Hello Moon and users! Thanks a lot for sharing.
I propose to abandon the roadmap. The idea of no roadmap in open source projects is from the book "Social Architecture". The author is Pieter Hintjen. As it's written: in 2007 he founded the successful and broad-ranging ZeroMQ community. There is the chapter "How ZeroMQ Lost Its Road Map" in [1]. Furthermore the current item list is more like "wish list". I believe one can rename the wiki-page to "Community wish list" [1] https://hintjens.gitbooks.io/social-architecture/content/chapter5.html Regards, Maksim Belousov From: Jongyoul Lee [mailto:jongy...@gmail.com] Sent: Thursday, March 08, 2018 3:10 AM To: users <users@zeppelin.apache.org> Subject: Re: [DISCUSS] Roadmap 0.9 and future Thanks for the sharing. As an enterprise user, it would help a lot to support Spark more concretely. And containerization became a default option to operate a framework. For provide immediate release of Spark, I suggest separate release cycle between zeppelin-* module and interpreters. This would help contributor see more frequent releases of interpreters they are using. Thanks, JL On Thu, Mar 8, 2018 at 4:08 AM, moon soo Lee <m...@apache.org<mailto:m...@apache.org>> wrote: I think there were couple of discussions about 0.8.0 release in the community. We have umbrella issue that tracks 0.8.0 release [1]. And i see not many blockers are left. Jeff volunteered the release manager for 0.8.0 and I think he can give better idea of 0.8.0 release estimate. Thanks, moon [1] https://issues.apache.org/jira/browse/ZEPPELIN-2385 On Wed, Mar 7, 2018 at 10:49 AM Paul Brenner <pbren...@placeiq.com<mailto:pbren...@placeiq.com>> wrote: [https://share.polymail.io/v2/z/a/NWFhMDMwOTE2OTMz/BRe7pNgtAmQwzQfWdRurKMu_L3CexD7fbzihUQ9bmqEuXHBjiuR2B50qLJ6JeUnOOIHtNGf9IAWHv9b8xv6g5da0m57iEDKSQZ1e3rLZr6y8VKN7JHEMrJDVbOGhHSngSnovAYT-1vZDQE_5ziuawEs6EUk=.png] Thanks for sharing the results of the meeting! Regarding the point "Need more frequent release”, was there any discussion around when 0.8.0 will be officially released? I remember a message a few months ago suggesting that it might be soon. [https://marketing.placeiq.net/images/placeiq.png]<http://www.placeiq.com/> Paul Brenner [https://marketing.placeiq.net/images/twitter1.png]<https://twitter.com/placeiq> [https://marketing.placeiq.net/images/facebook.png]<https://www.facebook.com/PlaceIQ> [https://marketing.placeiq.net/images/linkedin.png]<https://www.linkedin.com/company/placeiq> DATA SCIENTIST (217) 390-3033<tel:(217)%20390-3033> [PlaceIQ:CES 2018]<http://pages.placeiq.com/2017-Integrated-Marketing-Whitepaper_LP_Download.html> On Wed, Mar 07, 2018 at 1:21 PM moon soo Lee <moon soo Lee <mailto:moon+soo+lee+%3cm...@apache.org%3E> > wrote: Hi forks, There were an offline meeting yesterday at PaloAlto with contributors and users. We've shared idea about current state of project and future project roadmap and wishlists (meeting note [1]). It was really inspiring and exciting time. Let me try summarize, move this discussion to online. There were many ideas related to Interpreter. Especially, there were consensus that Spark support is one of biggest strength of Zeppelin and need to make further improvement to keep the strengths. * Spark * Immediate support of new spark release * Ramp up support of current Spark feature (e.g. Display job progress correctly) * Spark streaming support * Handling Livy timeout * Other interpreters * Better Hive support (e.g. configuration) * Latest version PrestoDB support (pass property correctly) * Run interpreter in containerized environment * Let individual user upload custom library from user's machine directly * Interpreter documentation is not detail enough And people in the meeting excited about ConInterpreter ZEPPELIN-3085 [2] in upcoming release, regarding dynamic/inline configuration of interpreter. And there were ideas on other areas, too. like * Separate Admin role and user role * Sidebar with plugin widget * Better integration with emerging framework like Tensorflow/MXNet/Ray * Sharing data * Schedule notebook from external scheduler Regarding scheduling notebook, Luciano shared his project NotebookTools[3] and it made people really excited. Also, there were inspiring discussions about the community/project. Current status and how can we make community/project more healthy. And here's some ideas around the topic * Need more frequent release * More attention to code review to speed up * Publishing roadmap beforehand to help contribution * 'Newbie', 'low hanging fruit' tag helps contribution * Enterprise friendly is another biggest strength of Zeppelin (in addition to Spark support) need to keep improve. I probably missed many idea shared yesterday. Please feel free to add/correct the summary. Hope more people in the mailinglist join and develop the idea together. And I think this discussion can leads community shape 0.9 and future version of Zeppelin, and update and publish future roadmap[4]. Best, moon Special thanks to ZEPL<https://www.zepl.com> for the swag and dinner. [1] https://docs.google.com/document/d/18Wc3pEFx3qm9XoME_V_B9k_LlAd1PLyKQQEveR1on1o/edit?usp=sharing [2] https://issues.apache.org/jira/browse/ZEPPELIN-3085 [3] https://github.com/SparkTC/notebook-exporter/tree/master/notebook-exporter [4] https://cwiki.apache.org/confluence/display/ZEPPELIN/Zeppelin+Roadmap -- 이종열, Jongyoul Lee, 李宗烈 http://madeng.net
