Please introduce new maintainers to the OMF team #788
Comments
|
If someone wants to take this over, keep in mind that there's numerous ways you could go about it. At the very least you would have to go through the issues. Some are simply open questions (like #780, where someone simply misunderstood the At the very least any questions should be answered, even if just with "sorry, we're not answering questions here, please ask at $somewhere". If you had the inclination, you could also tackle the bugs, and if you feel up to it the enhancements, but a lot of that is bound to be in the plugins people use. So there's a spectrum going from pure maintenance mode to actual feature work. An alternative is to officially stop OMF. Add a note to the README and installer declaring the project dead and archive the repo. That would be an open and honest way to go about it, if nobody steps up to it. If no work on OMF is being done anymore (and I'm sorry to say it doesn't currently look like it), that would really be in everyone's best interest. |
|
Interesting opinions, thanks for starting this discussion you two. I've started using fish a very long time ago and created OMF then to help the growing community it had back then. After some years maintaining it I eventually lost the energy to continue supporting the project and so the team reduced to @bobthecow @derekstavis @sagebind and @scorphus (sorry if I missed someone). It's sad to acknowledge that the project could be doing the opposite of its original intention which was to help new fish users. As I'm away from the community for such a long time, the answer for your questions will have to be from the people that are still connected to this. However if my experience dealing with open source is worth anything, I don't think that nominating a new maintainer would be the answer since I believe the process of finding a new maintainer is an organic process. As in, a user will submit pull requests or answer issues and eventually become a maintainer. As an OMF user, I'd like to thank every one that devoted their own free time to maintain this project for so long. ❤️ |
|
Hi @bpinto, thank you for contributing so much to the fish community. Fish for sure wouldn't be where it is without OMF. I agree that it's better if it's an organic process. Unfortunately, no one has stepped up to volunteer and certainly no one is helping answer questions/submitting PRs. At the very least, can we get someone approved PRs in package-main? I've been waiting for my PR to be merged since April 😂 |
|
Saw this thread since Bruno pinged me. 😉 I do feel a bit bad since I haven't been helping to actively maintain OMF for over a year now, since I simply don't have the motivation nor bandwidth to do so. I don't feel the need to go into detail here on my reasons; feel free to ask me directly though if you like. I am still a big fan of Fish shell, and to some degree OMF, and still use both daily for now. I concur that it would be ideal if more maintainers could join, but it definitely is best if maintainers are chosen naturally from contributors already active in the community and with OMF. I'm not sure how we would best solve this particular problem. I do not think that archiving the project is necessary, as it is still rather actively used. In addition, except for a few notable issues and feature requests the core framework is basically "feature complete" and doesn't seem to me like it really needs much active development to continue working. An exception to this would be the main package registry which appears to regularly receive submissions. Maintaining a quality package registry is not a small effort and does require at least a couple active maintainers. Note that |
|
This is definitely a project that I would certainly not like to see dying. And, v7 was released just a few months ago, by yours truly. I wouldn't call it dead. I use Fish with OMF on a daily basis. They help me working with so many other things. Things that keep me busy. OMF needs all our help. As it is important not only for us but also for the community, I think all of us should donate a bit more to the project. One doesn't have to be a maintainer to do so. PRs can be submitted and/or reviewed. Those issues are public and any of us is allowed to help a fellow user understanding why the installation doesn't work or why is it taking so long to load (might be a slow plugin with a faster alternative). We all benefit from any improvements made. I agree with Bruno that the inclusion of new maintainers is something that happens organically. Also, the addition of a new maintainer may as well change nothing. We’re all volunteers. Regarding packages, a plugin does not necessarily need to be added to IMO, Oh My Fish! lives and is far from being archived. And, even if it would ever be, redirecting to Fisher should not be an option — for historical reasons. |
|
Hi @sagebind, thanks for temporarily returning to chime in. No please don't feel bad. It's not like you were getting anything in return for maintaining OMF. Especially once it's feature complete, there's only a huge volume of questions and hard-to-pin-down bug reports that are high effort / low value. In light of what you said, I am changing my stance. I also think the only thing we really need to maintain with OMF is the package repository. As a plugin author myself, this was my main frustration. It felt so unofficial to tell users to install with a long GitHub url vs If we could update omf install to assume the https://github.com/ prefix, that would solve this problem and reduce the load of PRs coming in to packages-main. @scorphus Touche. Please accept my sincere apologies for claiming that OMF is dead. Yes, v7 was a pretty great release and it was only half a year ago. You've been carrying the burden of maintaining OMF on your back and you deserve the fruits of that, definitely at least to see OMF stay open. 👏 👏 👏 I rescind my vote for archiving it. |
I totally agree. The last release showed exactly that, with not so many new features if compared to previous ones.
That's a nice idea! |
|
Hi all, I apologize that this thread was temporarily hidden when I got flagged but that issue has now been resolved. Just wanted to give a quick ping here so we can continue talking if people have more to say. |
|
Hi everyone! I see that this is closed now, but I don't see if there is a conclusion:
I'm interested in it as the pisces plugin author. Please take a look at this discussion for some opinions on the current state of things for plugin maintainers: laughedelic/pisces#26 (comment). |
Hi! It was closed by the OP, it doesn't mean any conclusion was met — nor even necessary, TBH. 🙂
Contributions are more than welcome. And, as aforementioned, one does not need to be a maintainer to do so. This is OSS just like any other. Users are more than welcome to submit and also review PRs. They're open to everyone. Those with true intentions always step in with real, meaningful contributions.
Things remain as in “Oh My Fish! lives and is far from being archived” and also “Oh My Fish! is still rather actively used. In addition, except for a few notable issues and feature requests the core framework is basically "feature complete" and doesn't seem to me like it really needs much active development to continue working.”
Everyone is free and welcome to submit PRs to any project that either does not fit their needs or does not support one thing or another that they use. Most of the time, these needs and things are missed and used by many other, and they also benefit from that PRs. Espcially if one is versed in the tech stack used in the project. In cases like these, I don't see that many reasons why one wouldn't submit a PR to that project. I'd definitelly love to see more actions, less noise. |
Sure, they can open PRs, but they can't merge them. You have 23 open PRs, going back to 2016. These should be handled by either merging, requesting changes or closing them, ideally in a timely fashion. That's what maintainers are needed for, and that's why there's a request for them.
That's... not the point. Software can both be unmaintained and used.
There are a lot of things that could still be done with OMF to improve it. Also, there's just the simple work of handling incoming questions - things like #780, where someone just misunderstood which part of the instructions they were supposed to copy. It's been open and unanswered since september. This should also be handled. Note that I'm not talking about handling feature requests (tho if you're not doing feature work I think the term "maintenance mode" applies, but that's semantics), just simple user questions.
Okay, how about this: If someone were to go through all the PRs and review them, with just a simple "merge/don't merge" decision, would you act on that? |
|
Hi! Glad to see activity on here again. I closed it mostly because of inactivity. Let me reopen it to encourage discussion. I would love to see OMF become active again :) |
|
It's great to read your input, @faho.
Yes. But they can help by reviewing. Actually reviewing, requesting changes and even adding a few nit's
True.
Although the criticality of the majority of them is rather low, they should definitely be acted upon.
I'd like to know what. Honestly. Not that I disagree or think there's none. I simply don't know — a.k.a. pardon my ignorance here.
Agreed. Which can be done by any non-maintainer user. That happens to many OSS projects out there. Why doesn't it happen to OMF? I don't know. But I don't think it has to do with the number of maintainers it has.
I think a properly done review, with change requests if necessary, pinging maintainers for action (why not? a small reminder/nudge might be all someone needs) with a proper comment is beyond helpful. |
|
I would like to speak to this
I agree with @faho here. As a plugin author, I don't want to continue providing support for frameworks that are unmaintained and basically has no active presence from its contributors. Additionally, I felt very unwelcome to join the OMF community as a plugin author when my PR to get added to packages-main was commented on but unmerged for no less than 5 months. So when Fisher v4 no longer supported init and uninstall, I had to choose which framework I would support 100% and which one I would support only 95%, it was an obvious choice for me. |
As an OSS contributor I generally disagree with this. When I see a project with a long PR queue and old unanswered questions it definitely discourages me from contributing to it and as a user I start looking for more actively maintained/developed alternatives. Any potential contributor has some feeling of "is it worth spending my personal time to figure it out and implement" and in a situation like this it might seem that this effort won't get attention from those who can provide feedback and have the power to accept the contribution. This is the role of the maintainers, not other users. Sure, sometimes there are proactive users that monitor issues and might answer questions or give some feedback on a PR, but if they don't have at least triage rights they can't act on it. And if maintainers are late with their actions, it gives bad vibes to everyone involved. I suggest giving a shout out on Gitter (or wherever the OMF community is more active) and seeing if any of the previous contributors want to step up. This will at least give some visibility to the issue. |
|
So, since november things haven't improved at all. We have 8 issues opened after this one without any sort of developer response. Two of these are clear duplicates that simply need to have a theme removed (#801, #800), one is a problem in Jetbrains stuff that has nothing to do with OMF (#802), one is a simple user question (#803). These are all trivial. They take about five minutes each, but nobody has stepped up. Another is a certificate issue that can't be handled by non-maintainers because it needs access to the website. Then there are things like #685, a pull request that is already approved but waiting on a second approval 1.5 months later. I'm sorry to say, but this isn't a good situation to be in. Not for OMF and not for fish as a whole. Please think about improving things, even if that means admitting that you're done. I know that's hard, but it would be for the best. Or, if you wanted to continue OMF in maintenance mode, you all, collectively, would have to put in about an hour a week in total and maybe think about not requiring a second approval for every PR, because that is clearly not working. |
|
I haven't really used oh-my-fish much but is there a benefit to it over Fisher (which is very actively maintained)? I mean, apart from the fact its name is easily discoverable by zsh users. Wondering if it makes more sense at this point to archive this repo and direct users to Fisher in the readme. |
|
IMO I'd rather see all efforts be targeted into creating a built-in package manager inside fish shell. Fish shell is a mature project at this moment and having a built-in package manager would have a big and positive impact on all fish users. |
|
@oranja We probably need your help for the certificate thing. |
|
I wish I read more solution-oriented messages in this thread. Things like "Hey, guys! I see your oh-my.fish domain has an expired certificate. I guess the contact that has the credentials is no longer interested in the project. I happen to manage fishshell.com and I can create a subdomain to do the redirect to the install script. How about I create an Is that possible, @faho? |
That's a very valid point. I will accept it. |
We have. The problem isn't that the certificate is expired, the problem is that nobody is available to handle it in a timely manner, or to even answer bug reports about it. We've also said that either new people need to be introduced (which so far hasn't happened and I don't see any one stepping up, but I also haven't seen anyone ask where prospective candidates would see it - that might help), and I have mentioned that the old people could also come back, either fully or at least to keep the project in maintenance mode by answering issues (even if it's just either "not a bug in oh-my-fish, please look $here" or "yup, please send patches") and merging PRs. If none of that happens, I maintain that formally winding down the project would be, on the whole, preferable to having it around as a zombie that keeps being recommended to people who then end up asking here and not getting an answer. Also given the nature of OMF it needs active maintenance at least on the package front - #800 requires an actual change in the OMF repo even tho nothing in OMF changed, simply because the linked package disappeared. I know this is difficult for you, and please don't assume this is easy for me to write. It's just that I don't think the current situation is good for the users. |
That's not necessary, though. The certificate issue will be solved eventually. |
|
BTW, #800 is a clear indication that most of what motivated the OP to create this very issue is not a good idea (adding packages to |
What is packages-main for, then? It's not a list of all packages, because you aren't adding all of them. Is it a curated list of packages? Well, no, there's a bunch of archived things in there, some of which no longer work! So it's just a random assortment of "stuff". Which doesn't seem like a good state for it to be in, if I'm honest. And making it "all" packages (or close to it) requires accepting all the PRs that add packages, making it a curated list requires removing things that no longer work or that aren't "good" (for whatever definition of "good" you want to choose). So if you have packages-main, that requires ongoing maintenance. |
Do you think we should do that? And have some sort of automation that periodically removes unreachable/archived/untracked repos? Such as a GitHub Action. |
|
Oh god no! That just means all kinds of trash get added, and doesn't actually help users. If you want to make "everything" installable, do something like #824 and entirely get out of the process. Like... ditch packages-main entirely. But if you are to have a repository of "packages", it should have some sort of curation. You'll note that oh-my-fish/packages-main#151 removed packages that don't have an issue tracker. I'd also add the condition that
As the minimum vetting that actually helps uers. But... all of this means people need to do work - even writing automation scripts would be work. And the main issue here is that the work is not being done. |
I've been very much inclined to that. But sometimes I think that some curation — even if static or that seldom changes — is better than no curation.
That's pretty hard to guarantee. I mean, after inclusion. We don't do any version/sha1 lock nor anything else in that regard. |
I mean, yeah? But even just looking over it initially and then having a spot where users can report it is better than just... not doing that? |
|
yeah, removing packages-main would avoid maintenance problems. The community will curate packages automatically by using them. If there are broken packages in packages-main, and new ones are not accepted, that's not helpful to users. If you want autocompletion for plugins, why not just scrape GitHub for the top 100 most popular plugins? Though the url/repo syntax to install any repo should definitely be supported, to avoid the barrier of entry. |
|
There is now a security related issue with no response in 9 days: #873. |
|
This is now closed, but not actually resolved. OMF is still not in a good place. There are multiple issues open about it not being installable on macOS Monterey (#869, #870, #875 probably). #849 has had a "closeme" label for 8 months. There have been a total of 17 commits in 2021, of these 5 touched the actual main code and not auxilliary scripts or the README. 2020 had 32 commits. That's very very little activity. There has been a trivial PR open for a month (#883) and an accepted PR open for six (#863). I'm sorry to say this, but this isn't a good state to be in and I would like something to happen about it. Because it doesn't seem like new maintainers are lining up or the old ones coming back, the best thing to me seems to be to make the installer print a deprecation warning, after which the project could be archived unless somebody shows up to maintain it. This would allow the wider fish community to move on and find something else to line up behind. This doesn't diminish the work you've done over the years, it's simply an honest acknowledgement that you've moved on. |
Fisher filled that void long ago. Just wish it had a way to search for packages. |
Sorry this is off topic but I would argue that not relying on a central repository is actually better. Lower barriers to becoming a plugin and less work for maintainers to maintain the central repository of packages. An alternative I think is to start a convention of marking fisher packages with the tag #fisher-package or #fish-packages so one can just search GitHub using that tag. |
|
Sorry for reopening and chiming in so late. As a maintainer I'm open to adding more contributors that can merge PRs. I have not been able to dedicate the time I wanted to dedicate to OMF and I wanted to welcome anybody that want to become a maintainer. We should still keep a forum where we can discuss bigger changes before they are made. |
I'd like to throw my hat in the ring @derekstavis -- How do we make this happen? |
|
Hey, @gsornsen! It's great that you want to contribute to OMF. Like any other project, we obviously would love to welcome new contributors! And it is, like with any other project, something that should work organically. Making anyone a maintainer on GitHub is easy: just a flick of a flag, one button click away. But the real meaning of that is something else, and it goes a long way. Please, get yourself acquainted with the history of the project: https://www.google.com/search?q=oh+my+fish+takedown and one of the reasons why we take this a bit more seriously. Please, join our Slack workspace and start interacting with other contributors (mostly me, although @fidencio has just came onboard and has, likewise, shown very good intent 🙂). See you there! |
|
I'm sorry I have to write this, but it's been another year and things haven't improved. In the past year, oh-my-fish/oh-my-fish had 7 commits to the actual code (i.e. excluding CI and docs). The promised release hasn't happened (there have been zero code commits since that announcement). packages-main continues to collect PRs - including one that fails the extremely baseline policy I recommended, which is still open. I'm not sure there has been any maintainer activity on either of the two repos in the last month (we have nonsense issues remaining open), or much on the organization in general. The only thing I can see is a commit on theme-bobthefish in the last month and one merge in theme-agnoster about two months ago. plugin-bang-bang had a pretty serious bug report a month ago, with a simple PR the next day. All of this together leads me to believe that oh-my-fish is, in practical terms, as sad as that may be to people who've used and contributed to it over the years, dead. I'm sorry, I continue to say that it would be better for our users if you could admit it and formally wind it down. |
|
It's been almost three months since my last message and https://github.com/oh-my-fish/oh-my-fish/pulse/monthly shows there has been zero activity on this repo in the past month. Pull Requests on packages-main keep accumulating. This isn't a good experience for users who are lead to believe that oh-my-fish is the thing to install if you would like to see what fish has to offer (a lot of that is of course via the comparison to oh-my-zsh). I will add that we have removed the "third party tools" section from fish's documentation in large part because I believe leading people to omf in this state does them a disservice. I'm not blaming any person in particular, and I am trying quite hard to keep these messages as polite and on-point as I can, but it is very hard, because I emotionally do not enjoy any of this. I would quite like this situation to improve, whether that is by oh-my-fish starting back up or winding down, and I'm starting to believe that commenting here seems to not be helping. |
|
The last commit is now a year old and was to the README. It is the only commit since the announcement that there would be a new release. There has been no release in over 3.5 years. I am sorry to say that oh-my-fish is de-facto dead. I believe it is only honest to formally declare it so. |
Hello,
I would like to start by acknowledging that I was not around for most of OMF's development and am ignorant of how OMF got into its current state so I freely admit I'm not a great person to make recommendations for OMF.
That said, I do care for the fish community at large and am quite invested in it as both a user, a plugin developer, and a fish evangelist. And it seems most of the people in the OMF org are either completely inactive in the community now, or barely active. I am thankful for their contributions to fish in the past don't blame them as the most one should expect for something like this is to be nothing more than a year-long weekend hobby.
But I do think this is really sad for the fish community because it's really us back in so many ways. Likely, most users who check out fish first bootstrap their set up by using either OMF or fisher (which is also not super active) and probably get turned off seeing that OMF seems pretty much dead. And what's more no one is that no one is owning the approval process for packages in packages-main, which is also holding back the development and dissemination of new plugins in the fish community. Overall, these effects multiply: user growth is stunted -> less investment in the fish ecosystem -> less user growth -> fish ecosystem dies more. And we can make a huge dent in that vicious cycle easily, right here, by just injecting some energy into OMF with active owners. I'm sure there is a large pool of fish users to pick from who have the time, passion, and integrity to even partially own OMF going forward, that this is a huge lost opportunity.
In light of this, I would like to suggest to whoever is still around (@bobthecow?) and has the power to add new members to the OMF team to consider nominating a couple of new people. I don't know who to nominate or even how to choose them. I just want to get the ball rolling on this. Thoughts? Volunteers? Nominations?
- A concerned fish shell user
P.S. I am NOT nominating myself to be a maintainer. As much as I would love to, my bandwidth is limited and I've got my eyes set on another not-yet-released fish project.
The text was updated successfully, but these errors were encountered: