Re: [9fans] Software philosophy

[Keith Gibbs <k@pixelheresy.com>]

[-- Attachment #1: Type: text/plain, Size: 7442 bytes --]

Hej Eli,

I think, philosophically, there is a disconnect re: the status of 9front vs 9legacy by 9legacy folks. Even David suggested that 9legacy is “real” Plan 9, rather than a fork. Fundamentally this is semantics, but it leads to seeing them as fundamentally different things, when they really aren’t.

The true disconnect is what position within each hold in regards to each other. Historically, forks (after a point) stay forks and no longer contribute “upstream” to their parent for love and cookies. They become cooperative peers [in good cases] or competition [in not so good cases]. Even *if* 9legacy is the “true" torchbearer, the fork happened ages ago. Some in the 9legacy camp see 9front as a downstream project that should [be forced] to make patches for the “true” Plan 9. Most 9front devs and users see them as peer projects and feel that pointing to the source code and letting 9legacy folks make their own patch the reasonable answer. Most 9front people have no vested interest in porting anything to 9legacy, since it is not the one they use. To them, it is like Ubuntu being expected to upstream cherry picked features and bug changes hand selected by Debian developers who don’t want to do it. Pretty absurd.

Insofar as what David du Colombier said that 9legacy is a “continuation” of Plan 9 from Bell Labs, sure. However, cinap or hiro or Ori or a bunch of other people here can make that same argument with 9front. 9front came about because people were slow to fix things or reticent to change things. In this way, *both* are continuations. 

But in the end, looking at the project as a peer vs project as a subordinate offshoot frames how further dialog and cooperation is done. In the former, we can point at code, debate healthy re: what level of compatibility is worth it [i.e. what Plan 9 from Bell Labs version 2 software are you really wanting to run and not just update a few lines of code…], what provisional changes can be made to fix issues but maintain old interfaces while everyone catch up, bugs in the legacy code that can be fixed, how can we collectively showcase software tools [non-OS code] made by the collective community, etc. In the latter, it will typically degrade to finger wagging for doing something that steps over some invisible line or demanding that specific changes be ported to the “real” one… i.e. 9front contributors have the bulk of the emotional and physical labor supporting a version they will never use.

hiro made a bit of a tongue in cheek, shit talking quip re: “lol but it is” since 9front, for good or for bad, probably commits orders of magnitude more code than other 9family projects. And from the 9front ml and code discussions, the community does keep pretty high standards in not just committing crap, cruft, or flights of fancy into the repo. Design wise, both 9legacy and 9front stick to simplicity and cleanliness present in the software culture of the Bell Labs team. It isn’t like one is crazy bloat and the other is elegant… more that one is less adverse to pulling the out the whole engine to fix the car and the other is more adverse. One is less interested in backwards compatibility with versions from 20 years ago for backwards compatibility sake and one wants to not have anything not run that V4 can. Both are different strategies and have different benefits...

So not to belabour things further, I think we kind of need to come to somewhat of a consensus re: how these two project relate. I honestly think that imposing a “One Plan 9” or reframing 9legacy as the authoritative parent project will in fact harm 9legacy more the 9front, as the latter is more comfortable doing its own thing and honestly, 9front works better on more hardware and is more actively updated and supported. Agreeing that both projects are sister projects allows more dialog and actual sharing to happen.

My intention was not to spark some sort of holy war and I get the feeling most people in this community see the two are peer projects. When people float the idea of the P9F imposing a “One Plan 9” by dictum rather than the actual codebases, community members, etc. deciding how things should work, well, that needs to be called out…

-pixelheresy

> On 19. Aug 2021, at 1.00, Eli Cohen <echoline@gmail.com> wrote:
> 
> what is all the friction actually about here?? the most important
> philosophical question always ends up the same, how can we figure out
> a good formula for not being jerks?
> 
> I have ended up using 9front more and more, obviously. 9front was
> started specifically to address the fact that Plan 9 from Bell Labs
> didn't run on most computers... If I have any feeling at all about it,
> it's that there's room for another fork that is an even simpler
> research platform. in other discussions people say, why do we have
> things that aren't relevant? We all love catclock, email... some users
> may only want plan 9 for that... some people also discussed even
> removing compiled binaries as much as possible. mostly, I like the
> idea of plan 9 that runs on the computer I have... but I understand
> that for a lot of reasons other people don't necessarily feel the same
> way.
> 
> we wouldn't be here if we didn't agree Plan 9 is the best OS design.
> and they're all free software. 9front has some very interesting things
> that 9legacy can (and does) use as patches. it's just actually
> difficult to write software, for some value of difficulty.
> 
> there's a lot of shit going on in the world today... we all gathered
> here to agree Plan 9 is great software, then just be rude to each
> other because...? I really don't understand, I'm not exaggerating.
> what is the actual disagreement here?
> 
> On Wed, Aug 18, 2021 at 1:12 PM David du Colombier <0intro@gmail.com <mailto:0intro@gmail.com>> wrote:
>> 
>> Here are some clarifications.
>> 
>> 9legacy used to be a an experimental patch queue for
>> Plan 9 from Bell Labs, providing patches that were not
>> yet accepted into the mainline distribution. That's why
>> we didn't recommend to use 9legacy, unless you had
>> specific needs.
>> 
>> However, this isn't really the case since 2015, because
>> Plan 9 from Bell Labs is not maintained anymore
>> (last release was 2015-01-10).
>> 
>> Today, 9legacy is more of a continuation of Plan 9 from Bell Labs.
>> There are still experimental patches, but also a lot of fixes and
>> improvements that would probably be part of Plan 9 from Bell Labs
>> if it was still maintained.
>> 
>> Also, NIX is not maintained anymore. However, there are
>> some other variants of 9k (the 64-bit Plan 9 kernel), including
>> the one available as part of 9legacy, that are still in progress.
>> 
>> --
>> David du Colombier
> 
> ------------------------------------------
> 9fans: 9fans
> Permalink: https://9fans.topicbox.com/groups/9fans/T9ef6430f3025e731-Me55ae2eef0de0a39ecd205ad <https://9fans.topicbox.com/groups/9fans/T9ef6430f3025e731-Me55ae2eef0de0a39ecd205ad>
> Delivery options: https://9fans.topicbox.com/groups/9fans/subscription <https://9fans.topicbox.com/groups/9fans/subscription>

------------------------------------------
9fans: 9fans
Permalink: https://9fans.topicbox.com/groups/9fans/T9ef6430f3025e731-M51f0a671128e643ab96b6b11
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription

[-- Attachment #2: Type: text/html, Size: 31822 bytes --]