I'm just trying to get to the heart of the issue here Max. I don't necessarily disagree as stated earlier. Once we have a list of concerns and things that prevent folks from joining and/or posting on dev we can do something about it. As was mentioned earlier this is not something singularly affecting Hop so I'm sure we can learn from other projects as well.
Before we go there, let's pile on a bit more since it seems to me that this is treated as a major concern and I think the Hop devs see it as such. My impression is that the ASF, and as a consequence dev mailing lists as well, are seen as a Very Big Thing and that folks are in general quite reluctant, even afraid, to post something. I think this especially is painful and diametrically opposed to building great software where insights come around in an iterative fashion. So again, I'd be very interested in hearing from our mentors how we can alleviate this so that we can have more lively discussions on dev. I'm thinking along the lines of rewarding participation with swag and so on. We're not Google but we're not completely poor sods either are we :-) Thanks, Matt On Mon, Jan 11, 2021 at 12:36 PM Maximilian Michels <[email protected]> wrote: > The core problem is fragmentation. I would follow Hop development more > closely if I wouldn't have to monitor yet another chat client. Having > one source of truth, is what we should strive for. Again, this doesn't > mean getting rid of the chat but all design discussions should live on > the dev mailing list. Alternatively, a summary could be posted. > > Mailing lists and chats are two mediums. Generally, I find that the > mailing lists are better suited for archival because more thinking goes > into the writing. > > > On the other hand I found it surprisingly hard to find anything in the > Apache mailing list archives. > > @Matt I'm not sure which service you used to search the mailing list but > from my experience, > https://lists.apache.org/[email protected] works quite well. > > -Max > > On 11.01.21 11:08, Matt Casters wrote: > > Again, nobody is disagreeing in principle. However, I personally don't > > think this is as black and white is being stated by the mentors. > > > > First off, Julian made it seem like we're using some sort of locked down > > medium for the chat. Let me state here on the record that the Mattermost > > server is open source and that we don't have any limits on the discussion > > archive. It's as such quite different from free Slack functionality. > > For all practical intents and purposes we are actually archiving > > everything. It's easy to get on the system without limitations. It's > also > > very easy to find something by using the search functions. > > > > On the other hand I found it surprisingly hard to find anything in the > > Apache mailing list archives. Perhaps these need a software update as > well? > > I just think that the gap between both is too wide. Imagine having a > > "discussion" on look & feel, images and so on on the dev mailing list? I > > can tell you right away it's just not practical no matter who wants to > > decide what on the topic. > > > > Furthermore, now that you have me going on this tangent... the concerns > > that are being raised by ASF surely cover major decisions and not > > day-to-day small trivia? Who on Earth is going to argue about minor > details > > of any project? We trust people to commit to the source code but we > > somehow need to somehow see everything that was said and only via a > mailing > > list? > > > > Anyway, we'll do our best. I've reminded everyone to subscribe to this > > mailing list. > > The name of the chat server was changed to Apache Hop. > > We're looking into creating an archive for all chat channels. > > > > Matt > > > > > > > > > > On Sun, Jan 10, 2021 at 10:12 PM Julian Hyde <[email protected]> > wrote: > > > >> Yes, chat is excellent for discussions. That’s actually the problem. > >> Discussions (and therefore decisions) will naturally happen on chat. > It’s > >> great for people who happen to be on chat, but people who are not on > chat > >> will be excluded. > >> > >> So, we need to make a conscious effort to move discussions off of chat > and > >> onto dev@ if they look likely to result in a decision. > >> > >> I’m not saying we should ban chat. Just be careful how chat is used. > >> > >> By the way, a lot of Apache communities face this or similar problems. > For > >> example, I know of one or two projects that started in China and have a > >> strong temptation to use Chinese, which is efficient for them but makes > the > >> rest of us feel excluded. It’s difficult for those projects to grow > beyond > >> their original country. > >> > >> The other suggestions in my email were to create a public, read-only > >> archive of chat so that people can see what has been happening, and to > more > >> strongly encourage people to join the dev list, not just chat. I think > it > >> is important that we do those. > >> > >> Julian > >> > >> > >>> On Jan 6, 2021, at 1:54 PM, Matt Casters <[email protected] > .INVALID> > >> wrote: > >>> > >>> I don't disagree with Julian but for that specific example I indeed > >>> mentioned this first on dev when I stated how I felt about how > important > >>> these integration tests are ... to me. > >>> I'm not sure it warrants a specific discussion since the devs seem to > be > >> on > >>> the same wavelength on the subject. > >>> > >>> Where I do disagree is that chat is excellent for discussions and > >> throwing > >>> ideas against the wall to see if they stick. > >>> The way we typically seem to do it is to just ping an idea back and > forth > >>> and throw it in JIRA in some form. > >>> These cases indeed are always visible and remembered more easily than > on > >>> chats or mailing lists. > >>> > >>> On Wed, Jan 6, 2021 at 10:24 PM Julian Hyde <[email protected]> wrote: > >>> > >>>> I am a bit concerned about Hop's use of Chat. To be clear, Apache > >>>> projects use chat rooms (e.g. Slack, IRC) and they are a good way to > >>>> get questions answered quickly and to build a sense of community. > >>>> > >>>> First, I am concerned about the lack of a public archive. People who > >>>> want to read the chat have to first sign up. (Hopefully, I am > >>>> mistaken. If so, please post a link to the archive on the site.) > >>>> > >>>> Second, let's fix the branding on Chat. Currently it is under the > >>>> project-hop.org domain, and the project is called 'hop'. No 'apache' > >>>> in sight. > >>>> > >>>> Third, at Apache we have a rule 'if it doesn't happen on dev it didn't > >>>> happen', i.e. don't make decisions off the mailing list. The > >>>> conversation on chat is generally pretty benign, but I saw one > >>>> exception: this one from Matt Casters: > >>>> > >>>>> I would like to set a goal of having a substantial set of integration > >>>> tests > >>>>> for 1.0. The bare minimum seems to be all the popular most often used > >>>>> transforms and actions. I know that this far exceeds what P5o and > >>>>> Kettle had but stability is really important. > >>>> > >>>> That discussion should have been on the dev list. > >>>> > >>>> Last, according to the latest incubator report [1], a lot more people > >>>> are signed up for chat than for the dev list (122 vs 22). I am > >>>> concerned that, with such a disparity in membership, Chat will become > >>>> the de facto place that people discuss important matters. I think the > >>>> solution is to increase the number of people on the dev list, and to > >>>> continue to drive significant discussions onto off of Chat and onto > >>>> dev@. > >>>> > >>>> Julian > >>>> > >>>> [1] https://cwiki.apache.org/confluence/display/INCUBATOR/January2021 > >>>> >
