Ask HN: Difference in management styles of Ulrich Drepper and Linus Torvalds

60 点作者 tech6将近 11 年前

Linus Torvalds stewardship is often cited as critical factor in thriving linux kernel development Ulrich Drepper is often cited for preventing widespread community contribution to glibc and causing forks such as eglibc. For casual observer like me it appears both are technically brilliant individuals who have strong opinions and dont suffer fools lightly. What caused linux to thrive ? Was it the organizational heirarchy of having individual maintainers for different components in the linux kernel ?

13 条评论

JoshTriplett将近 11 年前

Linus accepts patches for just about anything, only discriminating based on patch quality and general development direction of the kernel, without discriminating against the specific thing someone wanted to add. Ulrich went on record several times as not caring about most architecture ports or various other areas, and he rejected patches in those areas, forcing them to live in add-on areas like the "ports" repository. Distributions had to maintain a large number of patches.Linus cares deeply about the quality of the kernel, but leaves it to developers to determine its direction. Ulrich had very specific thoughts on the direction of glibc, and screw you if you cared about anything else.Linus has a large number of trusted submaintainers, and delegated a large amount of trust and judgement over patches to them; Ulrich took everything on himself, and simply rejected patches in areas he didn't care about.Linus has been known to change his mind, graciously, in the face of convincing argument; he also readily admits being wrong. Ulrich, not so much; if he didn't like your patch, you might as well give up.While Linus can be unnecessarily awful to people he thinks he can expect better from, there's an undercurrent of doing it to defend kernel quality, and he usually has a point underneath the invective; I also don't think I've ever seen Linus go off on a new developer who's still learning and making reasonable mistakes. Ulrich was awful to everyone, with or without reason, and he was wrong far too often to get away with it.Finally, for every discussion in which Linus ranted at someone, I've seen several dozen where he gets involved in a complicated technical problem and helps find the solution, or works with someone else to do so; the balance is overwhelmingly positive.Don't look at either of their abrasive communication styles as worthy of emulation; do better than that. Linus successfully manages Linux despite being occasionally abrasive. Ulrich's abrasiveness just added to his mismanagement, rather than being the sole source of it.

pessimizer将近 11 年前

When Torvalds hurls invective at some idea and at anyone who would support it, it's clear that he read it, understands it, and is speaking directly to it. Drepper would just swear at people and call them stupid.The way to figure out the difference between the two is to strip out the choice of language and focus on the content of their responses. Drepper, if you strip out the invective, rarely has any content at all. Torvalds' sweary hateposts are often some of his most profound, detailed, and quotable.

breckinloggins将近 11 年前

If I remember correctly, the main problem with Mr. Drepper was that he was resistant to accepting changes to glibc... even (subjectively) good ones.Torvalds, like Drepper, is often criticized for a less-than-friendly attitude, but it is generally acknowledged that he lets the Linux kernel progress.

评论 #7954086 未加载

评论 #7954452 未加载

A_COMPUTER将近 11 年前

Bashing Ulrich Drepper has now taken on the flavor of bashing Nickelback, it's a fun and socially safe outlet for hate. The guy hasn't even been a thorn in anyone's side for years now, it's time to let it go.

forgottenpass将近 11 年前

Like all software projects, glibc and the kernel are resistant to change. The kernel not (often) unreasonably so. Drepper was notorious for being unreasonable.Gregkh gives talks about kernel maintenance, here's an excerpt:<pre><code> It is in my self-interest to ignore your patch [...] So your goal is, when sending a patch, is to give me NO excuse to not accept it. To make it such that if I ignore it, or reject it, I am the one that is the problem here, not you. </code></pre> <a href="https://github.com/gregkh/presentation-linux-maintainer/blob/master/maintainer.txt#L301" rel="nofollow">https://github.com/gregkh/presentation-linux-maintainer/blob...</a>Simply, Drepper's excuses sucked, he was the problem.For casual observer like me it appears both are technically brilliant individuals who have strong opinions and dont suffer fools lightly.Also, perception of Linus' attitude is more smoke than fire. He does suffer fools lightly, it's the people he expects better of or explicitly provoke him that get the angry responses that make LKML show up on social media.

krakensden将近 11 年前

The assholeness of Torvalds is oversold. Imagine that every conversation you had at work, and a good deal of your conversations with your friends off work, were publicly available on the internet and routinely scrutinized by journalists and bloggers, and any time they criticized you it was newsworthy enough to make it to all the big social news aggregators. Now imagine that you worked a retail job, and not only dealt with professionals with a sense of propriety, but randoms with varying degrees of anger, malice, and incompetence.

lifeisstillgood将近 11 年前

I cannot say I have studied either in sufficient detail but I will say there is a huge body of academic and less-academic work on the effectiveness of leaders and leadership style in commercial and corporate markets - and none of it is conclusive in anyway.However, in these cases there is a vast corpus of emails and comma - it's as if we had a stenographer following Lou Gerstner and Steve Jobs around, so I would be surprised that now you have mentioned it, there is not some academic attempt to quantify answers.Hell, if anyone fancies a middle aged student for a research thesis I would happily take a run at this :-)

colin_mccabe将近 11 年前

There really is not much similarity between Linus and Ulrich. Linus relies on his "lieutenants" to do most of the work in Linux... not only the code, but also reviewing the patches of others. Ulrich was a one-man show who would often rewrite patches rather than accept the work of someone else.Anyway, Ulrich was just wrong about a lot of things. I remember one bug that affected ARM that he tried to close because "ARM is not a real arch." He made a lot of very questionable decisions, like leaving out strlcpy and strlcat. To be fair, there is valuable stuff he did behind the scenes (like ELF TLS support and some of the dynamic linking documentation). But people (reasonably) focus on the negative things, because a negative maintainer can cause huge problems for downstream projects...

matt__rose将近 11 年前

Let me see: I came across a bug in adobe flash (I know, evil) on linux one day where it came down to a change in memcpy in glibc.<a href="https://bugzilla.redhat.com/show_bug.cgi?id=638477" rel="nofollow">https://bugzilla.redhat.com/show_bug.cgi?id=638477</a>comment 129 <a href="https://bugzilla.redhat.com/show_bug.cgi?id=638477#c129" rel="nofollow">https://bugzilla.redhat.com/show_bug.cgi?id=638477#c129</a>---Quite frankly, I find your attitude to be annoying and downright stupid.How hard can it be to understand the following simple sentence:<pre><code> THE USER DOESN'T CARE. </code></pre> Pushing the blame around doesn't help anybody. The only thing that helps is Fedora being helpful, not being obstinate.Also, the fact is, that from a Q&A standpoint, a memcpy() that "just does the right thing" is simply _better_. Quoting standards is just stupid, when there's two simple choices: "it works" or "it doesn't work because bugs happen".Standards are paper. I use paper to wipe my butt every day. That's how much that paper is worth.Reality is what matters. When glibc changed memcpy, it created problems. Saying "not my problem" is irresponsible when it hurts users.And pointing fingers at Adobe and blaming them for creating bad software is _doubly_ irresponsible if you are then not willing to set a higher standard for your own project. And "not my problem" is not a higher standard.So please just fix it.The easy and technically nice solution is to just say "we'll alias memcpy to memmove - good software should never notice, and it helps bad software and a known problem".---This is why Linus is widely considered a good steward, and Ulrich had to be removed from glibc.Yes, they both had abrasive mannerisms, yes they sometimes said things that wouldn't pass fortune 500 HR policy.The difference is: Ulrich seemed to care more about some kind of technical "correct-ness", and anything that didn't fit in his mental model was considered wrong, and nothing else mattered.Linus deeply cares about the user experience. the kernel has a strict no-regressions policy for this reason. If it used to work before and now it doesn't, this needs to be fixed in the kernelLike this case, most of the cases where Linus uses salty language comes down to various kernel developers not following this policy, then complaining when their patch doesn't get accepted.Edited for formatting

chris_wot将近 11 年前

I suggest watching Torvalds' take on his honesty:<a href="http://youtu.be/-ZRvHbHxr-k" rel="nofollow">http://youtu.be/-ZRvHbHxr-k</a>

nodata将近 11 年前

Sounds like homework...

评论 #7955198 未加载

评论 #7954176 未加载

zerr将近 11 年前

Well, we can conclude that being an asshole is irrelevant or not enough for running a successful project.

评论 #7953993 未加载

评论 #7954316 未加载

general_failure将近 11 年前

People want to work under linus. Nobody wants to work under Ulrich.Case closed?