free media alliance

 free software, free culture, free hardware



[lit] the real problem with redix is that it breaks important things. the most important thing it does is hobble the gnu[lit]/[lit]linux ecosystem-- its developers, and users. gui redesign follows design trends and fashion-- discoverability is higher, documentation is a pain. shell scripts ought to have greater stability than guis. constantly eroding the posix interface is what redix is all about. as an api, gnu[lit]/[lit]linux does evolve. but its stability and longevity is one of its strengths, and one of the surest ways to arouse the ire and scepticism of developers as well as users, is to push for breaking it. there is no better or more historical example of this than systemd. if systemd had developed more along the lines that busybox did-- as a lightweight alternative to a standard, rather than an effort to take over the ecosystem, the fsf would still be doing its job and redix would not be a threat. redix watch is a line in the sand, but we should all know that line will move. if you write a better tool, there is still plenty of reason not to break everything else. if you write a better tool, eventually it will be gradually adopted. but if you write your tools in a way that is pushy and monopolistic, which causes years of problems for the ecosystem to work around-- youve advanced redix. the following consists of notes, not formal categorisation-- not everything here deserves the label of redix. but it should help to build a theme around the idea. systemd inconveniences: everyone in the gnu[lit]/[lit]linux ecosystem who doesnt want it, or that considers it a security[lit]/[lit]stability problem breaks: the entire debian ecosystem, among others benefits: large corporations with big resources, who want the ecosystem designed in a way that gives them more (most) control over design and maintenance allegedly fixes: the linux ecosystem, old designs, networking, init-- everything that can possibly be brought under an opaque, convoluted, monopolistic umbrella foisted by: red hat, a few debian developers, allegedly gnome (contested by groups who have cheerfully bragged that you will be forced to use it) pulseaudio inconveniences: people who plug their headphones in and remove them, to find the speaker volume they muted is now helpfully turned up to 100 in the middle of the night (yes, you can turn that off. who is responsible for making it a default?) breaks: audio benefits: musicians, presumably, who would do better using jack and avoiding this garbage like the plague allegedly fixes: the very thing it breaks foisted by: mozilla, who made it a near-requirement for audio on the web (apulse may make it a non-requirement) iproute2 (ss, ip) inconveniences: command line users who benefit from the relatively friendly, straightforward interface of net-tools breaks: countless shell scripts, production benefits: faster, more complete, more up-to-date interface (could be true, but systemd claims similar) allegedly fixes: net-tools (netstat, ifconfig) using [lit]/[lit]proc to get network info foisted by: not necessarily anyone. ifconfigs output has changed too, if your scripts parse it home: [lit][lit]