Maintenance Working Group Meeting (2024-02-01 08:35 GMT-5) - Transcript

Attendees

Adolfo Brandes, Awais Qureshi, Brian Mesick, Feanil Patel, Felipe Montoya, Jeremy Bowman, Jeremy Ristau, Kyle McCormick, Maksim Sokolskiy, Michelle Philbrick, Navin Karkera, Piotr Surowiec, Robert Raposa, Sarina Canelake, Tim Krones, Tobias Macey, Xavier Antoviaque

Transcript

This editable transcript was computer generated and might contain errors. People can also change the text after it was created.

Feanil Patel: Need that later. Great.

Feanil Patel: Just keeps me alright.

Feanil Patel: that's hello It's good to see everybody welcome. I'm glad to see so many people here. The agenda of sort of planned items I have is not too large. But if you have other things you want to discuss, please add them to the discussion topics. The link is in the slack Channel maintenance and I've dropped it a couple of times in the message here, but I'll just drop it one more time. So please take a look at that.

Feanil Patel: So the first order of business

Feanil Patel: is I wanted to as the group sort of kicks off we've started by.

Feanil Patel: I've started by creating a priorities spreadsheet that you can take a look at I just dropped in the notes on the Wiki page and that spreadsheet is meant to prioritize sort of with an eye towards product which repositories I think we want to prioritize for maining. Sort of medium to long term within the open edx project. This was in consultation with genomicowski and Chelsea Rathbun from the product working group who work at accent and hopefully I'm hoping to use it to find maintainers for

Feanil Patel: all of those the repos sort of in priority order. So to back up a little bit. I think maintenance pilot was sort of working from this trying to find a maintainer for some things eventually and I want to operate more towards everything has a maintenance pretty quickly and I want the maintenance to help us maintain those repositories sort of for all of the things that means which is not only doing the day-to-day work but also to help support any upgrades or major changes that need to happen. I'll python upgrades Django upgrades react upgrades all those things so, That's kind of where I'm hoping the maintenance and maintenance program moves and I'd love to sort of hear any thoughts people have on that.

Feanil Patel: And if you want to take a minute to look at the spreadsheet, that's totally fine. Let's take a minute to do that.

Feanil Patel: And you want that second tab I think is the most maintenance priorities tab is the one you want to be looking at if you're in there.

Tobias Macey: and I don't know if this is the right time to bring it up, but another element of the maintenance work is that Might tie into some of this stuff but is some of the system level dependencies thinking in particular Mongo 4.4 is end of life this month now that we're in February.

Tobias Macey: But we're in the middle of the Quince release cycle. So just thinking sort of things like that.

Feanil Patel: Yeah.

Tobias Macey: And then also I think Ruby 3.0. Is either end of life or coming to be end of life and we're still depending on that for the Forum service? And I'm sure there are other elements in there that I haven't. Identified readily, but just thinking some of those more cross-cutting elements as well.

Feanil Patel: Yeah, I know the default Ubuntu that all of their tests are running on is going to be end of life later this year. I think we have till Redwood for that one or not. So red would tell sumac for that one. But Jang the python Ruby and Mongo are all things that we need to deal with sort of more immediately.

Feanil Patel: The Mongo stuff. I believe only impacts the forums and edx platform. So hopefully we can sort of get a focused update there pretty quickly.

00:05:00

Piotr Surowiec: So regarding Mongo, we are actually doing some testing right now. We have upgraded mango 27.0 and one on our instances and…

Feanil Patel: Okay.

Piotr Surowiec: we're doing some checklist Starbucks. I'm not sure how extensive it should be to call it done. generally

Tobias Macey: I think it was with the 5.0 release. They added a backwards compatibility layer so that they can be API compatible with I think 4.4 and forward regardless of which actual underlying version you're using.

Piotr Surowiec: Right so far. We haven't run into any compatibility issues and we decided to go straight to 7.0 because 5.0 is supported that deal October I believe.

Feanil Patel: Yeah, so switching the test over I think would be a really useful step for edx platform at least running the test with both versions would be a really useful thing that test infrastructure I think is currently managed to you Jeremy ristau. And so I don't know exactly what it would take to get Tested with both versions of Ubuntu, but it'd be useful to learn to understand sort of what that work is and if we can help with that in any way.

Jeremy Ristau: I'll find out.

Feanil Patel: awesome

Feanil Patel: But can you add that to the action items at the bottom of the Wiki page? Thank you.

Feanil Patel: Okay, Yeah, thank you for flagging that the Mongo stuff. I feel like we should get ahead of a thing that I will mention that I know we're behind on is el super behind on it at the version that we're on it's been out of support for a while. I think it's sort of been deemed fairly low priority by virtue of nobody being mad that it is the way it is. So if you have feelings on how behind we are, I think that's a thing that we should. Talk about and maybe figure out sort of how to help priorize. In the meantime. I think focusing on some of these other major ones like python Ruby Mongo feels like the right choice, but I do want to come back to elasticsearch and figure out how we can either move off of it or update it to a version of open search that we can sort of moving forward.

Tobias Macey: Yeah, I know that there was some conversation about what's actually involved in that and I think that there was some Discovery work that needed to be done as far as how many different pieces of the ecosystem are touching elasticsearch. I don't know if that has been completed and documented anywhere.

Feanil Patel: Yeah, I know you did some Discovery work and I'm looking at Robert to sort of see if you can find that. I know some of it was shared with me, but I forget how much of it was made fully public. So it'd be great to share. I think there was even a sort of plan of here's what we should do in the different places. We're discuss where elastic search is being used written up. It'd be great to make that public so that we can potentially prioritize that through the community to get different pieces handled.

Jeremy Ristau: I thought that there was a GitHub issue related to this and I added that content there at one point.

Feanil Patel: Okay.

Jeremy Ristau: I think I exported it to PDFs and it was attached to that GitHub issue. If not, Let me know I can take that.

Feanil Patel: Yeah. Yeah it'll probably be the Dapper issue. So let me

Jeremy Ristau: But yeah, you're right around discussion Infinity did some work and around course management. the TL team some work, so I think those were attached to that good issue already. it was a while ago, so

Feanil Patel: Yeah, I see the attachments Jeremy. Thank you. Let me link that ticket here.

Feanil Patel: cool

00:10:00

Feanil Patel: there next okay yeah, so Django for two is hopefully Done as of in the next couple of hours edx platform I think is the last outstanding repository that has master or not running Django for two. Coins I think is running for two which puts us in this weird dissonant position where our release is running one version and Masters on a different one, but I'm hoping that the testing that it's going out to production to you right now and assuming they don't find any major bugs with it. It'll hopefully remain on master.

Feanil Patel: And I believe that's the last place that needed it, but somebody can correct me if I'm wrong.

Maksim Sokolskiy: And Ecommerce also on  Django 3.2

Feanil Patel: commerce is still in Django 3.2.

Feanil Patel: Robert and Jeremy, do you guys know if that work is being prioritized by arbit bomb to upgrade that?

Robert Raposa: no for sure, but I don't think so. But I guess Jeremy.

Jeremy Ristau: Yeah, I mean you're correct. They're working on edx platform right this moment. So That's…

Feanil Patel: Okay.

Jeremy Ristau: what they're focused on. I don't see.

Awais Qureshi: Testing different PR step by step.

Awais Qureshi: So the PRS are ready. For testing. Yeah.

Feanil Patel: Okay.

Jeremy Ristau: Yeah, I think the PRS are the easy part there though, right and stuff. The validation is the larger effort.

Feanil Patel: Okay.

Jeremy Ristau: My list says that eCommerce worker and at a platform or the only three Repose left for Django, so that's great.

Feanil Patel: So PR is ready for testing but It sounds like we need some people to test because I believe that QA team doesn't exist anymore. Is that right?

Feanil Patel: To other organizations that use e-commerce. Do they have capability to test commerce? updates

Felipe Montoya: We have a team that USC Commerce or last thing to join in that effort, but I can promise that they will I don't know what their schedules are what their screens already.

Feanil Patel: Felipe will you add a little action item for yourself at the bottom to get back to me about that?

Feanil Patel: Thank you.

Feanil Patel: cool All right. Yeah, I mean longer term commerce worker and it's related mfes or in my mind things that we want to deprecate and remove from the open FX platform. We've done some work with the woocommce. integration as a proof of concept for making sure that other Commerce systems can integrate with the open edx platform. I know it's missing a lot of features still if there are things that you think you need that commerce provides, but that the new woocommerce system does I would love to get that enumerated somewhere and shared there's a deprecation ticket for e-commerce, which would be a great place to share that information.

Feanil Patel: and I think Yeah, so yeah, if you're using a Commerce and you want it to be around please chime in on that deprecation. Otherwise, if we can't find enough people to test it that might accelerate our deprecation of commerce because I'd rather get rid of it than Support testing and getting upgrades to A system that we don't plan on using in the future.

Felipe Montoya: It should recovers take a root of the deprecation. There are. Requirements or in it it is required for all things Discovery and pieces of Alexander price. Are we thinking removing those dependencies and…

Feanil Patel: Yes.

Felipe Montoya: going around that I would

Feanil Patel: Yes, I think any coherent removal of e-commerce will probably require a disc a discovery is such a knowing word Discovery into all of the places where it is integrated and either disabling or future flagging whatever capabilities it supports on the course Discy front. I am doing some research now to see if there are pieces of course discovery that we can

00:15:00

Feanil Patel: we can sort of serialize into the edx platform this is an super early draft PR So here it means.

Feanil Patel: so I just posted that the pull request which is an ADR to start serializing some of the course catalog data in the earth materializing some of it into the edx platform repo as new models. Which I think will help make Discovery less of a requirement for more features.

Feanil Patel: One of the many issues with Discovery is I think that it's models for courses and programs is very tied to the way to you represents and those things and the processes around those so having some simplified models in the platform that can be backed by the more complex thing. But that allows for most of the simpler use cases to be fulfilled seems like a useful Improvement. So That I finished writing it yesterday. So there's probably a lot of things wrong with it. If you want to have early opinions feel free. There will be more of an announcement once I get a little bit more time with it for sort of general public review as well.

Feanil Patel: Yeah, the plan is that I think e-commerce maybe goes out the door, of course Discovery maybe becomes more optional. And that should hopefully simplify a lot of the edx platform because you don't need as many of those Integrations for more features. This is all in the future. So it kind of depends on How many people are excited about those things how many people can help build out those things But that's the general Vision that I want to make more public and more get feedback on.

Felipe Montoya: Here. Thanks.

Feanil Patel: Yeah. Yeah, let's see. So the next major upgrade that is on the horizon is the python 38 to either 311 or 3 12 upgrade. I don't. and always maybe more about your Jeremy ristau know there was some research happening about whether we were ready for three twelve or not. And I don't know what the state of that is.

Awais Qureshi: If we started some discovery. Last month, but I will share the update with you, but we didn't find we didn't do any can It's a detailed research at the time because we were busy in some other tasks. But what we started some little bit debugging at the time.

Feanil Patel: Got it.

Feanil Patel: Yeah, so I think the sort of big question in my mind is whether we can jump straight to three twelve or if there are any compatibility issues that prevent that I know that a lot of the supporting packages haven't declared through 12 support, but as long as there are no breaking Backward Compatible changes between 311 and 312. I'd much rather jump to 312 that gives us all much longer maintenance window. Plus performance boost that I'm really interested see the impact of in edx platform.

Feanil Patel: 

Feanil Patel: that will need more Discovery and I think that this upgrade will happen before we have maintenance for everything and can distribute the work more easily across maintenance for all the impacted repositories. So I think this is one where We will need more support from the community to be able to get work over the line. So my hope is that once we've decided A version that we're targeting that we can take it up that work and that people across the community can pick up that work across various repos.

Feanil Patel: It should be pretty straightforward. The first step is probably going to be to enable testing on the newer version everywhere. And fixing any issues with and sort of identifying where the issues exist today.

00:20:00

Robert Raposa: This might be a silly question. But if we target 312 can we just Have the exceptions that can't make 312 go to 311 instead.

Feanil Patel: Yeah, it depends on…

Robert Raposa: Thank you.

Feanil Patel: where in the libraries it is. But yeah, my preference is to start Target 312. It might be easier just to start running tests on both. because that change shouldn't be that hard to make across the system if you're running tests on one running elimond both shouldn't be too hard because most of our system is toxified anyway,

Feanil Patel: So my thought is if we can do that then. We can hopefully make this sort of defer that decision and make it much in a much more informed way than we can right now.

Feanil Patel: Okay.

Feanil Patel: Thank you. The node.js upgrade.

Feanil Patel: And I think we're already running into issues with 18 not being new enough. One of the things we're seeing in the node ecosystem is that being LTS supported is not sufficient because our third-party packages we depend on are already dropping support for 18 because there's a new LTS. So I think as soon as there is a new LTS there are packages that start dropping support for older lts's. So this is an issue that might exacerbate are sort of update and maintenance problems in future as we may need to jump to 20 sooner than we expected as a result of just having basic functionality available.

Jeremy Ristau: Is there any way that you could enumerate? Somewhere what basic functionality means?

Feanil Patel: So to real quick we're seeing that things that platform build and the front end platform are depending on some of those. haven't dropped it yet, but some of the dependencies for this semantic release package that we use to release across the whole system. I've already dropped support for a node 18. So we've had to sort of pin back to older versions.

Feanil Patel: I think nothing that we directly depend on in front-end platform has dropped to support yet. but front end Engineers are starting to see it in other related packages that we're starting to pin back. I'll see if I can find a more definitive list.

Feanil Patel: Okay.

Feanil Patel: So yeah, I think there's sort of more maintenance stuff to talk about here. But I kind of want to get to Felipe's question which I think will be a much more interesting topic which is a cc requirements for new repos that need maintenance. Felipe you want to say a couple more words on that to get us going.

Felipe Montoya: Yeah, basically how it worked now is to be maintenance. You need to be CC with access to that report. which makes a logical sense but the process to get to a cc in repo requires that you contribute to every poor have something to show for it and that sort of creates a loop of there's no way to get an antenna if we don't have maintenance that can approve it first and

Felipe Montoya: And this let us in a promise that we made to maintenance report. That was a front-end component front and MFE repo and we got into the chicken and egg problem and so for me to solve it. We started by continuing to that report, but that contribution hasn't materializing Masters. So we have nothing to show for it. So I can't nominate the CC and container And this report had someone. On the inside that could approved years some of them because I'm seeing in the list. May not at this point and that could be immersive. only something needs to be like I don't know first track of this process or we need to do some changes.

00:25:00

Feanil Patel: Go for Xavier.

Xavier Antoviaque: yeah, I think that's a good mind actually and Especially given right now the number of repositors for which one to find maintenance sheep probably need something better than that. could we simply say that anything which doesn't have a maintenance by default can be requested by any existing or…

Robert Raposa: them

Xavier Antoviaque: at least an existing contributors or maintenance can vote someone You don't have a big historic Etc. Just because there is no maintenance right now. That's all the chicken and egg problem.

Feanil Patel: Yeah. other thoughts

Robert Raposa: I mean,…

Feanil Patel: remember

Robert Raposa: yeah, the only other thought is an additional possible exception of Here's some Metric of how many commits it has so that we can actually see things that are not touched at all and have Either what is that? It just recommended or you can simpler. Solution for that of any Quark contributor who's working in other places can just be like, all right. I'm happy to maintain this thing that almost never changes.

Feanil Patel: Yeah, yeah. my current thought on this is like,…

Robert Raposa: even easier

Feanil Patel: yeah, I think the current process will be a good long-term process, but it is not a good short-term process to solve this disparity between what we have right now. That is under maintenance and what we have that needs maintenance. So I am very open to something where existing core contributors can claim sort of anything. That is unmaintained. I think that's an easy one. My concern with them nominating other people or nominating people that are not currently for contributors.

Feanil Patel: the only thing I worry about is sort of Providence and trust with those people. So I think if they are part of organizations that we already have organizational cla's with I think that is a pretty easy thing to do but I think sort of nominating anybody off the street. I would be more concerned about sort of trust issues if that makes sense.

Xavier Antoviaque: Yeah, I didn't mean to completely bypass those processes especially for new people. I mean They would still in my mind whether it's now on the long term for the whole process is just that because there is not much history with that repository. there is nobody to review contribution. So nobody can acquire that experience. Then we live to the requirements of Prayer of your mission to that specific repository to become maintenance,…

Feanil Patel: Okay.

Xavier Antoviaque: but it doesn't mean that we don't review the contributor itself that thanks.

Feanil Patel: cool Yeah, I mean, I think essentially what we're talking about and I think sort of what I would like to see is a accelerated maintenance maintainer core contributor process where somebody can essentially ask for both on a repo that is listed as unmaintained on this list. And we can approve them based on.

Feanil Patel: any PRS across other systems just to show their sort of caliber even if not in this existing system ideally in the same language set that is sort of the majority of the repo. but Thanks, Jeremy for dropping in. but yeah, I think If we can sort of do caliber, it's a evaluation with other repositories that have either python or front end to code in it and totally okay with that. This seems like a really useful way to get more things under maintenance faster. Does anybody see issues with that?

Kyle McCormick: I just want to be clear. I suggesting we the Bar for people becoming core contributors in the first place or…

Feanil Patel: cool

Kyle McCormick: lowering the bar for an existing core contributor to get rights to a repo that is under maintained.

00:30:00

Feanil Patel: yeah, it's I think so. Yeah for existing core contributors. I think they should be able to ask for maintenance on any repos that exists for people who are not currently cork contributors. My suggestion is that we don't lower the bar but we sort of temporarily lift the requirement that they have to have worked in the exact code base that they are asking for access on so we would essentially judge their caliber in other pull requests across the open edx ecosystem. So if they worked in platform, but they want to sort of take maintenance for an X block we can look at their platform contributions, but we should still be evaluating their caliber at the level that we've currently been evaluating it just not in the location that we've been evaluating it.

Felipe Montoya: Does that mean we could also propose sorry?

Sarina Canelake: I think

Sarina Canelake: Just might my only concern is that we need maintainers. So badly that I am concerned that somebody's going to be I want to be a core contributor and maintain this repo and we're just going to be We need to maintain her. So I think just for the people who are core contributors who would be voting on that to just keep that in mind that we need to evaluate a potential person who wants to be a maintainer as a core contributor outright. we can't just accept people because they're volunteering to do a job that nobody else wants to do.

Felipe Montoya: I was gonna have something in the same direction, but just one step less and is what if somebody is not yet a core contributor, but it has shown caliber in other contributions before it has contributed to its better form. Or I'm calling them the big reports. But then once to maintain I don't know it's more library that is written in title. And that hassle similarities. Would that be? allowed in Disney process

Sarina Canelake: Yeah, I think that's exactly what finial was saying. I think that the difference is if that person who's contributed on a python back end code to edx platform wants to maintain an MFE. you don't have shown that you can have written a bunch of quality like JavaScript and react and stuff. And so that maintainership wouldn't make sense. But yeah if you've Contributed to edx platform and you've made really quality code contributions and you want to maintain a library that X platform depends on that seems perfectly reasonable.

Feanil Patel: Okay, Any other concerns people might have with this or changes or feedback?

Tim Krones: I would just be interested in knowing where the information about the new maintenance is going to be documented because right now I feel the information is very fragmented and there's overlaps and there's sometimes inconsistencies. And so just also from the ospr management perspective. I would like to know where I'm supposed to go to know who to pay for reviews right in that context.

Feanil Patel: Yeah, so I think the sort of source of Truth in my mind for that information should be backstage which is fed in by the catalog info.yamal information kind to ease the core can to sort of the work that you and Michelle do Tim. I'm also hoping to automate pulling it into this maintenance priorities list as a new column. so you can easily see which things already have maintenance maintenance and also who those maintainers are whether it be a team or a person I think But all of that should come from the catalog info dot yaml file, which is one of the first things when you're maintainer that you should do is make sure that that's up to date and it's already how to maintain a repository how to choose. the documentation

Tim Krones: And how does backstage poll information from those files? Is that something that it does on the fly or does it updated periodically, how does that work?

Feanil Patel: Backstage, I think updates at every couple of hours. So it automatically will pull catalog info updates every couple of hours. I think we haven't been making a lot of use of Backstage historically, but it's I think been a chick it's also been another chicken and egg problem. We're not enough things are there so we're not looking at it often enough so not enough things are there. So I'm hoping that as we accelerate the maintenance maintenance program and…

Tim Krones: right

Feanil Patel: get to 100% maintained all of that info will be in backstage for both you to look up and for maintenance to easily see what they maintain because one of the views is you can go into backstage and see all of the repos where you are maintenance either directly by name or on a team that is the maintainer of that. So it gives people much better their perspective. Michelle

00:35:00

Michelle Philbrick: I'm sorry if I missed this somewhere, but just adding on to Tim's question. It's kind of unclear right now. For the current maintained repos. for instance. I'm not sure if what squads or teams are still available and who I should be reaching out to so I feel like right now we're just in kind of a limbo because I know that you lost some capacity and we're looking into seeing who can maintain things. So I guess I'm just kind of wondering in the meantime while things are getting sorted out. Is there something Tim and I should be doing to help better push things through

Michelle Philbrick: 

Michelle Philbrick: should we post things that need review in this working group Channel or should I just want to make sure we're still pushing things through…

Feanil Patel: Yeah.

Michelle Philbrick: but we're not pinging people that aren't able to.

Feanil Patel: Yeah. No, that's a super valuable question. I think. Do other people have thoughts on this because I'll always have a thing to say so why don't you go first up yet?

Tim Krones: yeah, yeah, that was something that I kind of suggested internally to rcc's because before Michelle and I think we were operating on the assumption that we would always have to get approval from let's say that the two you owners or the two human maintenance to say that this PR can be reviewed and merged by a cc but I just suggested recently that CC's, could just offer to pick up the reviews in any case and then I guess see if someone from to you would still want to sign off on those but to not sort of lock the process right away. Which that has caused a lot of delays I think in the past.

Feanil Patel: Yeah, plus a thousand right? I think Sarina sent out an email earlier last week to essentially the same effect, which is if you're a CC on a repo, you should feel empowered to review things and merge things and I think Michelle and Tim if you've been prioritizing trying to get somebody to you, I don't think that that is a hard and fast requirement anymore. I think we should treat all CC's equally and anybody who is capable of doing a review should do it.

Feanil Patel: And if we're finding issues through that in deployments, I just asked that If things get reverted that the reviewers are also taking on the burden of helping figure out sort of what went wrong and how we can prevent that kind of reverts or rollbacks in the future so that we can sort of continuously improve our testing infrastructure and have more confidence in it. but yeah, I think any existing CC should feel empowered to start reviewing things in the places where they are

Sarina Canelake: Yeah, I'm going to say one thing on that though. I think if it gets reverted by somebody who wasn't involved in the pull request the person who did the revert needs to explain it's one thing for the reviewer and the merger of that pull request should understand why it was reverted but it's not on them to dig into it that's on the person who reverted it needs to explain why it was reverted.

Feanil Patel: Yeah, and I think Robert we could certainly use your help. I think on that in terms of culture over on the two you side. I think it's very easy to get things verted. But often we have to pull information out of the engineer who did the reverting after the fact that I think we could set an expectation that revert should contain information about why the revert happened that would help everybody.

00:40:00

Feanil Patel: You made it.

Robert Raposa: And that sounds very reasonable. In terms of a request. I'm also wondering if I could see us growing over a Time. a short list of things to look at while reviewing so for contributors of just higher risk areas and questions to ask and…

Feanil Patel: Yeah.

Robert Raposa: not the old list of a thousand things that one might be able to ask but just certain things that are just like this seems Safe for did I consider a toggle or not? and…

Feanil Patel: Yeah.

Robert Raposa: is the total, default something that

Robert Raposa: basically people who immediately get this whatever all the things that come up from these Retros of Roberts and World X could potentially go on to some shortlist.

Adolfo Brandes: Parents, there is one important checklist item. I'd like to

Adolfo Brandes: bring special attention to which works and Twitter

Adolfo Brandes: anything that goes into the platform? That is in the release has to work with tutor. So that's a very important one for everybody.

Feanil Patel: Yeah.

Feanil Patel: Yeah, Adolfo. I think I'd love to.

Feanil Patel: I think that's very reasonable. But also very amorphous in that I don't know how to validate that today and maybe we can think about …

Adolfo Brandes: It's easy past and tutor.

Feanil Patel: how we can

Feanil Patel: but I think if we had examples it would make it even easier for people to do the right things. But yeah, I think if test and tutor is an important requirement we should think about how we can add a version of that to our CI because there feels like a minimal version of that that we can do in our CI the images are building and can be run.

Adolfo Brandes: no, totally but I'm a big fan of automating things and

Feanil Patel: But yeah, I think that's the core point. There is a really important one which is tutor is how the community Open at X and so changes we make should be validated in that environment today.

Feanil Patel: Go forward Xavier.

Xavier Antoviaque: Change a bit the topic. So if anyone has anything else What does it take but I always wondering about to use side of saying do we have no Clarity and as to concretely what can be maintained what you is keeping what it is dropping or not. I know that some of that is in your spreadsheet anything but for example, I see a question mark on a text platform. that's kind of a geeky right. So do we know is still floating?

Feanil Patel: Robert do you want to answer?

Robert Raposa: you answer please and follow.

Feanil Patel: Okay, That sounds good. You can correct me around my understanding is that the answer that question is still in the air one of the things that I have on my to-do list is to take the list that is of things that are currently maintained by to you. And sort of reverify it with them. essentially doing an audit of the currently maintained items and confirming either that the existing maintenance teams or groups still exist and are still committed to maintaining them or if those things need new maintenance to sort of put them back into the pool for others to take on that work edx platform in particular is a thing that I think axim will sort of lead because of how big it is with myself and Kyle ideally taking on sort of the coordination role, but

Feanil Patel: My expectation with that ex platform is that we're going to lean extremely heavily on the ccs and if you're an edex platform CC, please help in any way you make sense and…

Robert Raposa: Hey.

Feanil Patel: we'll ask for your help in a bunch of ways that as we figure them out.

Feanil Patel: So that's kind of the sort of triage version of what the plan is?

00:45:00

Xavier Antoviaque: Actually, and so that's going to materialize progressively into that spreadsheet right like the other one gave on the huge.

Feanil Patel: Yeah, yeah. My goal is everything in materialize into that spreadsheet in the long term. I would love if that spreadsheet sort of got subsumed by backstage, but I know that a single view table is a lot easier to work with for now. So I'm gonna try to keep that up to date but the goal is that If somebody says I will take on maintenance of it. It's not true until it's actually in the catalog file. And I will pull from the catalog file to update the spreadsheet for those ownership data.

Xavier Antoviaque: All right.

Feanil Patel: before Kyle

Kyle McCormick: You said pull from a spreadsheet is that the spreadsheet we're looking at?

Feanil Patel: Yeah, it's The maintenance priority spreadsheet is it's linked in the first meeting item…

Kyle McCormick: Yeah, I'm looking at it.

Feanil Patel: if you want to take a look at it. Yeah.

Kyle McCormick: Is there like a column for current maintenance?

Feanil Patel: I'm going to add that. Yeah, no worries.

Kyle McCormick: Got Sorry for Miss that and That's probably superceding this page, which I think is what we were looking at before right?

Feanil Patel: Yeah, that's So the pilot data. Was what we were looking at before my goal is to populate the new column directly from catalog info. So any promises made to maintain something we'll go into the notes column. But then until it's materialized in actual changes on the repository. It's not true as a way of sort of like doing that validation up front.

Kyle McCormick: Got it. Thanks.

Robert Raposa: Step. Thanks.

Feanil Patel: okay good for

Robert Raposa: I call that my son.

Robert Raposa: So I haven't been a part of the actual. Action to you discussions yet learning about some of these details. I think there's a possibility loaded of

Robert Raposa: breaking down Maintenance within edit platform in Parts as well, but will…

Feanil Patel: Yeah.

Robert Raposa: but there's nothing that's been decided about that. sorry, as far as I know and

Feanil Patel: Yeah. I'm very open to sort of. People taking on sort of major ownership of major areas, especially high risk areas. and I think the perspective I have right now is if you have an idea, let's hear it and in the absence of ideas. It's going to be anybody who can help and feels comfortable helping can help.

Feanil Patel: Xavier I'd love your thought on a Communications and process thing sarinau as well like the temporary change for allowing people to take on maintenance My thought was I would just announce that on discourse along with a list of here the first tranche of repos that we're looking for maintenance on. And see volunteers that way are there other process things that I'm missing that I should be considering?

Feanil Patel: Yeah.

Feanil Patel: Yeah.

Feanil Patel: Yeah, and I think the python upgrade will help kind of put the prioritization through its Paces because if we can ify. Items that were low priority that we didn't necessarily have maintenance for that were blocking upgrades that'll help us better understand. what's sort of truly required to get a release out the door.

Sarina Canelake: Yeah, I think just when you post that being sure to explain a little bit about the expectation that you'd be need to become a court contributor and you need to have sort of this evidence of contributions. But it doesn't necessarily have to be in the same repo that you're volunteering to be maintenance like all that discussion that we had just Make sure it's not the case that somebody just raises their hand and…

00:50:00

Feanil Patel: Yeah.

Sarina Canelake: their person we've never heard of and then we have to break the news to them that we can't use their help.

Feanil Patel: yeah, do you see any concerns arena with the sort of secondary rule we were talking about of contributors from existing organizations that we have agreements with can be nominated if they have sufficient contributions within the open edx ecosystem, even if they're not currently a cc. Or do you have concerns with that?

Sarina Canelake: Yeah, I mean, I mean they should become a cc. I have no problem with that.

Feanil Patel: Yeah, yeah.

Sarina Canelake: I think that I was agreeing in the earlier conversation. I think…

Feanil Patel: but

Sarina Canelake: if you just make a post of here's a list of repos we need maintenance for we might get a person in the community who's never made a contribution being like …

Feanil Patel: yeah. Yeah.

Sarina Canelake: I can do that and then I don't want to be in that position of having to tell somebody who is volunteering in good faith to know.

Feanil Patel: No, I think being very clear about the requirements in that post is going to be huge. The other thing I'm wondering is there a place to ify That I should codify it more long-term than in the discourse post, but perhaps not in the moment. Robert

Robert Raposa: The piece that I'm not clear about in terms of what you brought up a couple of times is what's the difference between the current process to become a cc and what you're saying? You would have or is it just become a cc and then you can come over and

Feanil Patel: yeah.

Sarina Canelake: I think the difference if I'm understanding correctly is that you'd be nominated for cc and maintainer at the same time where I think formally what we've been saying is you're a CC on the repo you make contributions on that repo and then you graduate to becoming and maintainer and we have so many repos without CC's we kind of need to adjust this process.

Robert Raposa: so that

Feanil Patel: Yeah, there's also a second. The other difference is I think historically you ask to be a CC on a code base, which is usually a repository that you've already contributed to. And I think we're sort of loosening that to say if you've contributed to anything in the edx platform in the same sort of language or part of the stack that qualifies temporarily as sufficient contribution. And honestly the more it feels like maybe that should be a more General. Rule, but that's a later problem.

Robert Raposa: Yeah.

Sarina Canelake: I also think there's an argument to be made that if you've done substantive development on that repo in a fork. Know that you're knowledge of that repo by pointing at PR's in your own Fork.

Feanil Patel: Yeah, yeah.

Sarina Canelake: Could also be evidence.

Feanil Patel: That's a fair point.

Robert Raposa: 

Feanil Patel: So we've got a couple of action items. I think the next thing for the community to look out is it is the post from me about ideally a set of the first priorities. And if you're at organizations that have either CC's or Engineers that can help contribute to the maintenance burden. please

Feanil Patel: make sure they see it. I know and if there's anything I can do to answer questions, please let me know next time. I think we'll cover sort of the ideally currently this is a weekly meeting. I don't think that term that is how we will have this meeting but I think short term while we're still in this maintenance Carnival

Feanil Patel: meeting more often will be useful. If they're our thoughts about how or other topics you want to cover. I will be making the wiki space for this soon. But feel free to chime in the slack Channel if there are more things you want to talk about. Is there anything else anybody else would like to add?

Feanil Patel: cool All right, and I'll post the transcript and the recording to this meeting notes page. So, thank you everybody.

Michelle Philbrick: Thank you.

Xavier Antoviaque: that

Tobias Macey: Thanks.

Piotr Surowiec: Goodbye.

Meeting ended after 00:55:04 👋