пт, 21 мая 2021 г. в 18:48, Clark Boylan <cboylan@sapwetik.org>:
On Fri, May 21, 2021, at 7:28 AM, Andrey Kurilin wrote:
пт, 21 мая 2021 г. в 17:17, Jeremy Stanley <fungi@yuggoth.org>:
On 2021-05-21 16:30:40 +0300 (+0300), Andrey Kurilin wrote: [...]
Why everyone points to third-party solutions for those who don't like IRC? Why the modern chat-platform can be used as a main solution and those who want IRC should look for third-party bridges to make it work in the good old way? [...]
It's all a matter of perspective, and you're paying attention to how it's phrased by people who are already using IRC (the bulk of our current community). You could just as easily phrase it as "some projects are moving to Matrix, but taking advantage of the available Matrix/IRC bridge so that users of the old IRC channels aren't left behind." Technically the solution is the same one as "let's recommend a Matrix/IRC bridge to anyone who wants to talk with people on IRC without using IRC." The main difference is in how it's documented and communicated. -- Jeremy Stanley
It's all a matter of perspective
It is True without context. I may be wrong, but I do not remember any big change in OpenStack community (maybe only the 4 opens and nova-net -> neutron, but it's earlier days). If something was used/developed/decided 10 years ago, we will live with that forever.
I feel like a big part of this is lots of people have very grand ideas, but no time and willingness to invest in them. We have done a number of large changes since nova-net -> neutron including an entire Zuul v3 rewrite,
yes, Zuul is another thing that occurred to me as soon as I sent an email, but it was late to change anything.
the massive Gerrit upgrade last year, deployment and use of global-requirements and later constraints and their later modifications, fully automated most details of the OpenStack release process, and so on. I am sure there are many more, but I've got a bias due to the things I'm exposed to.
Most of the listed things are just required things to do. It doesn't simplify them or decreases greatness.
A key detail with all of those is they found champions who worked through them, got necessary consensus and implemented the changes.
Sure thing. I do not want to hurt anyone, I know that there are a lot of people in our community that implemented a bunch of great, necessary and important tasks. And thanks to everyone for that. The problem is in the over-complexity of doing such changes.
That is why I read all suggestions of using matrix as "if you don't like the chosen way, we are very sorry, but please find a way to leave with it. this is the way." :)
I would characterize it more as "if you don't like the chosen way and have no willingness to help change things then it is unlikely that anything will change". From my (again biased) perspective it seems more and more that when people show up with ideas there is an assumption that someone else (often me) will simply whip something together for them and when that doesn't happen it is because the idea is rejected upfront rather than needing investment.
I understand what you are talking about, but only one or two emails mentioned the complexity of implementation and load of infra-team. Most emails of the topic cover only theoretical point of view "whether we want one or another solution".
Matrix as an IRC alternative has been brought up a number of times in the past, but it has always lacked someone or a group of someones that are able to PoC it, determine what would be necessary to switch, make the necessary changes, then guide the project through a transition if the decision is made to move. This isn't as simple as registering on the service and joining channels either. You'll need ops/moderators, channel management, updates to existing bots that people want to keep, privacy policies may need to be considered, etc. The suggestion to use the matrix IRC bridge is a good way to simplify all of this though.
For this reason I think it would be useful to shift the conversation back to whether or not Freenode is viable going forward. If the consensus for that is "yes" then we start a completely separate conversation on whether or not we want to move to an alternative protocol and take our time. If the answer is "no" then it is probably best to make an "easy" move using consistent tooling for now, then start a conversation on whether or not a move to another set of tools longer term makes sense separately.
I dislike questions with binary answers. Such questions are too limited. For example, if you ask me for a binary answer for the topic - I would just ignore answering because I use IRC rarely and only for openstack purpose. Should we move from Freenode to another IRC network? I don't care much, that is my answer... But again all of these options require effort and effort requires humans.
Let's try to address the immediate problem first without conflating issues which only causes confusion and will make it more difficult to solve the problem in front of us. Then once that is behind us, bring up the other discussions in a productive manner (this includes acknowledging the other side might have an opinion worth listening to and that the other side doesn't make choices simply because they have grown long beards).
Note: I've addressed some of the other ideas in the larger thread in this response, but they aren't necessarily the views of those I am directly responding to.
-- Best regards, Andrey Kurilin.
-- Best regards, Andrey Kurilin.