Who controls glibc? [LWN.net]

Who controls glibc?

Posted May 8, 2018 9:02 UTC (Tue) by mjg59 (subscriber, #23239)
In reply to: Who controls glibc? by spacemachine
Parent article: Who controls glibc?

"This is likely to discourage some number of people from contributing to the project, does nothing to encourage additional people to contribute to the project, and serves as a distraction from the actual purpose of the work that contains it, therefore let's take it out"?

to post comments

Who controls glibc?

Posted May 8, 2018 9:28 UTC (Tue) by spacemachine (guest, #124210) [Link] (18 responses)

Yes, the decision making process of "this will discourage some small % of potential contributors so let's take it out" is precisely the wrong way to make decisions. The cost of setting that precedent is the organization around the project itself.

The right way to make decisions is "this will encourage a large % of potential contributors so let's add it." Glibc is very far from that. Like I said, RIP.

Who controls glibc?

Posted May 8, 2018 9:40 UTC (Tue) by mjg59 (subscriber, #23239) [Link] (17 responses)

Ok, so that sounds like you don't think adding it in the first place was justifiable. What's the problem with reversing that (incorrect) decision?

Who controls glibc?

Posted May 8, 2018 10:57 UTC (Tue) by spacemachine (guest, #124210) [Link] (16 responses)

The problem is that the justification for the original patch in question is weak. It's a solution for a problem that doesn't exist. When those types of changes start being encouraged, it's the beginning of the end. The project has no direction. People are now wasting their precious limited time looking for, preparing, and submitting one-line patches that have virtually no effect on the project. Then they go ahead and pat themselves on the back for having a productive day. There are hundreds of open bugs on bugzilla, but they are ignored in favor of these pseudo productive patches. This is the future of glibc.

Who controls glibc?

Posted May 8, 2018 11:02 UTC (Tue) by mjg59 (subscriber, #23239) [Link] (13 responses)

> It's a solution for a problem that doesn't exist.

The problem exists.

Who controls glibc?

Posted May 8, 2018 11:13 UTC (Tue) by spacemachine (guest, #124210) [Link] (10 responses)

Hypothetically yes, but not in any way that solving it will drastically improve the state of glibc for its users. It's not that it's not okay to fix small things but there should be good reason for it, not some hypothetical "what if," especially when so many other things are broken. Has any actual user ever complained about this joke?

Who controls glibc?

Posted May 8, 2018 12:38 UTC (Tue) by excors (subscriber, #95769) [Link] (1 responses)

Fixing small things should only require small reasons. If you have to think very carefully for days, consult dozens of people, do a cost-benefit analysis and write up an extensive rationale document every time you clean up a bit of code or a bit of documentation, it clearly won't be worth the effort. Then the project will collapse under the combined weight of a million tiny problems that nobody bothered fixing.

Efficiently fixing trivial issues seems like a requirement for sustainable software development, not a sign of its demise. (And it sounds like this patch was being handled efficiently until Stallman got involved). Sometimes projects will explicitly talk about paper cuts (minor usability bugs that are individually unimportant, but a user who encounters dozens of them will be strongly put off) and technical debt (problems that weren't worth fixing in the short term, but their cost will accumulate until they seriously impede development) because they're aware they need to deal with those minor issues - it's tempting to ignore them and focus on the highest-priority issues instead, but it's important for the project's long-term health to work on the little things too.

(Besides, in this specific case it's easy to find people like https://stackoverflow.com/questions/48445031/why-would-it... who seem genuinely confused by the statement and wasted an appreciable amount of time trying to understand it, so it's not a hypothetical problem.)

Who controls glibc?

Posted May 8, 2018 13:17 UTC (Tue) by spacemachine (guest, #124210) [Link]

An obscure stack overflow thread does not justify this as a problem that is holding back glibc in any significant way. Do you genuinely believe removing this joke has moved the needle at all?

Who controls glibc?

Posted May 8, 2018 12:58 UTC (Tue) by likryol (guest, #115542) [Link] (7 responses)

You're right, removing it won't (at least in the short term and likely not in the long term) drastically change anything for anyone. You've been saying this is a waste of time while other bugs are still festering, but consider for a second how much time it actually took to make this change happen. Someone saw the "joke", deleted it, committed it. What is that like 30 seconds of work? This isn't a bug that needed hours of time to discover and fix technically or socially, and the quick change could have some benefits for the project in the form of more contributors.

The real issue (and time waster) was when RMS came in and exerted authority he no longer deserves to have (I understand he is an ideological lead and not someone doing technical work, correct me if I'm wrong) thus creating an argument that anyone watching from the outside will see RMS as an overstepping grossly immature leader. Again, not for the joke but because of the authority he is trying to exercise over something so minuscule. If anything deserves "RIP for glibc" it's his behavior, not the 30 seconds of work it took to remove a dumb joke. And if jokes like that aren't turning people away, it will be seeing arguments like his unfold that will make potential contributors go "oooh...maybe I don't want to get into the middle of that culture..."

And all of that you could ultimately blame on someone making minor changes that don't need to happen, fine whatever maybe you're right. But the trigger for this argument and his excessive use of authority could have been more technical and we'd have the same argument at hand, it's probably happened before. This specific instance is easy to separate the technical from the not and so it ends up being reported on.

Who controls glibc?

Posted May 8, 2018 13:11 UTC (Tue) by spacemachine (guest, #124210) [Link] (2 responses)

Start to finish it takes much longer than 30 seconds for an outside contributor to get that change checked in. What kind of engineer goes through that just to remove a dumb joke? Those are the kinds of contributors glibc has these days. These aren't the types that make a project great.

Who controls glibc?

Posted May 8, 2018 19:15 UTC (Tue) by rra (subscriber, #99804) [Link] (1 responses)

Do you follow glibc development closely? I ask because to me this is a bizarre statement given how dramatically better glibc has gotten in the past few years, in ways that are the exact opposite of what you're implying. It makes me think you're out of touch with current glibc development.

Right now, glibc is blessed with numerous people who are tackling large-scale, impactful work in substantial patch sets on topics ranging broadly from better standards compliance to security improvements to Y2038 issues. I have never seen the project healthier, and I've been following it for over a decade. Even the Hurd port is being resurrected from the dead, which regardless of one's opinions of the importance of this to the broader community is definitely not a trivial or minor effort. And this is happening without, so far as I can see, any slowing down of other work, which speaks to the breadth and capacity of the current development community.

Who controls glibc?

Posted May 9, 2018 11:35 UTC (Wed) by nix (subscriber, #2304) [Link]

I concur. glibc's gone from being a project that was verging on moribund to a project that's so active that keeping up with the mailing list alone, let alone the patch flow, is quite a lot of work. It's *lively*, and the changes taking place there are extremely nontrivial. (Even some of the changes with no functional effect -- e.g. classifying every entry point according to its multithread-safety.)

Who controls glibc?

Posted May 17, 2018 22:37 UTC (Thu) by HelloWorld (guest, #56129) [Link] (1 responses)

I can't blame RMS for not wanting his project to be taken over by ideologists. It's called Free Software for a reason. Freedom includes the freedom to say and write potentially hurtful things. And for the record, I *don't* think that joke is actually hurtful, and so far nobody has been able to produce an example of a person who was actually hurt or offended by this silly little joke. I don't believe there is such a person.

Who controls glibc?

Posted May 24, 2018 17:34 UTC (Thu) by nix (subscriber, #2304) [Link]

If you think glibc is, in any meaningful sense, "RMS's project", you are deluding yourself. Before this thread, Richard had a total of five commits, all in 1995, all to either config.sub or config.guess, none more then 50 lines. (Its history starts in 1989, when the first RCS commit was made.)

This was Roland's project, then Roland and Ulrich's and a few others. Now it is a shared, community-governed project, and frankly RMS's trying to exert dictatorial control over it feels quite offensive, given that there is no sign of him in that community of developers at all.

Who controls glibc?

Posted May 25, 2018 18:54 UTC (Fri) by shmget (guest, #58347) [Link] (1 responses)

"You're right, removing it won't (at least in the short term and likely not in the long term) drastically change anything for anyone"

leaving it would not drastically change anything either, and would have taken 0 seconds.

"The real issue (and time waster)" is the removal patch. which created a Streisand effect on an obscure joke that pretty much no-one was aware of.

The stackoverflow link above, when I looked at it had been vewed 790 times.. and I bet most of it was because it was linked above.
but, I been consulting man pages for years, on myny different boxes and distro. I've never seen the 'joke' before
and it still not visible in any 'man abort' I've just ran on a few varied boxes.

iow: that joke has very little visibility, certainly epsilon wrt to 'floss user'. removing it will have 0 effect wrt to drafting new contributor, it it will have a small cumulative effect wrt to discouraging existing ones.... just like the wave of pronoun-war patches inflicted on floss, which had real effect of getting an actual maintainer to call it quit, for what SJW swear were hordes of 'potential contributor' that were not showing up because of it.... yeah .. how things are going in node.js world ?

Who controls glibc?

Posted May 27, 2018 0:52 UTC (Sun) by anselm (subscriber, #2796) [Link]

I've never seen the 'joke' before and it still not visible in any 'man abort' I've just ran on a few varied boxes.

That's probably because it isn't in the man page, it's in the Info documentation. Still doesn't make it funny.

Who controls glibc?

Posted May 17, 2018 22:19 UTC (Thu) by HelloWorld (guest, #56129) [Link]

I'm sure it does exist in your head.

Who controls glibc?

Posted Aug 1, 2018 11:57 UTC (Wed) by diegor (subscriber, #1967) [Link]

If people is offended by abort, we should remove the abort() function. The joke does'nt add anything.

So why we don't censor every reference to "kill children" (process). Maybe someone have lost his kid, and be reminded of her lost.

Who controls glibc?

Posted May 8, 2018 12:24 UTC (Tue) by anselm (subscriber, #2796) [Link] (1 responses)

There are hundreds of open bugs on bugzilla, but they are ignored in favor of these pseudo productive patches. This is the future of glibc.

So Stallman thinks that as the notional “project leader” he gets to decide that his lame paragraph (I won't dignify it by calling it a “joke”) must stay in, but he doesn't think that as the notional “project leader” he ought to see about getting those hundreds of open bugs fixed? Some leadership.

Who controls glibc?

Posted May 17, 2018 22:38 UTC (Thu) by HelloWorld (guest, #56129) [Link]

RMS always made it quite clear that he cares about freedom and not about software quality.

Who controls glibc?

Posted May 8, 2018 17:08 UTC (Tue) by fuhchee (subscriber, #40059) [Link] (9 responses)

Counterpoint: "The presence of less-than-100%-PC humour is likely to *attract* some number of developers to the project. Therefore let's leave it in."

Who controls glibc?

Posted May 8, 2018 22:12 UTC (Tue) by jubal (subscriber, #67202) [Link] (8 responses)

Re-counterpoint: unfounded assumption.

Who controls glibc?

Posted May 8, 2018 22:22 UTC (Tue) by fuhchee (subscriber, #40059) [Link] (5 responses)

Well the point is what is the burden of proof? I could point to ... myself ?! ... as someone who generally prefers a more relaxed type of community. I'm not the only one. So this is a reciprocal of the "if it turns someone, anyone away, it's bad" argument.

Who controls glibc?

Posted May 8, 2018 23:22 UTC (Tue) by nilsmeyer (guest, #122604) [Link] (2 responses)

Interesting point. I usually sometimes find myself re-phrasing and re-phrasing commit messages and even code comments over and over, especially when it's correcting a mistake someone made, at least at work. It's easier for private projects, but the potential to offend genuinely scares me.

Who controls glibc?

Posted May 10, 2018 11:16 UTC (Thu) by zenaan (guest, #3778) [Link] (1 responses)

To paraphrase you: "The work to avoid offending the various categories of snowflake safe-space-junkies, can be a real cost, can be significant, is a detriment, and the chilling effect this all has is real."

THAT is the reason this particular joke (by RMS) should stay in the glibc manual.

"The triggered" and "the oppressed" are redefining permissible speech - which is ironically apropos RMS' original joke.

The redefinition of allowed speech is dangerous and literally tyrannical in the underlying intent of doing so (whether conscious, or unconscious) - refer Dr Jordan Peterson who puts this exact point so succinctly.

Create your world, folks,

Who controls glibc?

Posted May 10, 2018 22:11 UTC (Thu) by tvld (guest, #59052) [Link]

You keep advertising this opinion throughout this comment thread. But do you actually know the glibc developer community? Do you know what they think works well for them, and what helped them make all the technical progress on the project in the recent years? Have you attended a GNU Tools Cauldron, for example, to see how they work together and what kind of environment the actual developers want? The glibc community *is* building the environment they want.

Speaking as someone who has contributed to glibc in the recent years, my impression was that nobody was or felt bullied. Developers just *wanted* to be friendly to each other. IOW, you misjudge what drives this.

There's nothing wrong with a majority wanting to be friendly people in the first place and not being interested in bothering with unfriendly behavior.

Who controls glibc?

Posted May 9, 2018 0:02 UTC (Wed) by likryol (guest, #115542) [Link] (1 responses)

I hear you but I think we can have a "relaxed community" without making questionable jokes in an official, long lived, and widely seen media.

Also I derailed the original point of this comment thread which is I think to say that if anything is adding up to "RIP glibc" it isn't the removal of OT content, it's the BDFL being excessively authoritarian.

Who controls glibc?

Posted May 10, 2018 16:43 UTC (Thu) by nilsmeyer (guest, #122604) [Link]

I think that's also something people in this debate often forget, the intent of the original committer is probably not to censor RMS or derail the project by starting this debate.

Who controls glibc?

Posted Jun 6, 2018 17:32 UTC (Wed) by clicea (guest, #75492) [Link] (1 responses)

Have a counterexample: I've stayed very far away from Rust just because they seem very heavy-handed on policing their developers.

Who controls glibc?

Posted Jun 7, 2018 16:27 UTC (Thu) by peter-b (guest, #66996) [Link]

Yes, I agree that the Rust community moderation policy works really well at fostering a productive and friendly atmosphere.