[__d]

So, compare this with say the Python2 to Python3 migration.

Similar motivations: the developers had some legacy decisions that were unfixable without breakage. But they were sick of it, and decided to just go for it.

Most end users didn’t care about those issues. The few that did were happy to pay the cost of switching. Everyone else clung to Python2 for years because migrating was high cost and low value.

It took about 15 years to complete the migration for most, and there are a small number of users who will never make it over.

Perl5 to Perl6 is another useful historical example.

FOSS development is managed by the developers, and so, compared to a commercial software project, the implementation issues get more weight. This sort of thing is very likely to happen again and again.

[Blikkentrekker]

Not only that, the situation with Wayland also made me kind of afraid of the future of open source because it dawned on me that many of the figureheads in open source are actually simply put mentally unstable and extremely zealous and lack nuance. It didn't occur to me before but look at all the figureheads in free software: Theo de Raadt, Richard Stallman, Ulrich Drepper, Lennart Poettering, Linus Torvalds, Drew Devault. They are all kind of extremely uncompromising people who refuse to listen to reason with many of them even being known for vitriolic Twitter rants.

The issue is that free software is fundamentally a political thing and it seems to attract very political people who treat software like an ideology rather than a product who are out to wage war.

[__d]

I wouldn't say "mentally unstable", but zealous is probably fair.

To create something like the GNU project, or OpenBSD, or Linux, takes serious levels of commitment. You really have to believe in it, and to a degree, you have to _will_ it into being. Along the way, you need to explain why your crazy idea is worth all the sacrifice, discourage those who would distract your team members, maintain your own and the team's focus through years of not actually having the thing you want in any useful form, etc, etc. You have to be an unreasonable person to take it on, and then continue it.

There are people who become "fans". They can be even more zealous than the project leader(s). Maintaining direction (aka control) of a horde of over-zealous fans takes aptitude and patience. It's easy, I think, for projects to devolve into vitriol, and denigration of those who think differently, even if it starts out from a good place.

All group endeavors are ultimately political. A group endeavor with a multi-year payoff period and no tangible rewards? It's bound to be very political.

That said, we all enjoy the fruits of their labors ...

[Blikkentrekker]

> That said, we all enjoy the fruits of their labors ...

Well, we also enjoy the issues. When you talk to them they are extremely uncompromising in practice and extremely tribalistic. I think “tribalistic” is maybe a better word for what I feel is an issue. “Not invented here syndrome” reigns supreme in open source and in general it's full of extreme fanboys who aren't willing to admit anything is wrong with “their tribe” and aren't willing to acknowledge any issue whatsoever and defend everything to the death.

The opposite is also just as true though. Many of the users and figureheads will believe everything is wrong with “other tribes” and refuse to acknowledge any of the merits and good ideas.

Proprietary developers have no allegiance but to money and there's something to be said for that. They just work for a company because it pays them and will switch to another company when they get a better contract there and in many ways that makes far less loyal and thus level headed about many things when talking to them.

[NetMageSCW]

For some reason this is reminding of the Gimp developers decision to default to saving in Gimp format, even if you opened a JPEG file, and there response to complaints was if you don’t like it, don’t use Gimp. So I don’t.

[Blikkentrekker]

Yes, that, or the “Use case for <extremely useful and obvious thing>?” memes. Ebassi once got father angry at me after finding out that I did not run Polkit or a system dbus on my system and alledged that I must not know what they do because everyone would want that.

[lostapathy]

Did perl5 to perl6 actually happen? I feel like perl mostly fell out of favor along the way.

[MBCook]

No. Perl 6 was renamed Raku (?) so people wouldn’t be confused that the 5 line was continuing development.

Basically, to the degree I understand, the language was effectively forked into two.

[__d]

Disclosure: I'm not intimately familiar with all this.

I think Perl5 was originally planned to be replaced by Perl6. Then Perl6 took much longer than anyone expected, and kinda ended up in a different place. Perl5 was re-anointed as the once-and-future Perl, and what had been Perl6 became Raku.

If I remember correctly, somewhere in the middle of all that there was talk of running Python (and other languages) on the new Perl6 VM.

[lizmat]

The Rakudo implementation of the Raku Programming Language uses the MoarVM, which is pretty much a generic VM. All you need to do(TM) is write a grammar and associated actions to build the right bytecode out of the given Python source.

[Polizeiposaune]

It's a good example of a migration that mostly didn't happen.

[__d]

From one perspective, the XLibre folks seem to be taking the Perl5 path, and hoping Wayland is Raku.