On Thu, 1 Aug 2024 at 09:49, Jonas Smedegaard <jo...@jones.dk> wrote: > Quoting Otto Kekäläinen (2024-08-01 07:30:18) > > You have for sure developed an optimal workflow for yourself. > > Then I have failed: I have strived towards a collaborative workflow. > > Just not a web-centered collaborative workflow, but an email- and chat- > based one.
Emails are actually a barrier against collaboration, and actively hinder it by preventing new people from joining in. Please understand that, outside of the demographic group of people who started using computers in the 70s and 80s, the vast majority of opinions on the topic of email as a collaborative tool range from "barely tolerated" to "deeply hated". Emails sucks, it's horrible and outdated stuff based on horrible and outdated procolos and ideas, and today it is by and large a system to deliver spam, phishing and malware, with occasional barely useful things like confirming registration on a new website or resetting a lost password. The vast majority of people who are forced to use emails do so for work via a work-mediated/administered interface that tries to make it somewhat tolerable, like Outlook or Gmail. By and large people born this side of the millennium look at people running their own email workflows as you would look at someone programming with punchcards - an interesting curiosity if found in a museum re-enactment or a history book photograph, and a strong reason to get the heck out of there otherwise. And outside of the tech bubble it's even worse. Some projects like the kernel can afford to let the curmudgeons continue to dictate its use because it's a de-facto monopoly and if one wants to get things merged in Linux there is no alternative, and even then there are continuous grumblings and attempts by the IT infra managers to build bridges with the 21st century with custom and bespoke kernel-specific kludges, to the point where some subtree maintainers have just gone "fuck this" and set up their own Gitlab repos for their own subtrees. We don't have that luxury: there are plenty of alternatives to Debian that work just as well for the same use cases. The number of people involved in programming, software engineering, IT and any other related field has absolutely exploded since the 90s. The number of Debian project members has remained pretty much flat at ~1000 people, and we just about manage to backfill retirements/MIAs. To pick a random example, a less well known, less used, less popular distribution like Nixos has 7000+ contributors listed on Github. It would be wise to stop and reflect on why this is the case every now and then. The passage of time and demographics are cruel and merciless. And the fact that, two decades or so after it became standard practice in the industry, there is _still_ resistance against having CI runs _before_ pushing things out simply boggles my mind. This is a no brainer: run a CI before uploading, even a very basic one is just fine, better than nothing. It's 2024, "but but but the Gitlab UI doesn't display nicely on my 80x24 green phosphor monochrome monitor" just doesn't cut it anymore, sorry.
