[dpdk-dev] DPDK Community Call - 16.04 Retrospective - Wednesday May 11th

[Glynn, Michael J]

Meeting Minutes for Community Call - May 11th 2014

Topic: DPDK 16.04 Retrospective 
Facilitators: Mike Glynn (Intel), John McNamara (Intel)

Attendees: Christian Ehrhardt, Hemant Agrawal, Jan Viktorin, Mauricio Vasquez, Mike Holmes, Stephen Hemminger, Thomas Monjalon, Naoyuki Mori, Tom Gall, Konstantin Ananyev, Mike Glynn, John McNamara, Mohammad Abdul Awal, Bruce Richardson, Declan Doherty, Roy Fan Zhan, Ferruh Yigit, Bernard Iremonger, Reshma Pattan, Remy Horton, Vadim Sukhomlinov, Shanmukha Sreedhar Theerthala

Minutes:
*	IRC chat channel was opened up the discussion (see notes captured below). IRC Channel: http://webchat.freenode.net/?channels=%23dpdk
*	John McNamara reviewed some statistics from 16.04, which included;
o	Number of commits per release - averaging ~700
o	Number of contributors per release - ~100 for the past two releases
o	Unique "Reviewed-by" contributors is very low (7 in least release) 
o	Number of "fix" patches - 301 
o	Number of patches with a "Fixes:" line - has grown from 45 in DPDK2.0 to 261 in 16.04 (attributed to greater compliance to using the "fixes" line)
o	Number of patchset revisions in v16.04 - vast majority at v1 (199) 
o	Days between Author Date and Commit Date - 50% of final patches are merged within 7 days, 75% of final patches are merged within 14 days, "Final" patch means the last version, when all v1..vN comments have been addressed

*	Discussion on Areas for Improvement:
o	John Mc: 'Reviewed by' numbers are low. Thomas: this is likely because people don't distinguish between 'reviewed by' and 'acked'. Agreed that there is still value in retaining the reviewed-by tag.
o	Stephen H: Need more active reviews in the community. Suggested that the Maintainers could delegate to others on the mailing list. Maintainers - please take note.
o	Thomas M: Having a large number of v1's being applied may not be a good thing...would prefer to see higher number of patch revisions since it shows community review process is working better. Reviews are the area which require the most improvement. 
o	Christian E: Packaging of DPDK - any community pushback to package something working almost everywhere but optimized where supported. No objections - Christian will post to the mailing list next week. http://udrepper.livejournal.com/20948.html 
o	Documentation gaps in (1) RTE table library, and (2) ACL need to be addressed. Please discussion on the mailing list, submit patches, or ping John McNamara (documentation maintainer)
o	Mike G: Is the RFC process working? There is usually very little feedback on RFC's. Thomas commented that RFC's are still useful, suggested using header file (with Doxygen comments) as an RFC format.

IRC Chat Notes:
[16:09] == reshmapa [c0c6972b at gateway/web/freenode/ip.192.198.151.43] has joined #DPDK
[16:16] <tmonjalo> Not enough public reviews
[16:17] <tmonjalo> Fixes: lines are increasing (good for maintenance)
[16:20] <tmonjalo> A lot of patches are committed late in the cycle
[16:22] <cpaelzer> just for the minutes - since the call mentioned some missing commit stats - If I didn't mistype that should be 2.2 -> 16.04 - Authors: http://paste.ubuntu.com/16363502/   Domains: http://paste.ubuntu.com/16363543/ 
[16:29] == nijopa [~nijopa at 72.246.0.14] has quit [Quit: Leaving.]
[16:33] == nijopa [~nijopa at 72.246.0.14] has joined #DPDK
[16:35] <tmonjalo> not for minutes: audio quality is bad from some speakers - really hard to understand
[16:36] == yliu [yliu131 at nat/intel/x-phwergujuqevebif] has joined #DPDK
[16:39] <cpaelzer> I wanted at least see how the idea at all reflects with the community - http://udrepper.livejournal.com/20948.html 
[16:39] <cpaelzer> that would allow Distributions to package something working almost everywhere but optimized where supported
[16:39] <cpaelzer> there were no hard opinions on it yet, I'll bring something to the mailing list (prob. next week)

Please feel free to reply if I missed something/captured incorrectly