I'd rather only have people who are mentally stable enough to withstand Linus's flames develop code for the Linux kernel.
To each their own. Personally, I'd rather have good Linux kernel developers develop code for the Linux kernel.
Also, your implication that anyone who doesn't want to work in an unpleasant environment is mentally unstable is obvious hyperbole.
This is not about a small experimental project where nobody cares about stability, but one of the largest truly collaborative endeavors ever made by humankind
And again, your hyperbole is unfounded. This is one of the largest collaborative endeavours ever made by humankind. This is one of the largest collaborative endeavours ever made by humankind. This is one of the largest collaborative endeavours ever made by humankind.
Linux is a relatively large software project. There are plenty of other relatively large software projects, written by plenty of other smart people who also care about stability and quality.
billions of dollars in worldwide economic growth hinge on its future development
In what way, exactly? Systems that include Linux surely have a collective value on that scale, but that value wouldn't disappear if someone made a small mistake in a kernel commit, or if Linus retired tomorrow.
To give you some perspective, I'm pretty sure that the astronaut program performs stricter tests for mental stability than being able to take some guy's rants not too personal
You are giving me some perspective, by comparing a programme that develops one relatively large piece of software to a programme that puts humans in space? Physician, heal thyself!
Oh, come on. Linus is not the boss of any of the people on LKML.
Are you suggesting that he's just an ordinary poster on the list, with no special authority compared to kernel contributors and no special status as a leader that newcomers might look to?
If you're a n00b who posts something stupid on LKML, you are not going to get massive old-school-Usenet-style flames.
I wonder how many n00bs never get that far, because they see how the leader of the community treats others and decide to go do something else instead. Maybe Linus does personally know the recipients of his infamous rants, but on a high-profile public forum not everyone watching might realise that.
If you walked into an office for a job interview, and the first thing you saw was some management type openly berating a subordinate, what tone would that set before you even started the discussion you were there for? Some conversations are best held privately, as much for the benefit of the community as for the participants themselves.
I have often seen this same "enforced politeness" tried on other mailing lists, and the result is always the same. The "wizards" soon migrate somewhere else
Then wish them well and send them on their way.
I find your implied association between smart people who get useful things done and rude people who can't act like adults unlikely. I know plenty of smart people, and the overwhelming majority of them would prefer to work with others in civilised fashion. Sure, when people are passionate about something then occasionally someone might cross the line, but then they apologise and everyone carries on.
I know plenty of blustery people as well, and a lot of them bluster to cover the fact that they aren't nearly as smart or valuable as they would like everyone else to believe. As with any bad apple, the best management decision for the project as a whole is usually to fire such people at the earliest opportunity rather than let them contaminate things any further or dig in any deeper.
Sometimes doing that will hurt in the short term, but no-one is irreplaceable. Once the bad ones are out of the way you can get on with bringing in other smart people to replace them. That can now include all the smart people you couldn't bring in before because they had no interest in working in a hostile environment and, being smart, they had plenty of other choices.
FWIW, I'm in the UK, in an area where the power supply is less than brilliant. We don't get many complete outages, but moderate surges and brown-outs seem to be happening all the time if the behaviour of a UPS is to be believed.
The working life of our electronic devices was surprisingly short across the board for the first few years after we moved here, with many formerly reliable devices all failing within a couple of years of the move, including (coincidentally or otherwise) multiple consumer-grade broadband routers. In contrast, in the years since installing a UPS for all the serious gear and at least basic surge protectors for everything else that plugs into a wall socket, we've seen almost no surprising failures of that kind.
Of course we don't know for sure whether it was really the dubious power supply that was responsible, and as other posters have mentioned there are several alternative explanations that would also make sense. But given how many things we saw fail within the window where the power supply was bad, and how few failed before and afterwards, the odds of the power supply being a factor seem quite high in our case.
Any attempt to raising a point about how you don't need to optimize everything but only few critical zones of your code (what matters)
To be fair, if you were debating with someone who writes applications that really do need the very top levels of performance, and you claimed that optimising trouble-spots would be sufficient to achieve that, then you were ignorant. For most software, being within a factor of 2 or 3 of hand-optimised low-level code is easily sufficient, and a bit of work with the profiler to identify the most serious flaws will go a long way. The rules change when you shift from that kind of performance standard to needing the very top levels, because then the emphasis on speed permeates everything.
Yes, that sounds unpleasantly familiar. It seems that these days a small business can't rely on any major vendor for a complete range of good quality, compatible gear.
Maybe the likes of Cisco and HP are worth it if you have 24/7 IT teams running dedicated servers room full of equipment and measure the cost of downtime per second. For those guys, the high-end gear and management facilities and expensive support contracts might justify the cost.
For the rest of us, it feels like the best strategy now is to build heterogeneous IT systems and networks. Look for recommendations of individual devices, often from smaller and more specialised manufacturers, that do one job well. Anecdotally, I've found that in recent years these boxes tend to be pretty good at supporting the major standards, so if you buy best-in-class for each device you need then compatibility doesn't seem to be a major issue as much as maybe it used to be. And if anything does go wrong or you do need help with some awkward configuration, you have more than a snowball's chance in hell of speaking to someone who can help without coughing up most of this month's revenues for a support contract.
Agreed, given the repeated failures here, the power supply might be less than wonderful.
It's also worth remembering that "enterprise" equipment is often more about the management features (which no home user is ever likely to need) than the hardware itself. Sometimes the low-end business gear actually turns out to be worse than decent consumer kit. For example, we bought a bunch of Cisco's small business branded equipment for a small office once, paying a premium for it but expecting that the quality and support would be better than some disappointing consumer grade equipment it was replacing. In fact, the NAS turned out to be a rebadged device from another vendor that Cisco never really supported properly, the wireless access point turned out to have buggy firmware that would just drop connections, and so on. It's a mistake we'll never make again.
Heuristics are bug ridden by definition. If they didn't have bugs, then they'd be algorithms.