org/meetings/2011MiniDevMeeting/notes.md

+= Short-term list =
+0.2.2.x-alpha --> stable
+
+increase numbers of bridges -> bridges by default -> tor appliances
+
+Review un-reviewed proposals
+
+modular transport (client -> relay) --> ip hiding, entry point
+
+less blockability
+
+bridge distribution
+
+= Long-term list =
+UDP (relay to relay for performance)
+
+better crypo -> ecc? multithreaded?
+
+ipv6
+
+scalability issues -> nodes, users
+
+== Sponsor F ideas ==
+torbutton enhancements, chrome integration, torbutton in mobile environment
+
+clean up after tbb users to zero footprint
+
+support camoflaged transport
+
+bridge user/operator usability
+
+performance improvements
+
+metrics - safely measuring tor
+
+== Hand-wavy ideas ==
+
+ * less centralized {dir auths | exit relays}
+ * consistently faster
+ * scalable: udp
+ * easy bundles auto-updated
+ * (not so tied to firefox)
+ * ui leads users to better anonymity choices
+ * better user feedback
+ * better relay operator feedback
+ * faster dev cycles from thought to code to deployed in binary
+ * lots of bridges (entry points) & bridgedb to handle them
+ * not blocked
+ * full ipv6 support
+ * users are trained to help one another
+ * well-maintained and visible docs
+ * "no docs needed"
+ * help those that need tor but don't use it now.
+ * hidden services that don't suck.
+ * lots of relays, like 100,000.
+ * better understanding of anonymity than today.
+ * better integration with academic researchers.
+ * better communications with and understanding of commercial filters.
+
+== declarative statements ==
+ * research on less centralized directory authorities.
+ * we have less centralized exit relays because running them is less painful.
+ * more vendor supported packages
+ * auto-updated packages with thandy and bittorrent
+ * work on low bandwidth environments.
+ * faster consistenly because of better metrics that feed into congestion control and latency management.
+ * better instrumentation and raw data collection to assist metrics collection and interpretation.
+ * more scalable because microdescriptors work and have datagram transport with userland code stack.
+ * We evangelize to the community to grow more coders and architects with bounties for proposal implementation and sexyness factor.
+ * computer human interface research into tor usability.
\ No newline at end of file