there's a middle ground between "Linux as a personality" and " it doesn't matter at all who controls the future of computing". be careful not to react against one so hard that you slide right into the other.

@tindall my main beef is that both of these extremes ignore the orthogonal issue where operating systems and source code aren't really important to fixing the power dynamics that make today's computing bad

@scanlime @tindall dbus is still the most pro-user faculty i would point to, the core difference.

everywhere else, software is monolithic, has no extensibility. in linux, it's not great, but there is some notion that software is just a tool, and even without recompiling the source, users are sovereign & ought have power & control over the software. dbus embodies that, by making the software programmable & adaptable & flexible & extensible, even without recompiling, even without the source.

it's still dev-oriented, but at least it allows extensibility. your music app exposes mpris interfaces, so you can add another scrobbling app, or a gamepad control app, regardless of your music player. there's a general system for building together, unlike everywhere else on the planet.

@jauntywunderkind420 @tindall hm dbus would not be my first choice. how about good old unix pipes?

@scanlime @tindall pipes are a good example, but they don't aggregate. pipes are deep spellcraft, are a developer-style wisening that empowers the same same "source is available" folk but not the every-man. pipes don't accrue, pipe's dont build up, except as we, case by case, hack them. dbus, on the other hand, accrues, builds, is an ambient growing system of systems that continually extends the userland. dbus is richening. pipes are shallow, ad-hoc, specific, instantaneous. they have none of the enduring might of dbus. dbus extends the user-lands.

@jauntywunderkind420 @tindall i don't get how dbus is a user-friendly way to extend anything

@scanlime @tindall as per my first example, it allows runtime extensibility. your media players all export mpris controls, so you the user can just have some audioscrobbling, some joystick control, some hotkey daemons running, watching mpris, controlling as they may. it's decoupling grease for programs.

@jauntywunderkind420 @tindall in that example why does dbus help? what the user needs is a set of apps that work together. I don't see anything that's intrinsically allowing more user flexibility than what developers specifically plan for and test.

@scanlime dbus has been the only accepted, commonly used bus for interoperability, extension, & service activation in free software with significant adoption for around two decades now.

yes, sure, fine, there's no reason other rendezvous methods wouldn't work. but the number of pid1's that support dbus & dbus service activation is infinitely greater than the number of pid1's that support any other attempt or protocol.

@scanlime mostly though i just think dbus embodies the right spirit of freedom. it embodies the act of making software ambient, available, is a fine way for software to exhibit itself, make itself accessible. i do not see these characteristics in mainstream operating systems, and yes, i do not see this spirit of user-freedom particularly embodied by most other open source operating system environments. it's flawed, imperfect, not wholly sufficient, but freedesktop's adoption of dbus has been a very good & moral thing, has been the right direction for things. i very much would enjoy more competition on this front, but dbus has been a maximalizing system-of-systems for user-flexibility for over 20 years now.

@jauntywunderkind420 dbus and this thread are both just great examples of tech junk that is totally orthogonal to any actual problems or solutions

@scanlime ok well that's a jerkwad response to some very kindly reasoned & i think semi-substantiated argumentation.

i see no other places where open-source has marshalled a substantial front for user empowerment. dbus speaks for itself as supremely successful & competent.

i'm not afraid of looking at other options. i dont only vote by popularity, proven success. but there is almost no one else in the runnings. everyone else has been playing with themselves, for far too long. freedesktop has contributed real value. this thread is indeed insubstantial, but not on my side.

@scanlime i just want you to say something positive about something, to make some kind of case. rather than just trashing endlessly the cases i keep making. which i think have some value. if not perfect value.

@jauntywunderkind420 i'm not trying to be a jerk i'm just trying to be real about what things actually are. and wow this is some out of touch nerd junk. what do people use to automate their computers? gui tools like apple automator, or shell scripts, or special-purpose macro apps, or in-app scripting

@scanlime there are some very good examples of dbus as an agent for extensibility in action. from The Architectures of Open Source, one of my favorite examples is Telepathy, which uses dbus to create a decoupled instant messenger architecture. this allows infinite combinations of messaging platforms, allows for extensible, easy ways to write bots that work with arbitrary different messaging systems. run pidgin? get access to dozens of different messaging platforms? that's dbus. it works.
aosabook.org/en/telepathy.html

i very much want to see more automation. dbus was not easy to integrate with shell scripting, and i think, yes, that is a critical failure, a miserable fault. but decoupled software architectures, extensibility, is something i see again & again in dbus. telepathy messaging, mpris media playing: just some examples. gpsd location is another common example. software decoupling needs grease. dbus has been there. for decades. kicking ass, taking names.

@jauntywunderkind420 are we talking about features for app developers or features that can be used to link apps you didn't develop with each other? because pipes have done the latter since forever

@scanlime in terms of "operating systems and source code aren't really important to fixing the power dynamics that make today's computing bad", i dbus has been a radicalizing force for de-monolithizing power-bases & disaggregating power & capability in a radically important way.

yeah, tons of credit to pipes. but they have been expert tools, forever, requiring deep not-user-friendly comprehension to swizzle things together intelligently. and unlike dbus, there is no ambient constructivist system, no means for softwares to see each other & grow together autonomically efficiently.

@scanlime ok well this doesn't seem to be very fruitful a discussion. fine. i'm unsure what at all is unclear in any way about any thing i've argued. i'd love a little more help refining & developing my argumentation. but i just have no idea what my weaknesses are at this point.

@jauntywunderkind420 I don't have the energy to spew paragraphs about this like you do, mostly I'm trying to say the choice of tech parts is irrelevant and none of this matters for actual interoperability challenges folks face, which are mostly solved by spreadsheets and macros and shit like applescript

@scanlime as the original post said, it's a spectrum, and i'm trying to testify for why i think there's a positive, developing power base in linux beyond mastery of the source code. especially as compared to everything else that exists in the mainstream.

Sign in to participate in the conversation
Beach City

Beach City is our private beach-side sanctuary for close friends and awesome folks. We are various flavors of trans, queer, non-binary, polyamorous, disabled, furry, etc.