!bQveTkYXFyxLJQwcLG:matrix.org

Malleable Systems Collective

98 Members
https://malleable.systems | Contribute to the catalog and blog: https://github.com/malleable-systems/malleable.systems | Logs: https://view.matrix.org/room/!bQveTkYXFyxLJQwcLG:matrix.org/16 Servers

Load older messages


SenderMessageTime
9 Apr 2021
@akkartik:matrix.orgKartik Agaram

Yes. Compatibility is a crutch for "users", but it also keeps them from growing into more. People with agency on a computer should not be constrained by it.

That does create problems for collaboration. We should be trying to meet those problems and solve them, not shy away from them.

My current best-guess synthesis in response to this tension:

  • Upgrades always happen on the receiver's schedule. We need communication channels to convey urgency, but hands off my machine.
  • Allow people to try out incompatible upgrades and easily roll back.
17:14:01
@akkartik:matrix.orgKartik Agaram

Gnuxie πŸ’œπŸ (Computing Maximalist): I didn't intend the link to Hyrum's law as some sort of ironclad refutation, just indirect circumstantial evidence in some favor of my position. I think this question isn't settled yet, and I still reflect on all the different opinions. It's interesting to think about situations where each side is closer to the truth. Thanks for the SBCL/quicklisp scenario.

One pattern I've noticed is that there's a "settling time" when it's easier to violate compatibility in a compatibility-focused community. But once it's been out there for some critical period, and has been in use by some critical mass of people, there's push-back if things break. Even if the adoption is tiny compared to all the people who might use something in the future. That feels sub-optimal.

17:15:10
@akkartik:matrix.orgKartik Agaram *

Gnuxie πŸ’œπŸ (Computing Maximalist): I didn't intend the link to Hyrum's law as some sort of ironclad refutation, just indirect circumstantial evidence in some favor of my position. I think this question isn't settled yet, and I still reflect on all the different opinions. It's interesting to think about situations where each side is closer to the truth. Thanks for the SBCL/quicklisp scenario.

One pattern I've noticed is that there's a "settling time" when it's easier to violate compatibility in a compatibility-focused community. But once it's been out there for some critical period, and has been in use by some critical mass of people, there's push-back if things break. Even if the adoption is tiny compared to all the people who might use something in the future. That feels sub-optimal.

17:19:40
@gnu_ponut:matrix.orgGnuxie πŸ’œπŸ (Computing Maximalist)i don't believe it does provide validation to your position and it's surreal but totally expected to me that you interpret it this way17:33:39
@gnu_ponut:matrix.orgGnuxie πŸ’œπŸ (Computing Maximalist)like it's surreal because you obviously shouldn't rely on behavior that isn't specified, sure it happens but it should be mostly easy to fix, and it doesn't really matter if in some circumstances it's found the behavior is desirable and becomes a part of the interface 17:36:11
@gnu_ponut:matrix.orgGnuxie πŸ’œπŸ (Computing Maximalist)to argue that therefor all behavior should be explicit is to encourage people to rely on irrelevant things and also encourage the situations where you can no longer change them because of that is absurd and i'm disappointed that the collective hasn't been more critical tbh 17:37:41
@gnu_ponut:matrix.orgGnuxie πŸ’œπŸ (Computing Maximalist)i know these last few sentences are harsh but that's actually my initial reaction17:38:26
@jryans:matrix.orgJ. Ryan StinnettIn my view, there are many different ways to approach the problems we want to solve. Everything is a series of tradeoffs. Each of us clearly values different attributes above others, which leads to different solutions, and I would hope everyone here would see such explorations as healthy and desirable.18:48:06
@jryans:matrix.orgJ. Ryan StinnettGnuxie, to be honest, I find many of your reactions overly harsh at times, to the extent that I hesitate to share things here at times because of that negativity... πŸ˜“ It’s good and valuable to discuss and critique things, yes... but I would also like us to accept that there may be quite a few (and likely many) ways to approach the challenges we come across, and that’s okay.18:53:56
@gnu_ponut:matrix.orgGnuxie πŸ’œπŸ (Computing Maximalist)well, yeah, I agree with that18:55:27
@gnu_ponut:matrix.orgGnuxie πŸ’œπŸ (Computing Maximalist)and I am aware I am harsh and often dismissive 18:56:05
@gnu_ponut:matrix.orgGnuxie πŸ’œπŸ (Computing Maximalist)hmm18:57:09
@gnu_ponut:matrix.orgGnuxie πŸ’œπŸ (Computing Maximalist)idk what to do about that honestly 18:57:49
@computably:matrix.orgcomputably
In reply to @gnu_ponut:matrix.org
to argue that therefor all behavior should be explicit is to encourage people to rely on irrelevant things and also encourage the situations where you can no longer change them because of that is absurd and i'm disappointed that the collective hasn't been more critical tbh
It could also encourage people to be willing to change the language implementation for their own purposes, because it's tiny and easy to do so, making it impossible for general libraries to rely on implementation details. Not saying that's actually likely, but it's possible in theory. C is kind of a counterexample but that's not really proof one way or the other.
19:04:42
@computably:matrix.orgcomputably
In reply to @gnu_ponut:matrix.org
idk what to do about that honestly
Personally I try to keep in mind (not always successfully, unfortunately) that everybody has a different perspective, and even if I think their conclusions are wrong, their experiences and personal preferences are definitionally valid (outside of *maybe* some moral issues which have never come up in this chat afaict).
19:10:59
@jryans:matrix.orgJ. Ryan StinnettI find this guide to discourse from the Merveilles Mastodon instance to be a useful mental checklist: https://wiki.xxiivv.com/site/discourse.html19:11:11
@jryans:matrix.orgJ. Ryan StinnettPerhaps I should reference it on the code of conduct page on the site...19:11:59
@gnu_ponut:matrix.orgGnuxie πŸ’œπŸ (Computing Maximalist)Ok yeah19:12:47
@gnu_ponut:matrix.orgGnuxie πŸ’œπŸ (Computing Maximalist)fundamentally my hostility comes from a fear that people will buy into the ideas I don't like and they will get resources allocated to them while the ideas I associate with suffer and are forgotten 19:13:34
@gnu_ponut:matrix.orgGnuxie πŸ’œπŸ (Computing Maximalist)specifically when they directly compete 19:14:21
@_oftc_contrapunctus:matrix.orgcontrapunctus Gnuxie πŸ’œπŸ: that's the healthiest thing I've heard in this room in a while 😁 19:15:30
@computably:matrix.orgcomputablySelf-awareness and reflection is always good19:23:36
@jryans:matrix.orgJ. Ryan Stinnett
In reply to @gnu_ponut:matrix.org
fundamentally my hostility comes from a fear that people will buy into the ideas I don't like and they will get resources allocated to them while the ideas I associate with suffer and are forgotten
That fear is understandable, but I think we’re very far away from such concerns here. Most things we discuss are early prototypes or research projects at the moment, and I think it’s far more valuable to let them flourish so we can all learn, and build something better next time. When you see something you disagree with, perhaps instead of saying β€œno” you could view it as a challenge to build something that matches your own vision, so we can examine that alongside all the other experiments.
19:23:41
@gnu_ponut:matrix.orgGnuxie πŸ’œπŸ (Computing Maximalist)well, I already am aiming to out do all of you but I'm keeping it secret for now 😸19:29:13
@gnu_ponut:matrix.orgGnuxie πŸ’œπŸ (Computing Maximalist)so you can't say no to me πŸ˜…19:29:30
@rrix:kickass.systemsrrixΒ πŸ• 🌎i usually don't have much to add in the conversations that happen in here because i think i'm more web-minded and don't have PLT/CS experience to pull on but i find it all quite interesting22:30:20
@rrix:kickass.systemsrrixΒ πŸ• 🌎i'm just quietly stacking org-mode documents on top of each other and calling it an operating system, y'all are working on a different scale than i am right now22:33:31
@rrix:kickass.systemsrrixΒ πŸ• 🌎 * i usually don't have much to add in the conversations that happen in here because i think i'm more web-minded and don't have PLT/CS experience to pull on but i find it all quite interesting22:35:14
10 Apr 2021
@berkan:matrix.orgberkan joined the room.13:22:54
@khinsen:matrix.orgkhinsenA comment on Hyrum's law: I think it matters if an interface documentation is a agreed-on standard with multiple implementations, as it is for Common Lisp.14:16:41

There are no newer messages yet.


Back to Room List