeterson and dougt working with Legal and PR about privacy notice
• Play Store release is delayed a few weeks, waiting for updates for privacy notice and PR
• Guilherme implementing "battery low" intent handling

Server

• Switched local testing from sqlite to mysql to better match RDS and use some mysql specific functionality
• Reworked sql connection pooling and test db isolation
• Added explicit master/read-replica connections and use read-replica for all read-only functionality
• Bunch of minor website updates, changed map to a heatmap
• Started reading up on celery as a potential async worker / timed job scheduler with configurable backends (probably mysql at first to avoid additional ops-complexity)

Other

• TBD: Hanno travel?
• "Location Estimation Methods for Open, Privacy Preserving Mobile Positioning", Master's thesis from Brendan Johan Lee

Next Actions

• Post stumbler apk builds in a semi-official Dropbox URL and email dev-geolocation when the apk builds are updated.
• Do we need to filter (some) duplicates from stumblers? Stats page says we have ~20 measurements per AP and ~200 measurements per cell tower. Hanno says no because having even 100-1000 measurements for the same wireless network is still useful. The main reason being, that a lot of environmental factors can influence the measurements (cell towers / AP's dynamically adjusting power levels based on devices in range, different data accuracy from device sensors, different levels of interference / blockage, ...)
• Hanno planning to visit MV and SF for the first two weeks of September. \o/