On 18/08/18 10:24, Ian Luo wrote: > Mark, > > Here's one relevant topic I would like to discuss with you. I understand > the Apache way encourages *open* discussion. In my opinion, the interaction > on GitHub issue is one kind of the open discussion, and many modern open > source projects leverages this as the major channel. What's your opinion on > this?
>From an ASF perspective there are no concerns if that is how a project prefers to communicate (primarily because all the discussion is echoed back to an ASF mailing list). Personally, it isn't my favourite but I think that is more to do with how it is integrated into ASF mailing lists than anything else (ASF mailing lists are the primary way I follow what is going on across multiple areas of the ASF). I don't like the way questions, bugs and PRs all end up in the same place and the lack of threading makes it hard to follow - especially with a high activity project like Dubbo. One of the things on my TODO list for ApacheCon NA is to sit down with the infra folks and figure out if we can improve the integration - particularly the threading. More generally, the criteria that the ASF looks for in communication channels used by projects are (in no particular order): - open to all - asynchronous - available off-line - full history - searchable - archived on ASF controlled systems - low bandwidth / minimal system requirements E-mail may seem a little 'old school' at times but it is one of the few technologies that meets all of the above. Which is why most of our systems are configured to echo stuff back to the relevant mailing list. These days I'm used to an always on internet connections with speeds in the 10s of megabits where I don't need to worry about the cost (even when I am out and about) but it is worth remembering that not everyone is in that position. It wasn't really that long ago that I could sometimes be found working on Apache projects via a 9600 bits per second dial-up connection that I paid for by the second. It was perfectly possible for me to follow what was going on by connecting for a few minutes every couple of hours, syncing my email, making a few commits if I had anything to commit and then disconnecting and continuing to work off-line. If all communication had happened via a GitHub like interface there is no way I would have been able to follow the project. Cheers, Mark PS I see there is a Dubbo session at ApacheCon NA. I'd be more than happy to sit down with anyone that is interested and chat about any questions, concerns, etc. they may have about how the ASF works, why things are the way the are and any other ASF related questions. > > Thanks, > -Ian. > > On Sat, Aug 18, 2018 at 10:31 AM jun liu <[email protected]> wrote: > >>> >>> To provide some examples: >>> >>> I am employed by Pivotal and Pivotal employs the committers on the >>> Spring Boot project which embeds Apache Tomcat. From time to time I >>> receive a work e-mail, slack message or similar along the lines of "We >>> think we have found a bug in Tomcat. Can you look at it?". My response >>> is invariably "Sure. Please create an issue in the Tomcat issue tracker >>> and I'll take a look." >>> >>> I also receive direct email from Tomcat users asking for help with an >>> issue they are having. This happens often enough that I have a e-mail >>> template for the reply that directs them to ask their question on the >>> users@ mailing list. >> >> These examples impressed me a lot, actually, I am experiencing this now. >> People from work sometimes report issues about Dubbo through internal >> communication channels (IM or work email). And I have to admit that in some >> cases, I have chosen to directly discuss problems with them but forgot to >> bring them to the community, which means I misused the two roles of work >> and open source. >> >> Most of the times, language becomes the excuse for making the wrong >> decisions, because some colleagues and users of Dubbo in china are not good >> enough in English. For those users, maybe we can encourage them to provide >> both the Chinese and English descriptions when reporting issues, for >> example, write the Chinese version first and then directly translate to >> English using Google. >> >> I think that the way Mark's been doing is right for running an open source >> project, I will also try to “Redirect everyone reporting issues to the >> mailing list or Github issue tracker". >> >> Best regards, >> Jun >> >>> On 16 Aug 2018, at 19:17, Mark Thomas <[email protected]> wrote: >>> >>> On 15/08/18 14:09, Jerrick Zhu wrote: >>>> Hi, mark >>>> >>>> Sorry for disturbing all of you. >>> >>> No need to apologise. The additional traffic wasn't, and isn't, a >> concern. >>> >>> My concern was that it appeared that there was some sort of organisation >>> going on that the project wasn't aware of. >>> >>> A secondary concern was that multiple teams seemed to be writing PRs for >>> the same issue. >>> >>>> This is an activity for students to participate in open source project, >>>> it's held by department named BaiJi, Alibaba. They came us and asked us >>> >>> If by us, you mean "the Alibaba employees who work on Dubbo" then that >>> request should have been redirected to the Dubbo community - which means >>> the dev@ mailing list. >>> >>> If by us, you mean "the Dubbo community" then I don't recall seeing that >>> request on this list. >>> >>> My concern here is that folks appear to have mixed up their "employee" >>> hat and their "Apache committer" hat. It is easy to do and so is >>> something to keep in mind. A good general rule is that whenever you find >>> yourself discussing anything related to the project at work (or anywhere >>> that isn't the project mailing lists), ask yourself "Why isn't this on >>> the dev@ list?". In my experience it is nearly always the case that the >>> conversation needs to move to the mailing list. >>> >>> To provide some examples: >>> >>> I am employed by Pivotal and Pivotal employs the committers on the >>> Spring Boot project which embeds Apache Tomcat. From time to time I >>> receive a work e-mail, slack message or similar along the lines of "We >>> think we have found a bug in Tomcat. Can you look at it?". My response >>> is invariably "Sure. Please create an issue in the Tomcat issue tracker >>> and I'll take a look." >>> >>> I also receive direct email from Tomcat users asking for help with an >>> issue they are having. This happens often enough that I have a e-mail >>> template for the reply that directs them to ask their question on the >>> users@ mailing list. >>> >>>> to >>>> provide some simple issues that students can fully engage OS project, >> and >>>> we agreed. We also wants more guys to join Dubbo, to contribute. >>> >>> Please be aware that some people read "guys" as referring exclusively to >>> men. I recommend that you try to use a more inclusive term. I tend to >>> use "folks". "people" usually works as an alternative as well. >>> >>> I do think this is an excellent way to increase interest in Dubbo and >>> expand the community. Please don't take anything I am saying as >>> discouragement of this effort. I am fully supportive of it. >>> >>>> Now we have noticed that the PRs came together and generate so many >> emails, >>>> which had disturbed you. We will consider other more effective ways, >> such >>>> as one team fix issues separated from each other. >>> >>> It bears repeating. The volume of email was not a concern. It was the >>> appearance of some sort of organisation of project effort going on that >>> the project community was not aware of. That rings alarms bells for me >>> in my role as a mentor. >>> >>> Regarding separating issues between teams, there are pros and cons of >>> multiple teams trying to fix the same issue. The work might be >>> duplicated but, equally, they might learn from the different approaches >>> that the other teams took. I don't have a view one way or the other. All >>> I suggest (and this is more for the people managing the students) is >>> that the issue is thought about to ensure that the students get the best >>> possible experience. >>> >>>> Do you guys have any other suggestions? >>> >>> More of a comment than a suggestion. Given the volume of activity on the >>> notifications@ list, the activity on dev@ seems rather low. I'd expect >>> to see more discussion, more commentary, more planning given the >>> activity levels. It is possible that this discussion, commentary and >>> planning just isn't happening but I do find myself wondering if it is >>> happening off-list. If that is the case, it *really* needs to start >>> moving on to dev@ list. >>> >>> Cheers, >>> >>> Mark >> >> >
