Re: Future Direction of GNU Hurd?

[Richard Braun]

On Sun, Sep 13, 2020 at 03:01:44PM +0200, Paul Boddie wrote:
> Yes, you have been lucky to get a couple of good people via GSoC to work on 
> the Hurd. Evidently, last time I checked, there wasn't the bandwidth in the 
> project to mentor people for that or similar initiatives any more, which 
> should be a pretty worrying sign.

I considered it, and decided not to go ahead, because the student in
question wasn't convincing enough to me.

> [Opportunities for non-experts; volunteer culture and burn-out]
> 
> One of the most frustrating things about certain projects and certain 
> workplaces is the neglect for systems that retain and disseminate 
> organisational knowledge. Many people have presumably had the experience of 
> being told that they should ask this person or that person about some piece 
> of 
> code or other, and since nothing gets written down, they have that experience 
> over and over again. Eventually, this wears them down, makes them feel 
> inadequate, and they wonder how anyone gets anything done.

OK, but I'm not saying not to note things down, on the contrary.

> It is one thing to say that people should be able to "look up things for 
> themselves" - after all, we have Google these days - but it is another to 
> actually curate resources that allow people to find these things without 
> spending days of their own time doing so (or, frequently, just digging into 
> the source code, running programs to stress that code to figure out the 
> corner 
> cases, and so on). With the Hurd, plenty of documentation enquiries end up on 
> wiki pages showing copy-pasted IRC conversations in their usual tedious style.

That way of "noting things down" probably ended up being used because it
was cheap enough for the people who did it, considering their available
time. I'm all for improving that.

> If you think that people are inconveniencing you or others by asking 
> questions 
> whose answers are obvious to you, you might want to consider the 
> inconvenience 
> those people are experiencing and the impact that has on their motivation. 
> And 
> again, a project of this scale needs a broad range of contributors, some of 
> whose questions might seem superficial or "lazy" to you, but which actually 
> might be quite reasonable more generally.

I have never ever said such a thing. I wouldn't have acted as a mentor,
on GSoC and elsewhere, and a teacher, if that was the case. What I call
"bad lazy" is negligence and fear of effort.

> You are so utterly wrong about volunteer burn-out that I can only assume you 
> are unaware of the social dynamics around many Free Software projects. And 
> while I would normally steer clear of using such absolute words as "wrong", I 
> can use them with confidence in this case because I have followed various 
> significant projects over many years and have seen well-known contributors in 
> such projects document their own burn-out experiences. So to dismiss such 
> concerns as things that do not apply is actually rather offensive to people 
> who did burn out working on Free Software.

I'm not dismissing their experience, but you seem to consider them the rule
rather than the exception. You're questioning *our* decency as project
contributors/maintainers as if *we* were responsible for the fact that
other people would burn out from *their* own issues. I can't recall a
single instance where contributors were pressured into getting things
done, only regular reviews about their contributions (again, except for
Svante Signell who tried to boss people around but everyone dismissed as
noise).

> You have already spent quite a bit of time defending what I would call 
> "audition culture", meaning that people have to meet a particular set of 
> admission requirements before you will apparently even interact with them. 
> Many people would recognise this as a counterpart to internship culture in 
> wider commerce where people are encouraged to do unpaid work with a vague 
> expectation of being offered paid work after a certain period of time 
> "proving 
> themselves". In recent times, such practices have seen more scrutiny because 
> people have realised that they might be a form of exploitation.

*Before* ? Absolutely not. I interact with them, I really do try, and when
I feel it's not good enough, then I quit. That's what happened with Almudena
Garcia and why I mentioned him.

> Now, I am not accusing you or anyone in particular of exploiting people, but 
> I 
> want you to be aware of the social factors involved. People who are 
> enthusiastic about contributing to an initiative that has been promoted as 
> worthwhile, and in alignment with their own views or beliefs, are more likely 
> to "go the extra mile" to make it succeed and to try and make their own 
> contributions count.

Agreed. Not sure how it's relevant though. Again, I've never dismissed
someone for being motivated, just for not willing to do things right or
being really incapable of doing it.

> What one sees quite a bit in the Free Software arena is a very casual 
> attitude 
> towards the well-being of those people: from the perspective of a project who 
> just "needs labour", somebody quitting is just "too bad for them" and there 
> are always more volunteers to be found; nobody seems to consider the 
> perspective of the person who had to decide whether to abandon the investment 
> they had already made in a project due to the costs of remaining committed to 
> it. It is no wonder that interpersonal issues arise: does anyone even 
> consider 
> why people start behaving in a way that is then framed as irrational (often 
> to 
> justify excluding such people)?

As someone who invested a lot of time and work into the project and then
moved to something else, and who knows others who did, I perfectly know
that perspective. Virtually all people who contributed a lot and stopped
did so because life happens. And by that I mean they grow up, get a job,
sometimes a family, other interests, etc... All the interpersonal issues
have occurred with people whom I consider are not the right kind of
people to do the job. Hm, maybe there's a pattern there.

And yes, if people leave, it is "too bad", and if the project "dies" or
really, stagnates, then so be it. I don't see why the Hurd should
absolutely "succeed". It's but one possibility path among many others.
But if it is to succeed, it needs the right stuff.

> > The only exception would be Svante Signell who's the main reason for me not
> > working on the project any more [1].
> 
> See above.

Svante's case is special. He's clearly not the right person for any job
in the project. He's a troll who's unaware of being one. Despite that,
he's done contributions. They're usually extremely simple, full of mistakes,
and very likely caused negative productivity on Samuel who reviewed and
corrected everything, sometimes getting insulted on the way. I cannot
answer "why people start behaving in a way that is then framed as irrational"
in a general way, but for Svante, it's very clear. He's just stupid and
disrespectful. That happens.

> Nobody seriously claims that a "mere website" will "save" anything. However, 
> quite a few people would seriously claim that a "mere website" and the 
> associated assets of a project would certainly improve a project's viability.

Obviously, I exaggerated the idea to turn in into a parody. Again, I'm
not against completely reworking the website and adding assets. As usual,
I want the right stuff. So not a "modern" looking website with content
that changes as you scroll down just because it seems like a "cool" idea.
That's not going to work.

> I'm sorry, but I have looked at your own projects and I have read your 
> messages on these lists, and then I wonder why you are so negative towards 
> people whose only mistake is not to offer "top-level" contributions. But I 
> feel like I am not really getting through on this point.

But where the hell do you get your facts, or how the hell do you
understand what you read ?? I've just insisted many times in my previous
message that I just want people who are right for the job, defining that
as people who are able to both grasp the big picture and pay attention to
detail. Where have I mentioned I'd dismiss people who don't offer
"top-level" contributions ? Unless you define "top-level" the same as I
define "right for the job". But that's clearly not the same thing to me.

> Which misunderstandings are these? You want "top-level developers" who don't 
> spend their time pestering you and others with the organisational knowledge, 
> but you don't want people who are willing to improve the systems that provide 
> such knowledge and make it possible for other kinds of people to make a 
> difference.

Well "top-level" would be one misunderstanding. Everything else you're
mentioning in that paragraph is just wrong; that's absolutely not what I
have in mind. It's getting more than annoying.

> And since it would rather appear that the "top-level developers" are not 
> likely to improve those systems (which is an industry-wide phenomenon, of 
> course), and since they generally aren't showing up anyway, I actually do 
> wonder what kind of future the project has, whether someone else will be 
> having this conversation on this or another of the lists in a decade's time.

Like I said, if it doesn't get to be, so be it, it's really not that bad.

> Honestly, are you interested in welcoming new contributors or not? Is it 
> welcoming to take fairly valid observations about the way the project is 
> portrayed and communicated, frame them as superficial and then to tell such 
> people to (in your own words), "Grow up and focus on what matters."
> 
> https://lists.gnu.org/archive/html/bug-hurd/2020-08/msg00111.html

Here is some context for that reply :

"When I talk about Hurd with friends, my friends are afraid because the
website is old, ugly, out-of-date, and very confusing."

See [1].

"Ugly" is subjective. "Old" is irrelevant. "Out-of-date" is mostly wrong.
"Very confusing" is the problem and I agree, that is not superficial.

My "grow up and focus on what matters" comment referred to Almudena
describing the website as "90s style" [2] and later "90s ideas" [3]
as if it were a real problem, and claimed the website had to appeal to
how "younger people thinks nowadays". No, the website has to be good
at telling you how and where to find information, and it doesn't need
javascript bloat and moving stuff to do that. That is the superficial
stuff in this story.

Also, you should realize the only real reason I'm replying to you is so
that people actually don't get the wrong impression about the project.
So yes, I do care about the impression new contributors would get.

> Sorry again, but I didn't assert or even imply anything about people being 
> "too dumb" or "drama queens", although one can certainly find drama in the 
> mailing list archives. My only suspicion was that people decided that various 
> L4 kernels weren't suitable and then may not have reviewed their assumptions 
> when newer kernels emerged. Instead they went off and started new microkernel 
> projects, which the other Richard feared would be a bad idea. Well, I guess 
> he 
> was right about something else.

That's the impression me and others got when reading your message.

And being someone who started another microkernel project, I was aware
of the developments of other L4 kernel variants, and I determined it
wasn't good enough for what I had in mind. But since my project is also
stalled now, it really doesn't matter at all.

> Since the enquiry mentioned L4, being on a list somewhat concerned with its 
> potential role in the wider Hurd endeavour, I felt that someone should inject 
> some slightly more up-to-date information about L4-based systems, even though 
> there are plenty of people in the L4 community who would be in a better 
> position to do so, were they even paying attention to the Hurd any more.

And I didn't blame you for that, I blame you for the invalid portrayal
of the social aspects of the project.

> You might not appreciate my remarks about the apparent culture of the wider 
> Hurd endeavour, given that in your own words "L4 variants of the Hurd are 
> dead 
> projects", which in itself is hardly a motivating signal to those who might 
> see things differently and be looking for encouragement, but I honestly think 
> instead of pointing the finger at apparent critics and claiming to be 
> misunderstood or misrepresented, you might wish to review that culture and 
> consider realistic measures that might actually get people to contribute to 
> the effort and, crucially, to continue to do so.

As I wrote earlier, "dead" to me means that it stagnates. In the case of
Hurd on L4, there is simply no activity at all. Projects are not people,
they don't necessarily stay dead. I assumed that was obvious.

I actually question such things all the time. I don't think the culture
itself has a problem, which is why I've been so adament to defend it.
I'm actually very afraid of it changing into something a lot less
demanding. Software development, where formal solutions are still
very rarely used, is a craft, and system development is a difficult
one. People should thrive for excellence, nothing less. That's the
main point I wanted to defend.

As for realistic measures about the Hurd, there is only one way
really : getting rid of GNU Mach IPC and signal handling. That's huge
work and what triggered my work on my own project. I felt that the Hurd
wouldn't evolve from what it is into what I'd like it to be as fast as
a project from scratch, and personal preferences weighed in as well.
Obviously, if someone else can do that with an L4 kernel, or anything
else, I'd be delighted, and that's why I'm still following this list.

So yes, the only thing left for me to do is point the finger at the guy
who spreads soft FUD to newcomers.

-- 
Richard Braun

[1] https://lists.gnu.org/archive/html/bug-hurd/2020-08/msg00107.html
[2] https://lists.gnu.org/archive/html/bug-hurd/2020-08/msg00112.html
[3] https://lists.gnu.org/archive/html/bug-hurd/2020-08/msg00114.html
[4] https://lists.gnu.org/archive/html/bug-hurd/2020-08/msg00082.html