| *** emagana has joined #openstack-tc | 00:04 | |
| *** emagana has quit IRC | 00:09 | |
| *** emagana has joined #openstack-tc | 00:20 | |
| *** emagana has quit IRC | 00:25 | |
| *** emagana has joined #openstack-tc | 00:36 | |
| *** emagana has quit IRC | 00:41 | |
| *** emagana has joined #openstack-tc | 00:52 | |
| *** emagana has quit IRC | 00:57 | |
| *** emagana has joined #openstack-tc | 01:08 | |
| *** emagana has quit IRC | 01:12 | |
| *** emagana has joined #openstack-tc | 01:24 | |
| *** emagana has quit IRC | 01:29 | |
| *** emagana has joined #openstack-tc | 01:40 | |
| *** emagana has quit IRC | 01:44 | |
| *** emagana has joined #openstack-tc | 02:12 | |
| *** emagana has quit IRC | 02:16 | |
| *** emagana has joined #openstack-tc | 02:27 | |
| *** emagana has quit IRC | 02:32 | |
| *** emagana has joined #openstack-tc | 02:43 | |
| *** emagana has quit IRC | 02:48 | |
| *** emagana has joined #openstack-tc | 02:59 | |
| *** emagana has quit IRC | 03:03 | |
| *** emagana has joined #openstack-tc | 03:14 | |
| *** emagana has quit IRC | 03:19 | |
| *** emagana has joined #openstack-tc | 03:30 | |
| *** emagana has quit IRC | 03:35 | |
| *** emagana has joined #openstack-tc | 03:46 | |
| *** emagana has quit IRC | 03:50 | |
| *** emagana has joined #openstack-tc | 04:18 | |
| *** emagana has quit IRC | 04:22 | |
| *** emagana has joined #openstack-tc | 04:33 | |
| *** emagana has quit IRC | 04:38 | |
| *** emagana has joined #openstack-tc | 04:49 | |
| *** emagana has quit IRC | 04:54 | |
| *** emagana has joined #openstack-tc | 05:05 | |
| *** emagana has quit IRC | 05:10 | |
| *** emagana has joined #openstack-tc | 05:21 | |
| *** emagana has quit IRC | 05:25 | |
| *** emagana has joined #openstack-tc | 05:36 | |
| *** emagana has quit IRC | 05:41 | |
| *** emagana has joined #openstack-tc | 05:52 | |
| *** emagana has quit IRC | 05:57 | |
| *** emagana has joined #openstack-tc | 06:08 | |
| *** emagana has quit IRC | 06:13 | |
| *** emagana has joined #openstack-tc | 06:24 | |
| *** emagana has quit IRC | 06:28 | |
| *** emagana has joined #openstack-tc | 06:40 | |
| *** emagana has quit IRC | 06:44 | |
| *** emagana has joined #openstack-tc | 06:56 | |
| *** emagana has quit IRC | 07:01 | |
| *** emagana has joined #openstack-tc | 07:12 | |
| *** emagana has quit IRC | 07:17 | |
| *** jpich has joined #openstack-tc | 07:20 | |
| ttx | johnthetubaguy: Colette posted an email on the SWG, you might want to follow-up with your reviving plans. If you're not around I can post a quick something about it | 07:22 |
|---|---|---|
| *** emagana has joined #openstack-tc | 07:27 | |
| *** emagana has quit IRC | 07:32 | |
| *** emagana has joined #openstack-tc | 07:43 | |
| *** emagana has quit IRC | 07:48 | |
| *** emagana has joined #openstack-tc | 07:59 | |
| *** emagana has quit IRC | 08:04 | |
| *** emagana has joined #openstack-tc | 08:15 | |
| *** emagana has quit IRC | 08:19 | |
| *** emagana has joined #openstack-tc | 08:31 | |
| *** emagana has quit IRC | 08:35 | |
| *** emagana has joined #openstack-tc | 08:46 | |
| *** emagana has quit IRC | 08:51 | |
| *** emagana has joined #openstack-tc | 09:02 | |
| *** emagana has quit IRC | 09:07 | |
| *** emagana has joined #openstack-tc | 09:18 | |
| *** emagana has quit IRC | 09:23 | |
| *** emagana has joined #openstack-tc | 09:34 | |
| *** emagana has quit IRC | 09:39 | |
| *** emagana has joined #openstack-tc | 09:49 | |
| *** emagana has quit IRC | 09:54 | |
| *** emagana has joined #openstack-tc | 10:05 | |
| *** emagana has quit IRC | 10:10 | |
| *** emagana has joined #openstack-tc | 10:21 | |
| *** emagana has quit IRC | 10:25 | |
| *** emagana has joined #openstack-tc | 10:37 | |
| *** emagana has quit IRC | 10:41 | |
| *** emagana has joined #openstack-tc | 10:53 | |
| *** emagana has quit IRC | 10:57 | |
| *** emagana has joined #openstack-tc | 11:08 | |
| *** emagana has quit IRC | 11:13 | |
| *** emagana has joined #openstack-tc | 11:24 | |
| *** emagana has quit IRC | 11:29 | |
| *** emagana has joined #openstack-tc | 11:40 | |
| *** emagana has quit IRC | 11:44 | |
| *** emagana has joined #openstack-tc | 11:56 | |
| *** emagana has quit IRC | 12:01 | |
| *** emagana has joined #openstack-tc | 12:12 | |
| *** emagana has quit IRC | 12:16 | |
| *** emagana has joined #openstack-tc | 12:27 | |
| *** emagana has quit IRC | 12:31 | |
| *** emagana has joined #openstack-tc | 12:43 | |
| *** emagana has quit IRC | 12:48 | |
| *** emagana has joined #openstack-tc | 12:59 | |
| *** emagana has quit IRC | 13:04 | |
| *** emagana has joined #openstack-tc | 13:15 | |
| *** emagana has quit IRC | 13:20 | |
| *** emagana has joined #openstack-tc | 13:30 | |
| *** emagana has quit IRC | 13:35 | |
| *** emagana has joined #openstack-tc | 14:02 | |
| *** emagana has quit IRC | 14:07 | |
| *** emagana has joined #openstack-tc | 14:18 | |
| *** hongbin has joined #openstack-tc | 14:18 | |
| *** hongbin has quit IRC | 14:22 | |
| *** emagana has quit IRC | 14:22 | |
| *** hongbin has joined #openstack-tc | 14:22 | |
| *** emagana has joined #openstack-tc | 14:34 | |
| *** emagana has quit IRC | 14:38 | |
| *** lbragstad has quit IRC | 14:40 | |
| *** emagana has joined #openstack-tc | 14:50 | |
| *** emagana has quit IRC | 14:54 | |
| fungi | looks like it's that (tc office hours) time again folks! | 15:00 |
| dtroyer | heyo! | 15:01 |
| fungi | the last one (01:00 wednesday) was a bit of a bust. ended up talking to myself again | 15:01 |
| dtroyer | I totally missed it…and fingured out why my calendar sync wasn't working, so there is that | 15:02 |
| ttx | oooh, office hours | 15:02 |
| fungi | and looked like the 09:00 tuesday this week wasn't a whole lot better | 15:02 |
| ttx | happy to discuss the "how to deal with confusion" thread directly here | 15:02 |
| ttx | mordred: that could save us a few return trips | 15:03 |
| fungi | i'll be honest, i'm still catching up on mailing list posts this morning | 15:03 |
| fungi | though i think i'm caught up on that thread now | 15:03 |
| *** emagana has joined #openstack-tc | 15:05 | |
| ttx | fungi: thanks for reiterating why maintaining two orgs is painful | 15:08 |
| fungi | i probably missed some points, and it's not free from editorializing on my part, but hopefully it's close enough to an faq answer that people won't keep asking the question over and over | 15:09 |
| ttx | if it wasn't painful we just wouldn't be having this discussion at all :) | 15:10 |
| *** emagana has quit IRC | 15:10 | |
| *** zaneb has quit IRC | 15:11 | |
| fungi | right, we're at the point where any "easy" solutions have already been tried years ago | 15:12 |
| *** lbragstad has joined #openstack-tc | 15:14 | |
| fungi | though sdague's done some awesome work in the past ~12 hours starting to curate the openstack org page on gh | 15:14 |
| ttx | That's great, even if I don't think anyone looks at that page to find openstack projects, given the sheer size of it. At least those who do will be pointed to "good" repos | 15:16 |
| dhellmann | o/ | 15:17 |
| ttx | The issue I'm trying to fix is the one I mentioned on that email: | 15:17 |
| ttx | "It's about someone googling for | 15:17 |
| ttx | machine learning on openstack, finding Cognitive on git.openstack.org | 15:17 |
| ttx | and wiki.openstack.org, assuming it's an official openstack project | 15:18 |
| ttx | based on those domain names, trying to check it out, realizing it's only | 15:18 |
| ttx | 4 commits and dead for two years, and assuming OpenStack has pretty low | 15:18 |
| ttx | standards and is a bunch of dead crap. | 15:18 |
| ttx | " | 15:18 |
| dtroyer | I think we do need to had a very thin layer to the non-official project space that lets us clean out repos/projects that clearly are not making any progress and are not 'done' (as in completed work) | 15:18 |
| ttx | dtroyer: it's actually more difficult than you'd think. There is a reason why the stackforge repos are still around | 15:19 |
| ttx | dtroyer: what would be the rule ? No commit for 2 years, you're out ? | 15:20 |
| ttx | Or would you go and curate by hand those hundreds of projects ? | 15:20 |
| dtroyer | maybe it can be that sinple, for some period of time | 15:20 |
| ttx | Been startinog doing that on https://etherpad.openstack.org/p/negative-space-analysis and it's not exactly nice work | 15:20 |
| *** emagana has joined #openstack-tc | 15:22 | |
| dtroyer | of course when you set a minimum bar (ours is currently at 'it exists') there will be those that will meet that bar then stop | 15:22 |
| fungi | well, at least when we did the mass namespace move, any then-dead projects in stackforge got retired so their master branches now just contain a readme like http://git.openstack.org/cgit/stackforge/healthnmon/tree/README.rst | 15:23 |
| dtroyer | ttx: so much of the work we have is not nice work, it is part of the cost of maintaining a big system, especially one that wants to be many things to many people. The number of combinations of stuff we try to think about is absurd | 15:23 |
| fungi | i suppose we could do a mass ping of the maintainers of presumed dead repos and get either explicit or tacit approval to "retire" them | 15:23 |
| fungi | doing that is not free though, needs someone with bandwidth to periodically do that | 15:24 |
| dtroyer | it would simplify the task for that subset that will answer 'go ahead', which I suspect is small but non-zero | 15:24 |
| ttx | Just approved Fuel demotion, fwiw | 15:25 |
| fungi | dtroyer: which is why i also mentioned _tacit_ approval | 15:25 |
| dtroyer | so are we solving for less confusion for those outside openstack or for less work for those inside? | 15:25 |
| ttx | I'd like to solve for less confusion for those outside openstack while minimizing the work inside | 15:26 |
| ttx | which is why I put a few new options on the table | 15:26 |
| *** emagana has quit IRC | 15:26 | |
| dtroyer | I've had the stance with OSC that developers need to go the extra mile in order to provide the best user experience, that may have cost us some progress but I felt was worth it in that case. I don't get that feeling here | 15:27 |
| fungi | right, it was a bit of a blind spot in the original set of options where we assumed we couldn't solve for reducing outside confustion _and inside labor cost by sacrificing something else | 15:27 |
| fungi | s/confustion/confusion/ | 15:27 |
| fungi | i also don't think that kicking unofficial repos out of our infrastructure is a great solution, but we're basically weighing a lot of terrible choices at this point in an attempt to figure out which might be the least terrible | 15:29 |
| ttx | At the very least, if the consensus is that we should continue to support extra projects, acknowledge the fact that it's more costly (in terms of confusion and active work to limit it) than we first thought | 15:29 |
| ttx | The trick being, that never gave us extra resources | 15:30 |
| fungi | it's still possible that the least terrible solution is to let the current external confusion continue because the costs of making a significant dent in that are higher than the alternatives, but we still need to give them all consideration at least | 15:30 |
| dtroyer | I do think we should continue to host non-official projects (optionally adding that they can make even the slightest connection to the openstack mission). The cost of maintaining a clean ecosystem goes up marginally then, we're falling short in that regard either way. | 15:31 |
| dtroyer | I've been doing a bunch of random openstack + <common term> searches, the top results are almost always into the *.openstack.org domain space | 15:31 |
| ttx | The main reason for the confusion is that the solution is currently decided by people who do not suffer from it | 15:33 |
| * mordred waves to ttx | 15:34 | |
| ttx | between spending a lot of effort on policing it/implementing workarounds, dropping our free hosting and maintaining confusion, we choose to maintain confusion | 15:34 |
| * mordred reads scrollback | 15:34 | |
| dtroyer | but are the people who must implement the solution, as poopsed to those who are confused being able to decide on a solution that they bear no cost of implementation | 15:34 |
| ttx | But as I spend more time talking to random people outside of our bubble, I can see that people just don't get it | 15:34 |
| mordred | I do not believe there is a technical solution to this | 15:34 |
| mordred | this problem pre-dates the big tent | 15:34 |
| mordred | it predates the integraed release | 15:34 |
| ttx | oh, it has nothing to do with the big tent | 15:35 |
| mordred | it's the problem we have never solved | 15:35 |
| dtroyer | ttx: I'm reading 'it' as what OpenStack even is - that is is trying to be too many things? | 15:35 |
| mordred | yah - so kicking out a few deadbeat project is going to have zero impact - I don't even care about the cost of doing it - it's going ot have like literally no impact on the actual problem | 15:35 |
| mordred | which is that people want us to tell them what openstack is | 15:35 |
| ttx | it's just that since the "big tent" had this connotation of "anything goes in" it just reinforced people to believe that if they find crap with openstack marked on it, it's probably the big tent fault | 15:35 |
| ttx | mordred: I think we are not talking about the same problem | 15:36 |
| mordred | and, as a body, we have categorically decided that we're not in the business of passing value judgements on projects - no amount of git repo reorganzation is going to fix that | 15:36 |
| ttx | Let me read your answer first though | 15:36 |
| fungi | i still want to find ways of turning that perception around from "anything goes in" to "anybody goes in" | 15:36 |
| fungi | we're not governing software, we're governing people | 15:37 |
| ttx | mordred: not true -- as a body, we said that we are uin the business of judging if "you're one of us" and if you are helping with the mission | 15:37 |
| mordred | yes | 15:37 |
| ttx | The issue is that there are things that do not fit, and that people can mistake for things that fit | 15:37 |
| mordred | http://paste.openstack.org/show/613988/ <-- this is the set of repos that are unofficial btw | 15:37 |
| *** emagana has joined #openstack-tc | 15:37 | |
| ttx | aka https://etherpad.openstack.org/p/negative-space-analysis | 15:37 |
| fungi | defining and describing the software is orthogonal to the people and workflows used to produce it, and i think that human governance is the wrong place to be answering existential questions about computer programs | 15:38 |
| ttx | mordred: you're missing a few. Stackforge repos are unofficial too | 15:38 |
| ttx | so are openstack-attic ones | 15:38 |
| mordred | ttx: right - I filtered attic and stackforge | 15:38 |
| mordred | since the discussion was about confusion in the openstack/ namespace | 15:38 |
| fungi | everything in stackforge and openstack-attic namespaces is retured though. no contents except a readme saying explicitly it's dead | 15:38 |
| ttx | The etherpad is an effort at categorizing it | 15:38 |
| fungi | s/retured/retired/ | 15:39 |
| mordred | I just replied to jimmy suggesting that we add a retired flag to projects.yaml and have cgit not include those projects in the web ui | 15:39 |
| ttx | fungi: you can still find them from the description unfortunately | 15:39 |
| mordred | and similarly delete the remote github mirrors of those projects in jeepyb | 15:39 |
| fungi | ttx: maybe we need to reset the descriptions on them too in that case? | 15:39 |
| mordred | the code will still be in the git farm and in gerrit - but it would be less findable - and both should be trivial to implement | 15:40 |
| ttx | mordred: gerrit's projects.yaml ? | 15:40 |
| mordred | yes | 15:40 |
| ttx | mordred: I guess then the question is, who has authority to set it | 15:40 |
| fungi | there's a pretty major tradeoff to consider between preserving history and steering people away from dead ends | 15:40 |
| fungi | but flattening descriptions for retired repos seems like an okay compromise there to me | 15:41 |
| fungi | and relatively easy to do | 15:41 |
| ttx | fungi: yeah, that's yet another workaround we could implement | 15:41 |
| *** emagana has quit IRC | 15:42 | |
| mordred | fungi: do you have any problems removing mirror publication of retired repos? that seems fairly easy to me as well and could maybe help with google searches | 15:42 |
| dhellmann | ttx: let's just get the TC to set a time-based rule. We have the authority. | 15:42 |
| * mordred has no problem cleaning up abandoned projects | 15:42 | |
| mordred | dhellmann: ++ | 15:42 |
| fungi | mordred: that's a hard one to engineer, i thought. gerrit will queue up replication trying to put them back unless we switch to explicit whitelisting in the replication config | 15:43 |
| dhellmann | if a project goes dormant for a year and we retire it, then they come back, they can restore the repo | 15:43 |
| mordred | fungi: I'm just talking about the cgit config | 15:44 |
| mordred | fungi: the git repo would still be on the git farm | 15:44 |
| fungi | mordred: but still leave them on github? | 15:44 |
| dhellmann | is there no flag to control whether a repo is mirrored or not? | 15:44 |
| mordred | fungi: won't gerrit only try to put it back on github when a new commit is landed? | 15:44 |
| mordred | fungi: if we set it as deleted after we land the retired commit, then the removal of the github mirror should have no impact - or, we can probably test that assertion pretty easily | 15:45 |
| fungi | i mean, there's a lot we _can_ do on the cgit side of things to stop publishing retired repos there, or stop listing them in the index, or even replace the index with something based on classifications and tag groupings built out of governance metadata | 15:45 |
| ttx | mordred: leaving aside the "dead" for a moment. What about a not-so-hypothetic open core project. People find openstack/opencore by googling something, assumes it's official, and therefore thinks open core is OK "in" openstack | 15:45 |
| fungi | mordred: i believe gerrit also tries to refresh remote repos any time it is restarted or a blanked replicate api command it issued, but it's worth a small-scale test to confirm, sure | 15:46 |
| fungi | s/blanked/blanket/ | 15:46 |
| mordred | ttx: can you make that less hypothetical if it's not hypothetical? | 15:46 |
| ttx | We have made the choice of not curating based on liveness, but we made the choice of curating based on principles | 15:46 |
| fungi | dhellmann: the configuration for whether to mirror a repo isn't kept with the repo configuration, you configure gerrit to replicate and can optionally tell it not to replicate everything by providing a list of repos or regular expressions | 15:47 |
| mordred | ttx: and I wouldalso have no problem in removing things from our unofficial host that violate the 4 open | 15:47 |
| mordred | gah | 15:47 |
| dhellmann | fungi : ah | 15:47 |
| mordred | or being more clear that it's not free hsting for ANYTHING | 15:47 |
| dhellmann | that would be a big regex | 15:47 |
| fungi | yup | 15:47 |
| ttx | mordred: then why have unofficial space, at that point ? Why not make everything official ? | 15:47 |
| fungi | 1049 repos we need to whitelist at present unless we decide not all official projects get replicate to gh | 15:47 |
| ttx | if you put governance on the unofficial space anyway | 15:47 |
| ttx | Currently there is ungoverned space and governed space | 15:48 |
| ttx | under the same branding | 15:48 |
| ttx | which dilutes the value of governance | 15:48 |
| ttx | and makes people just say "why suffer the TC" ? | 15:48 |
| ttx | (see the questions on the original thread) | 15:48 |
| mordred | ttx: I think there is a diffrence between hosting ungoverned and hosting closed-source | 15:49 |
| ttx | open core is not closed source :) | 15:49 |
| mordred | yes it is | 15:49 |
| ttx | it's just partial hosting | 15:49 |
| mordred | I have no problem making the distinction | 15:49 |
| ttx | What about level-playing-field then ? OpenStack is about open collaboration, we have plenty of projects in unofficial space taht give unfair advantage to one vendor. OK, not ok ? | 15:50 |
| dhellmann | do we really have "plenty"? | 15:50 |
| ttx | In the end, it boils down to this: | 15:51 |
| ttx | We said "openstack is this community with this set of principles" | 15:51 |
| mordred | yah. we don't have THAT many unofficial projects | 15:51 |
| mordred | that's my main point | 15:51 |
| mordred | the actual number of 'problem' chiildren is tiny | 15:51 |
| mordred | and it feels like we're trying to solve what is actually a very small problem by completely reworking how we structure ourselves | 15:51 |
| ttx | and then we allow things that do not follow those principles (and therefore are not official) to appear to be openstack | 15:51 |
| mordred | the bigger problem that is bother folks is that there are 1000 REAL projects in openstack/ | 15:51 |
| ttx | There are less unofficial repos than there are official repos, for sure, but it's the same order of magnitude | 15:52 |
| mordred | those of us who understand what's going on are not bothered. the only people who are bothered are people who do not understand. those people are going to be just as bothered when they learn that the official set of projects in openstack/ is over 1000 | 15:52 |
| ttx | Also there are more unofficial PROJECTS than there are official ones | 15:52 |
| ttx | i.e. teams | 15:52 |
| ttx | so it's significant | 15:53 |
| mordred | but that doesnt' bother anyone | 15:53 |
| mordred | the only people who cares about that part is us | 15:53 |
| mordred | if you get far enoughto KNOW about project teams, you're engaging with the governance structure | 15:53 |
| ttx | mordred: I don't think anyone is bothered by the fact that we have 1000 git repos | 15:53 |
| ttx | that's not where the confusion comes from | 15:53 |
| mordred | I categorically disagree | 15:54 |
| dtroyer | when we talk about the confusion in making sense of the openstack/ repo namespace, we're talking about repos not teams | 15:54 |
| mordred | yah. ANY discussion about things being in openstack/ is based on people finding the number of thigns there confusing | 15:54 |
| ttx | The confusion is that people see a talk about Blazar at a summit and can't find it in the openstack releases page | 15:54 |
| *** emagana has joined #openstack-tc | 15:54 | |
| mordred | that will not be solved by reorganizging git repos | 15:54 |
| dtroyer | that particular one is a summit talk acceptance problem? | 15:55 |
| dhellmann | ttx: why are unofficial projects presenting at the summit? | 15:55 |
| ttx | while everything they can find about Blazar on the Internet screams it's an official project | 15:55 |
| dtroyer | there is a HUGE carrot in becoming official, only offical proejcts are on the program | 15:55 |
| ttx | openstack/blazar on GitHub or cgit, Blazar wiki page on our wiki etc | 15:55 |
| fungi | the confusion i keep hearing described is that people who aren't sure where to figure out what openstack is or does are getting bad information from unreliable sources and aren't given the means to ascertain the reliability of the information they're presented with | 15:55 |
| ttx | dhellmann: because we are inclusive of our ecosystem | 15:55 |
| mordred | but why was it on a sumit schedule and using the word openstack | 15:55 |
| dtroyer | so if namespacing is going to help, let's start by namespacing stuff in the wiki, that'll make it clear, right? | 15:56 |
| ttx | mordred: why is it on our cgit using the word OpenStack ? | 15:56 |
| fungi | similarly, everything on dockerhub is officially part of docker | 15:56 |
| ttx | The only places it doesn't appear are governance.o.o and releases.o.o | 15:56 |
| mordred | because there is an explicit trademark carve out for that | 15:56 |
| mordred | fungi: ++ | 15:56 |
| mordred | ttx: we have an explicit carve-out around git repos and development - precisely because we need to be able to work on stuff before it's ready to be official | 15:57 |
| ttx | mordred: I would argue the community usage of the trademark doesn't extend to unofficial projects | 15:57 |
| mordred | well - then let's go to fungi's suggestionand get rid of the git namespace altogether | 15:57 |
| mordred | it holds no meaning for us | 15:57 |
| ttx | ack, that would help | 15:58 |
| mordred | and let's just stop mirroring to github, which also holds no meaning - and is literally the only reason we have the prefix | 15:58 |
| dtroyer | really? | 15:58 |
| fungi | getting rid of the openstack.org domain is probably a more effective solution. similar to how the new pypi is pypi.org not pypi.python.org, so that people stop assuming everything on pypi is an official part of python | 15:58 |
| ttx | That is exactly what I mean by addressing the issue from the outside | 15:58 |
| mordred | what we CAN'T do is mirror some things to github in opentsack/ and some things in not-opentsack/ | 15:58 |
| *** emagana has quit IRC | 15:58 | |
| ttx | mordred: it won't remove the "openstack" from http://git.openstack.org/cgit/adjutant/ | 15:59 |
| mordred | ttx: right. so we make an "opengate.org" and do our dev there | 15:59 |
| mordred | remove all openstack mentions from any git repo name | 16:00 |
| mordred | and have the only place a thing gets to add the word openstack into its git repo name is via governance | 16:00 |
| fungi | we actually _can_ selectively mirror to gh, for increasingly interesting definitions of "we" if it can be determined that gh social media isn't something the infra team needs to take care of. it's possible for someone in the community with the blessing of the foundation to run a read-only pull+push mirror from our official upstream repos with whatever filtering they want to impose, and ideally also take | 16:00 |
| fungi | over the pull request closer at the same time | 16:00 |
| ttx | we'd have to also do wiki.opengate.org, but yeah, that would address it | 16:01 |
| mordred | fungi: that's a good point | 16:01 |
| mordred | fungi: oh - you know what - that's a GREAT point | 16:01 |
| * mordred hands fungi a cookie | 16:01 | |
| fungi | treat it the same way we treat the openstack twitter and openstack org on ohloh | 16:01 |
| mordred | ttx: I would like to write up a suggestion based on the above - that will not result in github mirrors going away | 16:02 |
| fungi | i honestly don't see why the upstream infra team even needs to dirty its hands with gh integration at all | 16:02 |
| mordred | but that I think will make more sense in a longer-form than IRC - AND I think is potentially actionable | 16:02 |
| ttx | I fear that opengate.org would create a confusion of its own, though | 16:02 |
| fungi | it's moby time ;) | 16:02 |
| dtroyer | it exists? | 16:02 |
| mordred | well - let's talk about that too - but let's put together all the pieces in one place | 16:02 |
| ttx | Moby wazsn't exactly a smooth ride for Docker | 16:03 |
| mordred | so we can read them - and then we can find what's wrong with it all | 16:03 |
| ttx | but yes the options actually are... | 16:04 |
| mordred | fungi: basically - if we stop having gerrit replicate to github at all and move to a push/pull bot someone runs - that bot can TOTALLY read goverannce/projects.yaml for the list of things it should mirror into the github openstack/ namespace | 16:04 |
| jeblair | let's not moby ourselves if we can avoid it. | 16:04 |
| ttx | 1/ Keep governed and ungoverned in the same space, and try to damagecontrol on the edges | 16:04 |
| mordred | that can be step one and can be done regardless of any of this - and gets infra out of the github game | 16:04 |
| ttx | 2/ Move ungoverned out | 16:04 |
| ttx | 3/ Move everything to something else, and then apply the trademark more selectively on governed stuff only | 16:05 |
| ttx | (3) sounds like it's more disruptive and confusing than the other two, but it's also on the table | 16:05 |
| mordred | yah. I like 3 because it allows us to be inclusive as a community - but restrictive for messaging/branding/trademark purposes | 16:05 |
| mordred | because inclusive development is very important | 16:05 |
| mordred | but so is keeping branding clear | 16:06 |
| ttx | totally | 16:06 |
| mordred | ttx: I will write up a detailed thing on 3 so we can all react to it | 16:06 |
| ttx | mordred: that is actually what I was hinting towards when i said 'we should brand the infra team' and proposed Opium in the first thread | 16:07 |
| ttx | If our infrastructure had a name, official and unofficial could share that same space | 16:07 |
| mordred | exactly | 16:07 |
| mordred | and then publication of things (git, artifacts, docs, whatever) can be segregated/namespaced | 16:08 |
| ttx | Still means a bit of enforcement on edges, like policing wiki.o.o or openstack-dev | 16:08 |
| mordred | we've had the inverse issue with docs for forever - we don't let projects publish docs to openstack.org because it's docs.openstack.org | 16:08 |
| mordred | ttx: we could always kill the wiki ;) | 16:08 |
| ttx | oooo great idea! wait | 16:08 |
| ttx | mordred: one issue with the Mobyification option is taht it's likely to be a lot harder to get infra resources from openstack sponsors if it's seen to support a lot more | 16:10 |
| *** emagana has joined #openstack-tc | 16:10 | |
| jeblair | we've always killed the wiki. we will always kill the wiki. | 16:10 |
| ttx | (key word being "seen") | 16:10 |
| fungi | killing the wiki would knock a huuuuge entry off my to do list that i keep not getting around to picking back up | 16:10 |
| mordred | ttx: indeed. I think we need to be clear/careful on that topic | 16:10 |
| ttx | mordred: or the name of the software forge could be close enough so that it's still *seen* as mainly supporting openstack development | 16:11 |
| fungi | opendev! ;) | 16:11 |
| * ttx slaps fungi | 16:11 | |
| mugsie | is the docs team not moving the Ops guide to the wiki ? | 16:11 |
| mordred | god are they? | 16:12 |
| mugsie | I am pretty sure they are | 16:12 |
| * ttx slaps himself | 16:12 | |
| * ttx sobs | 16:12 | |
| fungi | i heard the _ops_ team was going to try moving the ops guide to the wiki | 16:12 |
| fungi | the _docs_ team doesn't have the bandwidth for that | 16:12 |
| fungi | so they left it up to the ops team to do whatever they feel is best with it | 16:13 |
| mugsie | Oh, OK - i saw an email from alex saying they were moving, so i asusmed it was the docs team | 16:13 |
| ttx | I still hope I could find a wiki that actually enforces page ownership so that you just couldn't create things outside of owned hierarchies | 16:13 |
| ttx | and migrate us to that | 16:13 |
| mordred | ttx: it's closed source | 16:13 |
| mugsie | confluence does actually do that ... but ... | 16:13 |
| mordred | ttx: the wiki that does that | 16:13 |
| ttx | I don't mind the docs team using a wiki for publishing stuff. I mind random pages being created at top level | 16:14 |
| ttx | and then be nobody's job to update | 16:14 |
| mordred | I do not understand using a wiki when git exists - but clearly I'm not the relevant audience here | 16:14 |
| *** emagana has quit IRC | 16:14 | |
| ttx | mordred: yes you are not | 16:14 |
| fungi | we still use a wiki for the infra team weekly meeting agenda | 16:14 |
| mordred | ttx: we could just move the entire ops guide to the openstack wikipedia article ;) | 16:15 |
| fungi | but i think that's my only real interaction with wiki.o.o these days other than helping maintain it | 16:15 |
| ttx | I'm maintaining the TC status page on it | 16:16 |
| ttx | that is all I do on the wiki | 16:16 |
| *** emagana has joined #openstack-tc | 16:26 | |
| *** jpich has quit IRC | 16:28 | |
| *** emagana has quit IRC | 16:31 | |
| sdague | late to the party, but a couple of things that I've been thinking about in the conversation thus far | 16:32 |
| sdague | 1) this is what an ecosystem looks like. Some beautiful tomato and pepper plants, a few flowers, a lot of dirt, worms, weeds, insects, and cow poop to keep it going | 16:33 |
| sdague | you just can't have an ecosystem that's only beautiful things, that's called astro turf | 16:34 |
| mugsie | looks good from a far, but feels wrong, and hurts when you fall? - yeah, thats a pretty good analogy | 16:34 |
| sdague | 2) confusion comes when depending on where you get information you get different or oddly incomplete answers. And the first answer people get will stick in their head and shape/pollute the rest of their understanding. | 16:35 |
| sdague | which is one of the reasons I think cleaning up descriptions and putting some organization into github is useful | 16:35 |
| sdague | 3) when you communicate in a forum, you need to use the conventions of that forum to get maximal buy in. The conventions themselves prevent you from having to explain certain things. | 16:36 |
| *** zaneb has joined #openstack-tc | 16:37 | |
| sdague | which relates to the github info | 16:37 |
| sdague | but also any other place we engaged | 16:38 |
| sdague | engage | 16:38 |
| sdague | because when you don't, you come across as "I'm too cool to follow your conventions". Which puts a social barrier to interacting with other communities. | 16:39 |
| sdague | anyway, just some thoughts. I don't pretend that they build a complete picture. | 16:39 |
| *** emagana has joined #openstack-tc | 16:42 | |
| sdague | 4) oh, and, a lot of these problems are information development ones. A good wiki needs librarians. Good coherent documentation and story telling needs information developers/tech writers. | 16:43 |
| *** emagana has quit IRC | 16:43 | |
| *** emagana has joined #openstack-tc | 16:43 | |
| sdague | the bulk of those folks we had in the community got laid off recently, and we've got a huge void there | 16:43 |
| mordred | sdague: ++ to all of the above | 16:55 |
| *** emagana has quit IRC | 17:31 | |
| *** emagana has joined #openstack-tc | 17:31 | |
| *** emagana has quit IRC | 17:36 | |
| fungi | the https://review.openstack.org/479058 infra spec i just proposed may be of general tc interest as it impacts the technical election process | 20:43 |
| fungi | rendered version is http://docs-draft.openstack.org/58/479058/1/check/gate-infra-specs-docs-ubuntu-xenial/f059983//doc/build/html/specs/gerrit-contactstore-removal.html | 20:43 |
| *** zaneb has quit IRC | 21:56 | |
| *** johnthetubaguy has quit IRC | 23:18 | |
| *** hongbin has quit IRC | 23:19 | |
| *** johnthetubaguy has joined #openstack-tc | 23:28 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!