On Wed, 2021-05-19 at 18:22 +0200, Artem Goncharov wrote:
Yes, pool would be great.
Please do not take this offensive, but just stating IRC survived till now and thus we should keep it is not really productive from my pov.
Why is everything what OpenStack doing/using is so complex? (Please do not comment on the items below, I’m not really interested in any answers/explanations. This is a rhetorical question) - gerrit. Yes it is great, yes it is fulfilling our needs. But how much we would lower the entry barrier for the contributions not using such complex setup that we have. well its significantly better then using a pull request model that is used in github or email list when it comes to code review tools or lack of them in the email case. so if gerrit is your low barriier then you have set the mark pretty high. there are few tools that actully do this as well as gerrit does.
- irc. Yes it survived till now. Yes it does simple things the best way. When I am online - everything is perfect (except of often connection drops). But the fun starts when I am not online (one of the simplest things for the communication platform with normally 60% of the day duration). Why should anyone care of searching any reasonably maintained IRC bouncer (or grep through eavesdrop logs), would should anyone pay for a simple mobile client? - issue tracker. You know yourself... i personlly would have prefer to use githubs issue tracker or just go back to launchpad for all project but we all have different prefrences. i personally prefer how we track issue upstream to how we track them downstrema for example with a mix of bugzilla and jira and trello concuccnetly.
having one tracker for everything or possible 2 if you work on multiple project that span both the story boad camp and launch pad camp is better then many but i still like githubs issue tracker alot. since we dont use github for development though using it for issue tracking would send mixed messages.
Onboarding new people into the OpenStack contribution is a process of multiple months (so many times done that, also with all the Student programs we do). Once you are in it for years - everything seems to be absolutely fine. But entering this world is nearly a nightmare.
that depend on where you are coming from. my experince with on boarding interns who were on workplacment was we coudl normally get the up to speed in 1-2 weeks. most of the time was not spent on irc or email or gerrit. that was the simple bit that they got more or less stait away (espically if you teach them to use git review) the challange with onboarding was always the scope and explain how the different parts interact and get ther first verions of openstack installed so they could start working with it. although we normally started smaller with just cloning a project and runing unit test with tox. then a devstack install and then the rest.
I do not want to say - let’s change everything at once (or anything at all), but if we have chance we should not abandon idea of doing things better this time. In a daily work we all swim in workarounds we did for nearly everything.
improvment are good but many of the alternivies i genuwebly dont think would actully be an improvment over what we have today. irc is proably the case we might have the most to gain in but many of the tools like slack would be a regression in fucntionality since we would loose the much of the ease of use and isntead gain unwanted feature (posting documents and images inline in messages) and higuer resouce requriemetn to run the clients. i agree that searchabilty fo the irc logs is non existing but sendign a simple url to the irc logs when you know wehre it is simple. i.e. putting a link to the converstaion i just had with someone about a feature i a gerrit review comments is trival i go todays logs and scoll down to where the converstaion was. most of the alternitive loose that share ablity or require you to have an account to view the logs or partisapate. anyway i do think this is off topic and if we make any desisin on this it need the tc involvment as it will affect all projects. for now i dont think we should change all our ways of working and i dont generally think the tooling we use is bad. we can make imporment but i generally think that we are using some of the better options that are avaiabel today already.
Cheers
On 19. May 2021, at 16:56, Kashyap Chamarthy <kchamart@redhat.com> wrote:
On Wed, May 19, 2021 at 01:49:33PM +0000, Jeremy Stanley wrote:
[...]
In past years when the stability of Freenode's service came into question, we've asserted that OFTC would probably have been a better home for our channels from the beginning (as they're more aligned with our community philosophies), but we ended up on Freenode mostly due to the Ubuntu community's presence there. We'd previously been unable to justify the impact to users of switching networks, but there seemed to be consensus that if Freenode shut down we'd move to OFTC. The earliest concrete proposal I can find for this was made in March 2014, but it's come up multiple times in the years since:
http://lists.openstack.org/pipermail/openstack-dev/2014-March/028783.html
Honestly I'd be concerned about moving to a newly-established IRC network, and would much prefer the stability of a known and established one.
Yeah, moving to OFTC makes a lot of sense. FWIW, I've been participating on #qemu and #virt channels on OFTC for more than six years now and I've rarely seen glitches or random drops there.
(Also, agree with Dan Smith on "move one step to the left", i.e. low-to-no friction.)
-- /kashyap