Hello Terry, Terry Cocksworth wrote on Tue, Sep 23, 2025 at 11:44:15AM +0000: > Crystal Kolipe <[email protected]>: >> On Mon, Sep 22, 2025 at 11:14:30PM +0000, Terry Cocksworth wrote:
>>> So what help would be needed? >> From what I can gather, a plane ticket for Stuart and a cordless >> vacuum cleaner to prevent the problem from re-occurring. >> https://marc.info/?l=openbsd-misc&m=173507343430607&w=2 > I appreciate the lightheartedness Indeed, in the cited posting, Stuart neither says that he is willing to spend the *time* to travel from England to Canada (even if someone paid for the ticket) nor that something that can be fixed with a vacuum cleaner is connected to the problem in any way. To the contrary, the thing with the "cleaner" was a mere joke by someone who is not even a project member. > but surely something this important > could not rest on the shoulders of one sole project member and their > agreeing to extended travel in order to unclog some server? Is there > no one else - project member or server host staff - I would be somewhat surprised if there would be anything akin to (paid) "host staff" for any OpenBSD-affiliated server. Personally, i have remote access to OpenBSD-affiliated servers in three different cities (all thousands of miles away from where i live and also more than a thousand miles away from each other), and according to my best knowledge, all those servers are maintained entirely by volunteers. Some are in public infrastructure datacenters (those datacenters of course are run by staff of those public institutions, but they don't fix the servers), some are in corporate datacenters (those datacenters of course are run by corporate staff, but they don't fix the servers), some are in private residences (with adequate, highly redundant and highly capable network access). But the servers themselves are all run entirely by volunteers for all i know. > who can perform this task? This comes off as trifle. > It's positively absurd. But I am asking in all seriousness > with no intent to come off as a critic. When it comes to free software, it is not at all unusual that lots and lots of things hinge on individuals, and that is more true for OpenBSD than for some other projects because OpenBSD is run by a relatively small group of people with an extremely low, almost non-existent administrative overhead. Almost everything that is done by OpenBSD directly results in code improvements. > After reading the other thread I now realize the project is coming up > on a bloody full year of inaction and Arm64 debacle. What prevents the > "management" The "management" consists of exactly one individual, Theo de Raadt, and the one thing our "management" is not entitled to doing (in stark contrast to the role of management in a commercial company) is ordering any project member to do any particular work. Theo already does insane amounts of work in all kinds of areas including software design, code implementation, code maintenance, code review all over the tree, insane amounts of continuous testing on literally all supported hardware platforms, extensive amounts of technical counseling of developers in almost all areas, server purchasing, setup, operation, and maintenance, hackathon planning and organization, release planning, execution and publishing, and this list is almost certainly still incomplete. Needless to say, when Theo approaches any developer pointing out any problem, which is quite a common occurence given the amount of testing and code review he habitually does, the typical reaction of that developer is to take care of the problem as fast as they can, but it does occasionally happen that simply nobody can afford spending the time required. When there is a critical bug, it will get fixed quickly: someone competent will make the time no matter the personal cost for them. While it does not seem entirely unreasonable that some may regard lack of arm64-stable packages as an unfortunate problem, it does not rise to the level of severity, like a serious vulnerability in the base system would, that developers put everything else on hold (including their lives and jobs outside OpenBSD) to fix it. In general, keeping -current in good shape is regarded as *vastly* more important by OpenBSD developers than -stable ports on any platform, even on amd64. > from stepping in and sorting this out? Can anyone help? > Do you need more economic donations to allocate a replacement set of > Arm64 servers? Another year of extra time? The donation goals suggest > that there's at least plenty of funds to pay for flights. Well, https://www.openbsdfoundation.org/campaign2025.html says that right now (Sep. 23), about $140k of the $400k of the 2025 fundraising goal have been raised so far, which means that it is possible that some projects have to be postponed due to shortage of funds - i don't know any of the details, though. Please understand that the Foundation does not typically accept donations conditioned on being spent for any particular purpose, both to keep administrative effort down (remember that there is no administrative staff and any administrative effort would directly reduce the time that could be spent on development), but also to prevent external forces from influencing OpenBSD project priorities. And yes, developers prioritising the desires of companies contributing funds to those projects over sane development principles and priorities is a real problem in some other free software projects. In many cases, limited free time available to developers is an additional constraint in addition to lack of financial resources. Sometimes, it is possible to free some developer time by paying them for OpenBSD work such that they have to do less other work for hire - but that's not always an option, and it's easy to see you won't get all that far with a budget of $400 a year that also needs to cover hackathon and travel costs, hardware, electricity, redundant high-performance network access, and whatnot. Even if we had unlimited money to hire additional staff, that's not always practical. In particular, for running stuff like build servers, trust is essential, so that can only be done by fully trusted developers, not by some random employee (who doesn't exist anyway). So the best people can do is: 1) Contibute patches, such that over time, we get more developers and hence more people directly helping out by volunteering their time and reducing the workload on existing developers. 2) Contibute donations, such that fundraising goals are reached and those things that *can* be done with money actually do get done. Thank you for your understanding, Ingo
