What are the plans for Hubot?
joeyguerra opened this issue · comments
Hi. I was just wondering if there are any plans for continuing Hubot or not?
My guess is "no", can any Hubot dev confirm? @mistydemeo
I emailed this to the contact email displayed at https://github.com/hubotio, May 6 2021. I will update if I hear back - I should have linked to this issue and hubotio/hubot-redis-brain#40
To: opensource@github.com
Subject: Hubot - next steps
Date: 6/05/21, 3:40 pm
Hi there
A commonly used Hubot package (hubot-redis-brain) has recently started
failing npm audit checks due to a dependency that needs updating.
There hasn't been any update from the Hubot team in a long time, and I
believe Github's focus is now on the Probot / Botkit interfaces. There
is still a community of Hubot usage though, and I'm interested if you
have any plans you can share for how this community might proceed?
Would Github consider handing the reins over? I don't want ownership
myself, but I do want to see Github support a process to move towards
community ownership if that's plausible.
Thanks
Chris
cc @technicalpickles @mistydemeo - are you able to give any guidance / contact which would support some kind of community organisation here?
I'm afraid I no longer work at GitHub, so while I'm still a Hubot maintainer I'm not able to drive the conversation from that perspective.
Thanks for the reply! It might be fine to have any active maintainer, if there are ways we can support you in the issue queue etc.
Is the project still on your radar sufficiently that this would be helpful? How could some kind of community support work best for you (and/or other @hubotio maintainers)?
EDIT: Sorry, this takes the angle of question from "is there anyone home" to "will you take this on", and I'm trying to make the project more sustainable, which is at odds with pinning it on one person.
If there's anything people can do to help, let us know!
Is there another possibility then forking this repo in order to keep this alive?
I e.g. see lots of hello world issues by users already deleted and PRs to bump versions etc. I guess if some of the users could get the permissions to at least triage issues this could get a few updates. Or is there a successor which should be used instead of hubot?
@dominik-ba I'm also interested in ensuring Hubot gets at least the maintenance to continue operation. Simply unblocking version bump PRs would be a step in the right direction. Like you I'd prefer to avoid forking the project if there's some way to build on what exists.
Who might do something?
It looks like @xurizaemon and @dominik-ba are sufficiently interested to be asking these questions and thinking about alternatives. (Everyone in/on this thread, incl 27 👀, has expressed some interest!)
But why?
It's a working tool, with a small community around it. It's good for such a community to be able to land small fixes and security updates.
- https://www.npmjs.com/package/hubot - 19K weekly downloads
- https://www.npmjs.com/package/hubot-slack - 35K weekly downloads
- and others, just some random quick numbers there
What might work?
- Some kind of shared ownership / community governance?
- What might Github (license holder in LICENSE.txt) / original authors / current maintainers require to allow this?
- Could community volunteer maintainers step up in a way that didn't require ownership transfer?
- I don't even know what questions to ask about ownership transfer, I don't care so much who owns it provided that it is possible to maintain.
- Might Github / maintainers permit community sufficient access to merge changes?
Or ...
Or, as suggested, we might establish some kind of community fork. But nobody is leaping to that solution, I think because it divides what energy there is between multiple projects, and this project would still be the default target for those PRs.
Howdy folks! It's really cool to see so much community interest still in Hubot. I used to use Hubot a lot as well, and I worked on the docs site a bit many years ago, and I still think of Hubot as one of the simplest ways to spin up a basic bot for chatops.
Also FWIW, as a former Hubber I'm assuming that it is not likely for GitHub to release maintainership of Hubot anytime in the immediate future, if for no other reason then because there are still probably a number of apps and deployment processes running at GH on this repo (speculating a bit here, so I could be wrong about this).
So it may be worth seriously considering the fork option if you'd like to drive Hubot into the future (could even call it hubot-future or hubot-next. 😉 ).
The other option, of course, is to get on Twitter and get lots of folks to ask new(ish) GitHub CEO Thomas Dohmke to release this project to the community! (That used to work pretty well with Nat Friedman, I hear 😄). Again, my hunch is this approach is probably a long shot, but still might be worth a try.
There: I tweeted with a request for a response. Feel free to add your comments to the thread, or tweet Thomas in your own thread!
I think you're right, and it's fine for Github not to pass on stewardship of Hubot - that's a big ask tbh.
Feels like there's room for a community fork here, a nice thing about which is a fork can keep some parts (PRs) and not others (what's with all the hello-world issues?).
& in hubotio/evolution the project even comes with a (school-age, free to ignore) roadmap, which is a rare feature in unmaintained projects 😆
My brain parses hubot-future (hubot-dash-thing) as "a Hubot plugin that provides Future functionality" which is going to get confusing. nubot? yubot? Are we picking a name here?
A (docs) PR was merged recently in #1511 - good to see some activity!
Forking sounds like the right next step. I'm new here, so a bit clueless on how to create a fork (not git clone, but the social aspect of it) of an open sourced project and bring the community along. Any recommendations?
I love Hubot and have been using it since the Hipchat days. Later when the world moved to Slack I brought the Hubots along. Slack has written a kind of tutorial to migrate Hubot apps to Bolt, but it seems nobody's published a complete example of a Slack bot running Bolt for JavaScript with clear examples.
Since Hubot seems to have been abandoned...anyone have a good standalone example of a basic Bolt app that has all the main features of Hubot?
Just to clarify, I'm still using Hubot. I forked it and have been updating it.
@joeyguerra i'm Just Some Guy, so take my suggestions with several grains of salt, but I'd say next steps would be:
- rename your fork so that
a. it's obvious that development has moved elsewhere
b. you can create a new organization for it and not have the name conflict with this org - reach out to all the people listed as devs on this repo and invite them to contribute to your new fork
- reach out to anyone who has submitted a PR to this repo in the last ...18 months? and discuss merging their PRs into your fork
- reach out to anyone who as committed to their own fork in the same time period as above and invite them to submit PRs to your fork
- announce the fork on various platforms (eg HN, maybe some Reddit communities)
I'm on board to help out with this kind of administrative work (i have next to 0 modern JS experience, so that's all I can contribute)
@joh6nn this plan is actually quite helpful. Thank you.
@joeyguerra if you're keen to take it forward - I'd be much more confident if the changes on your fork had clear intent, eg by taking the changes as they are and reworking them into MRs for each feature or bug fix applied.
(What I'm referring to here is the current commit history when I last looked at the fork - i can't find it from my phone right now, so will add a link later! Hope this makes sense.)
If it's time to create a new org, we can do that by lifting and reviewing the changes into that new location. You've clearly done lots of work, and this process would document that, ensuring users could know with confidence what they were evaluating in switching over, and that the changes were reviewed.
I realize this is extra labor which you don't owe anyone, but it might really help viability of the fork going forward, which I'm enthusiastic about. Are you up for that?
Like @joh6nn, I'd be prepared to help with this.
Yes. I think that’s the right approach as well. To clarify:
- new org
- Fresh fork into new org
- MRs to collaborate on changes
Any ideas on the new org name? Naming is hard.
Btw, #hubot on the fediverse.
Let's just call the new org "hubot-new" or something, and make an issue for naming over there. Github organisations can be renamed (right?), so we can defer that and deal with it later IMO.
That also gives us something for the new roadmap.
Ok. I intend to create an org called hubot-new, add you to it, and fork the repo today.
Here's the new org/repo https://github.com/hubot-new/hubot-b
@levenleven since you recently merged to this project after a long period of no updates or response from Github, I am pinging you into this discussion to make sure you're aware of plans to explore continuing Hubot support in the new organisation linked above.
I think Hubot users would love to see Github's engagement, and we're also prepared to put some work in ourselves. If Github has plans, now is a great time to clarify them :)
Update: In touch with @technicalpickles to review, merge #1581, and create a new release.
@joeyguerra reached out over email, and we've been discussing what to do there w/ @xurizaemon and @joh6nn
They have been working over in https://github.com/hubot-new/hubot to merge PRs, update dependencies and etc. After reviewing, I didn't see any reason to not just add them as maintainers, and get that work merged here before it diverges.
Since then, @joeyguerra started #1581 with all the changes, and we've tested, and I just merged it earlier today. I (manually) cut a release of v3.4.0 (because Travis wasn't working, and not all the commits followed semantic-release).
I've also added those 3 as maintainers. There is some followup to do post merging that, but I can make issues for tracking them.
While there isn't a plan per se, there is work being done. Is that enough to close this? Perhaps.
I think so. I was also thinking that we could continue "the future of Hubot" discussion in a new issue or in a PR with a ROADMAP.md file. Or continue discussions in the evolution repo.
I intend to the close this issue. I'll wait a few minutes :)