18:34:58 RRSAgent has joined #exi 18:34:58 logging to http://www.w3.org/2016/12/06-exi-irc 18:35:00 RRSAgent, make logs public 18:35:00 Zakim has joined #exi 18:35:02 Zakim, this will be EXIWG 18:35:02 ok, trackbot 18:35:03 Meeting: Efficient XML Interchange Working Group Teleconference 18:35:03 Date: 06 December 2016 18:43:10 DP: I started with gzip, deflate and brotli. 18:43:28 DP: I am still not sure about brotli number yet. 18:44:15 DP: Pre-compression seem to work well. 18:47:56 DP: I tried several options to increase buffer to make brotli. 18:48:12 DP: That mau have caused to make brotli work better. 18:50:12 https://opensource.googleblog.com/2015/09/introducing-brotli-new-compression.html 18:53:08 DP: Brotli uses dictionary for Web data. That may be helping brotli work better. 18:53:13 compared perf of brotli, zopfli, deflate... http://www.gstatic.com/b/brotlidocs/brotli-2015-09-22.pdf 18:59:15 DP: I would leave it open so that we do not define any compression scheme. 19:02:17 TK: I agree to make it open. There will be an issue of what should be the default. 19:03:00 DP: We can use a header option to identify compression scheme. 19:04:12 DP: I read that brotli is a bit slower than gzip in encoding. 19:07:13 CB: I think we should continue measurement. 19:07:31 CB: Nate may also have some thoughts on this topic. 19:13:51 CB: Instead of letting people decide which one to use, we should provide a guideline. 19:31:39 https://en.wikipedia.org/wiki/Brotli 20:02:04 rrsagent, create minutes 20:02:04 I have made the request to generate http://www.w3.org/2016/12/06-exi-minutes.html taki 20:51:43 Zakim has left #exi