5 ms·

Just use IRC ... the problem with writing down all ideas and all discussion like with Git issues or a forum is that the few good ones get buried in all the shit

by z3t4 8y ago

Just use IRC ... the problem with writing down all ideas and all discussion like with Git issues or a forum is that the few good ones get buried in all the shit. With IRC you can only have one or two discussions at the time, so you won't have the problem of following 30+ simultaneous discussions. IRC is a form of grinding, after some time, the community will have reached a consensus, and all arguments for and against something will be known by most regulars, so when someone new joins and say something that already has been discussed a thousand times, one of the regulars can say that these are the trade-offs and this is why it has not been implemented. Good ideas or important issues will be repeated over and over until they are either implemented or fixed, so prioritizing issues becomes natural.

marcus_holmes 8y ago
IRC (and Slack) are really vulnerable to people missing out on the discussion altogether, though. Especially in a global organisation, where only part of the team are awake at any one time, so it's impossible to have a real-time discussion regardless of platform. The advantage of forums is that it's not a real-time discussion and everyone gets to contribute.
- Terretta 8y ago
  > forums Why not revision controlled design docs as readmes with parallel discussion talk pages (discussion threads) Wikipedia style?
MaulingMonkey 8y ago
> Just use IRC ... the problem with writing down all ideas and all discussion like with Git issues or a forum is that the few good ones get buried in all the shit. Buried is still searchable and linkable, IRC generally isn't. Poorly rehashing old discussions generally doesn't help improve the signal to noise ratio either IME. It's not necessarily obvious where all discussion should take place either. I started writing a VisualRust patch to add stdlib natvis files to MSVC projects, which instead morphed into a better rustc patch: https://github.com/rust-lang/rust/pull/43221 https://github.com/rust-lang/rust/pull/43221 Thanks to this buried gem I stumbled across in a search: https://github.com/rust-lang/rust/issues/36503#issuecomment-250405902 https://github.com/rust-lang/rust/issues/36503#issuecomment-... If that was an IRC discussion somewhere instead, it probably wouldn't have gotten done (or at least not nearly so soon), because I'd be chatting and listening in the wrong channel/silo.
- z3t4 8y ago
  The human brain already filters out sounds (noise) that are not important. It filters out vision that is not important. It filters out smell. etc. And it does so automatically. But we can not yet filter out information, and too much information debt can cause anxiety. And issues can get old very fast in software as things change, but there is no-one to delete or update posts that are no longer relevant. I think you where lucky that the 9 month old post was still relevant. Some organizations solve the organizational dept by ferociously closing down issue/threads and only work on the open ones, while also making sure there are no more open ones that they can handle.
  - MaulingMonkey 8y ago
    My experience is that desired features always vastly outpaces the ability for software to keep up with it, resulting in the very triaging of issues you mention - closing them quickly. The flip side of this is that most ideas never get worked on, and a lot of them do stay relevant. Sometimes the code that feature would operate on has changed significantly, sometimes the need for it has been mooted or enhanced by other changes, sometimes the change has been made significantly harder (or easier!) by other changes. But often the idea on it's own is relatively unchanged, and closed is just another form of buried - still searchable, reopenable for the right reason.

by z3t4 8y ago

marcus_holmes 8y ago

IRC (and Slack) are really vulnerable to people missing out on the discussion altogether, though. Especially in a global organisation, where only part of the team are awake at any one time, so it's impossible to have a real-time discussion regardless of platform. The advantage of forums is that it's not a real-time discussion and everyone gets to contribute.

Terretta 8y ago
> forums Why not revision controlled design docs as readmes with parallel discussion talk pages (discussion threads) Wikipedia style?

MaulingMonkey 8y ago

> Just use IRC ... the problem with writing down all ideas and all discussion like with Git issues or a forum is that the few good ones get buried in all the shit. Buried is still searchable and linkable, IRC generally isn't. Poorly rehashing old discussions generally doesn't help improve the signal to noise ratio either IME. It's not necessarily obvious where all discussion should take place either. I started writing a VisualRust patch to add stdlib natvis files to MSVC projects, which instead morphed into a better rustc patch: https://github.com/rust-lang/rust/pull/43221 https://github.com/rust-lang/rust/pull/43221 Thanks to this buried gem I stumbled across in a search: https://github.com/rust-lang/rust/issues/36503#issuecomment-250405902 https://github.com/rust-lang/rust/issues/36503#issuecomment-... If that was an IRC discussion somewhere instead, it probably wouldn't have gotten done (or at least not nearly so soon), because I'd be chatting and listening in the wrong channel/silo.

z3t4 8y ago
The human brain already filters out sounds (noise) that are not important. It filters out vision that is not important. It filters out smell. etc. And it does so automatically. But we can not yet filter out information, and too much information debt can cause anxiety. And issues can get old very fast in software as things change, but there is no-one to delete or update posts that are no longer relevant. I think you where lucky that the 9 month old post was still relevant. Some organizations solve the organizational dept by ferociously closing down issue/threads and only work on the open ones, while also making sure there are no more open ones that they can handle.
- MaulingMonkey 8y ago
  My experience is that desired features always vastly outpaces the ability for software to keep up with it, resulting in the very triaging of issues you mention - closing them quickly. The flip side of this is that most ideas never get worked on, and a lot of them do stay relevant. Sometimes the code that feature would operate on has changed significantly, sometimes the need for it has been mooted or enhanced by other changes, sometimes the change has been made significantly harder (or easier!) by other changes. But often the idea on it's own is relatively unchanged, and closed is just another form of buried - still searchable, reopenable for the right reason.