Hey, I was away for a vacation and had some pressure at my daily job before that, now all of that has cleared up. Yaniv and I started integration right before my vacation on the recent re-implementation of the Python SDK and the PySpark SDK. We work on integration in the level of configuration files prepared by the leader and pulled by the executor, these files are used to configure storage, logging and the generation of the Amaterasu runtime. We are really a few steps away from finishing it. I expect that if we can put an effort into this in the upcoming weekend, we can finally close this feature and move on to the next task.
Cheers, Nadav On Tue, 2 Oct 2018 at 10:28, Davor Bonaci <da...@apache.org> wrote: > Any comments? Anyone? > > Option 1: start a vote to retire the podling and move the project into your > own repository. > Option 2: keep things as-is for a few months and re-assess. > > I'd say Option 2 requires a minimum of 3 people explicitly saying that they > want to continue trying and contributing. > > On Sun, Sep 23, 2018 at 8:13 PM Davor Bonaci <da...@apache.org> wrote: > > > Thanks Yaniv for your comments. > > > > - After the release of 0.2.0 the community became very quiet. I think > >> that at this point in the life of the project it is natural, as we > all > >> doing this in our free time and the release was a major effort that > >> all of > >> us (after talking to members in the community) had to compensate for > >> in our > >> day jobs and families. > >> With that said, we shouldn't have gone so quiet. I think we can all > >> agree this is not acceptable for so long (if at all). > >> > > > > Not sure I agree: it is not natural for projects in the Incubator to be > > quiet. It does happen to projects that are getting obsolete/irrelevant, > > often after many years as TLPs. The release usually *increases* activity > > around the project as new users come, ask questions, start contributing, > > etc. > > > > On the other hand, totally fine for people to go quiet. The problem isn't > > around anybody going quiet, but the fact of nobody new arriving. Is there > > any evidence of any usage of the release? Anybody hitting any problem? > Any > > lack of documentation? Any bugfixes? That's the core of the problem. > > > > > >> - It is very critical at this point to grow the community. Going back > >> to > >> my first point, as long as we are such a small community, efforts > like > >> releasing a version will set us back, and the last release is a good > >> example for that danger. > >> > > > > Not sure I agree: releases usually pick up the activity, pick up new > > users, as new features now make the project more attractive. I don't > think > > I've ever seen an argument where "releasing a version sets us back". > > Especially the *first* one. > > > > - Grow the community. BTW I think this is one reason we should > consider > >> staying an Apache project, I think that with the release, we should > >> also > >> shift some focus to growing the community. This is an issue I see > other > >> projects struggling with, this includes TLPs such as Apache Arrow > (in a > >> recent thread on their dev list) and I don't think there is one > answer > >> on > >> how to do it, and I spent some time on other lists to see if they > have > >> solutions. I think we can do many things to fix this, and it's more > of > >> a > >> trial and error process for most projects. Things we can (and should > >> start > >> doing immediately) includes doing more public presentations (and I > >> have to > >> give a shout-out @Nadav Har Tzvi <nadavhart...@gmail.com> that > >> presented > >> in two conferences recently), write blog posts, and we should all > >> invest > >> time in doing so. But one thing we also need to do is actively > looking > >> for > >> more contributors. If anyone here has someone they think is a good > fit, > >> let's try to get them onboard. > >> > > > > Outreach (blogs, talks, etc.) can help, but they help you *scale*. I > think > > the project hasn't demonstrated early user fit -- and trying to scale > > before establishing that often doesn't yield results. For example, if you > > were to throw Amaterasu in front of 1000 people, how many would join the > > community? If only a few, it is probably a bad idea to do it. (I worry it > > is less than a few.) > > > > The problem is likely with the user fit, and can be solved only by user > > development -- most of which often happens before scaling, before > building > > the community, and before joining the Incubator. > > > > It is really, really, really hard to build the community before early > user > > fit. > > > > I think that the next few months are more about staying in the incubating > >> or not, it is do-or-die for Amaterasu. We need to fix the situation so I > >> wouldn't rush in this situation to consider retiring quite yet. > >> > > > > I'm totally fine with leaving things as-is for a few more months. But, I > > don't think it is realistic to expect changes to the degree necessary to > > graduate. As a result, I think you can use *your time* better. > > > > * * * > > > > I'm really sorry to be the messenger of bad news. I don't want to paint > > Amaterasu (or your work) in any bad way. I do want you and the project to > > be most successful as possible. > > > > Please note that I don't gain anything by driving this conversation. In > > fact, I lose a lot. Time is a precious resource of everyone -- and I'd > like > > to make sure that the time *you* have for the project is spent in a way > > that is likely to yield results, not trying to achieve various ASF goals > > that may not achievable. (If this is not obvious now, I trust that over > > time direct feedback and tough conversations will be appreciated, instead > > of letting you waste time for something unrealistic -- which is what most > > people in my shoes would do.) > > >
