Re: [9fans] Software philosophy

[Keith Gibbs <k@pixelheresy.com>]

> On 18. Aug 2021, at 14.28, Lucio De Re <lucio.dere@gmail.com> wrote:
> Let me put it this way: German and Italian motorcycle manufacturers
> eventually figured that the gear shift should be on the same side as
> Japanese manufacturers preferred.

This is my point exactly in one sense. In your example, there were alternate designs, and the German and Italian manufacturers conformed to the Japanese ones not by the Motorcycle Gearshift Foundation (MGF) sounding off on it or imposing a design for a “One Motorcycle Gearshift” [1MG] initiative. Instead, the experience of users and/or the better ergonomics of the design and/or functional logic dictated it was easier to just make them the same. 

Thought exercise: Fast forward 5 years from now… the NIX project delivers on a new version of the Plan 9 kernel that is brilliantly optimised for multicore/multinode distributed computing. Within a few weeks, 9front says, “this is the dopest thing ever” and folds it into their fork, along with some monkeying they have done to the kernel. Both get a shot in popularity and draw additional people into the Plan 9 community space. People develop software with this set of features in mind.

Let’s even say that the P9F takes leave of their senses and dictates a new Plan 9 from Bell Labs V5, which is basically 9legacy with standard patches and a version number. Great. Would “official” Plan 9 want to take in the cool addition to be more compatible with closely related peers or would people balk about the other distress “not contributing to common cross-compatibility” and come up with it’s own thing and stick to their own thing. I would venture the former would happen for the same reason in your motorcycle example. 

> What I am proposing is that where some code will run on one flavour of
> Plan 9 and not on another, which is annoying, that somebody be
> entrusted with the common sense to suggest which of two
> implementations should be favoured and for what reasons.

In the case of plurality, you have a push and pull, where people may intentionally make a departure from compatibility for the sake of bettering things. That is a risk, sure, but that is where you get growth. And if it is the best, since it is free and open source, the better code can be worked into sister projects. It’s like when there is a win, everyone can win. And mind you, most software is more about different tastes, workflows, etc. and don’t affect capital C “Compatibility”.

Similarly, a breaking change could be a problem and could even harm a project in the long term. That is where the risk comes in. 

> It seems to me that the paranoid individualist assumes malice behind
> such an obvious proposal.

I’m not sure malice is the right word. It is “motive”. It is clear you don’t like 9front as a distribution. That is your choice. But this isn’t the first time you have either directly or indirectly brought up the desire to propose that there should be a single Plan 9 and that what is official is dictated by editorial decision by a governing body. 

We obviously have very different views on top down vs bottom up development. However, most OSS development comes from a volunteer putting out a pull request that he or she things would make things better or a little standalone tool that people could find useful. The absense of authority, even in your motorcycle example, is not chaos. We have a community with some very smart and talented people. 

-pixelheresy

> Lucio.

------------------------------------------
9fans: 9fans
Permalink: https://9fans.topicbox.com/groups/9fans/T9ef6430f3025e731-Mdec4678ae5d02a14e4d6de91
Delivery options: https://9fans.topicbox.com/groups/9fans/subscription