Few things that was tame for him. You must have a set of balls and be the ultimate narcissist to send code to Linus that does not even build.If you are stupid enough to try and ship code that does not build you should be perma banned from touching anything thing. Sounds like these people did not do a single test or had some ultra weird setup requiring vague modules,api to build right.
Linus Torvalds lashes devs who 'screw all the rules and processes' and send him 'crap'
Linux 4.11's first release candidate has been released, but not without a little friction after Linus Torvalds railed at the quality of some code sent his way during the merge process for the new update to the platform. Torvalds has a few gripes, writing that “if you cannot follow the simple merge window rules (this whole two- …
COMMENTS
-
-
Monday 6th March 2017 05:28 GMT Anonymous Coward
@kain
Assuming it actually happened.
I'm in no position to judge on that part but I do consider it a bit odd to read that someone wouldn't even test their own stuff. Making you all the more curious to who Torvalds is shouting at yet today this detail wasn't being shared. Even though he was previously all the more eager to directly address RedHat representatives and call them out for their horrid update. Yet now he choses to keep up anonymity?
I have no reason not to believe him, but it would have made more impact on me if he called out the people who actually did it.
In my opinion: seems some are good enough to be scolded at in public and some are special enough to be kept safe. Takes away the impact and reeks of double standards.
Of course: calling 'm out in public also gives the other party a platform to actually talk back.
-
Monday 6th March 2017 08:22 GMT Anonymous Coward
"If you are stupid enough to try and ship code that does not build"
Probably it does build somehow. On their systems exclusively. It looks Linux is undergoing the same issue Windows underwent some time in the past too - the arrival of lame developers. It happens to any platform as soon as it becomes widespread enough.
-
-
-
-
Monday 6th March 2017 11:13 GMT Doctor Syntax
Re: "If you are stupid enough to try and ship code that does not build"
"usually paid employees of the hardware maker."
And these are donated to Windows.
And if you look who contributes code to Linux it's mostly paid employees of H/W manufacturers or distro makers such as Red Hat.
Either both cases are charity or neither is. Is that beyond your understanding?
-
-
-
-
-
-
-
Monday 6th March 2017 02:59 GMT jonfr
Not surprise reaction from Linus
This is not a surprise from Linus, bugs do happen and due to complex Linux is something A not working with something E, G and D is hard to avoid at all. Lack of quality of code happens due to pressure and rushing things along, even if its just in rc builds.
Disclaimer: I'm not yet a programmer, but I've been looking into it as a next step in learning more about computers.
-
-
Monday 6th March 2017 10:33 GMT WatAWorld
Re: Grasshopper ...
"... allow me to impart some of the wisdom of the Ancient Coders: Do not begin to write code without first understanding full extent of the problem."
That is ideally true in the vast majority of cases, especially those with recently written in-house software at companies who've always had high quality staffing and procedures.
Then there are:
- The poor bastards stuck in shops where they're afraid to ask questions because they'll either be called stupid or ostracized as newbies.
- The poor bastards told to fix the current problem because it is urgent and that ramifications can be taken care of laster, and
- The poor bastards stuck in shops where most documentation was destroyed by people worried about their personal 'job security'.
More applicably to operating systems and massive shrink wrapped applications: There now there are systems installed in such a vast variety of companies all around the world, each using it for different purposes, different alphabets, each customizing it in their own way, each with different unforeseen needs, running on a vast variety of imperfect hardware, and then you'll realize that a detailed total understanding of the full extent of the problem is not always possible.
Understanding the full extent of the problem with mega complex massive multi-user software will come with sitting in your ashram chamber and accepting that your knowledge is not detailed that there will never be bug fixes, future releases and future versions.
You try to insist on get a sufficient understanding of the problem that you won't introduce bugs, while accepting that in such complex situations nothing is 100%.
(I think maybe that is how 'security researchers' see the world -- as a bunch of small shops with simple specs and simple human interactions, something that a human being can totally understand with a couple of months effort. That would explain why they think project scheduling, analysis, coding, unit testing, system integration testing, regression testing, beta testing and production roll-out, done over 15 countries with 45 different cultures and delivered to 137 countries in 55 languages should never take more than 90 days.)
-
Monday 6th March 2017 12:30 GMT Anonymous Coward
Re: Grasshopper ...
@WatAWorld
If the extend of the problem is unknowable, they should be clear on what problem they think they have solved, and what the gaps/risks are i.e. what they have not tried to solve.
I agree it is not always possible to understand everything when approaching a solution, but it is normally possible to frame your effort in what you are, and are not capable of achieving and what the constraints are.
-
-
-
-
Monday 6th March 2017 10:42 GMT Doctor Syntax
Re: Not surprise reaction from Linus
"Lack of quality code comes from getting your code second hand, built by people paid to fulfill some paying companies own project"
The biggest single contributor is usually Intel. Their "project" is the Intel processor line. Of course, given some recent experiences you may have a point....
https://www.theregister.co.uk/2017/03/03/netgear_recalling_hardware_with_bad_intel_atoms/
-
-
-
-
Monday 6th March 2017 07:06 GMT 9Rune5
Arguably, the amateurs who assembled your computer should have made sure that all the components came from respectable vendors supplying drivers for all major operating systems.
And the amateur buying said computer should have verified that the vendor was in the business of selling proper computers and not mere toys.
Good device drivers is mainly the hardware oems' responsibility. Many Windows installations have gone the way of the dodo due to dodgy third-party drivers. (MS writes very few drivers themselves)
-
-
-
Monday 6th March 2017 10:39 GMT Anonymous Coward
Re: "supplying drivers for all major operating systems"
Is Chrome Linux? Can you install and run any Linux software on Chrome? Google will of course write drivers for the Chrome hardware, getting all the needed info from the OEM, signing all the required NDAs, and maybe won't publish them either. Android too is not Linux, is a Google proprietary OS built on top of the Linux kernel.... just like macOS is not FreeBSD. There's a reason why the macOS driver framework is not the FreeBSD one...
-
-
-
-
Monday 6th March 2017 10:53 GMT Doctor Syntax
If and only if you are normally paid for your work are you a professional.
If you work for free you're an amateur.
Could you explain a little further.
Are the devs working for Intel being paid or not? Assuming they are that makes them professionals. Intel donates their work to the Linux kernel for free. Does that make them amateurs?
-
Monday 6th March 2017 11:20 GMT WatAWorld
"Are the devs working for Intel being paid or not? Assuming they are that makes them professionals. Intel donates their work to the Linux kernel for free. Does that make them amateurs?"
As covered in one of my other posts in this topic: The amature contributions to Linux and then the professional contribution to Linux.
The amateurs work for free, for experience or as part of class work.
The Intel, AMD, nVidia, etc developers are professionals paid by their employers. So are the developers working for banks, governments, consulting firms, and other companies that donate code to Torvalds.
Those professionals are employed by those other companies, not Linux.
With some exceptions, the objectives of the projects they are working on are not to improve Linux, but rather to get THEIR product to work with Linux, or to change Linux so it will play nice with THEIR internal app, or to eliminate some bug in Linux that affects their company and their clients.
- Their jobs are to fulfill the needs of their employer and their employer's clients.
- Their loyalty is to their employer and their employer's clients.
- They surely don't want to hurt Linux. They probably all want to play nice, because of professionalism and so their companies aren't banned.
- But their loyalties and objectives are not to make sure that Linux runs fine and bug free on obscure stuff at other companies. When their boss assigns them to a new project, regression testing of the old project ends.
So this stuff is created so Linux will work with their equipment or their internal applications, and it is donated to Linux for free. It is written by them so their stuff will work. And they're donating it to Torvalds for free. Even if it has bugs, it is arguably worth the price Torvalds pays for it.
Is Torvalds not even giving them a registered charity donation receipt so they can claim their effort on their income taxes?
-
Monday 6th March 2017 11:34 GMT Doctor Syntax
"With some exceptions, the objectives of the projects they are working on are not to improve Linux, but rather to get THEIR product to work with Linux, or to change Linux so it will play nice with THEIR internal app, or to eliminate some bug in Linux that affects their company and their clients."
Upvoted because this is actually an interesting comment.
The point at which I think we diverge is this: you're seeing Linux as something separate from these contributors (and here I'm including the employers such as Intel as contributors), as some external product which they have to improve.
I think you're looking at it wrong. Linux is, collectively, their product. Torvalds could step away from it today and people (including those companies) would still go on contributing because they see it as worth their while to have that product there and continuing to evolve.
One other interesting aspect of all this discussion is the role of the companies who contribute vs those who don't. If some H/W manufacturers such as Intel find it essential to contribute to Linux by sharing IP what, really, does it say about the self-image of those who seem to think it essential not to? Do they really think they know something Intel doesn't? Or do they not know something Intel does? Or, as someone said in another comment, are they ashamed of the quality of their code?
-
-
-
This post has been deleted by its author
-
-
-
-
-
Monday 6th March 2017 08:54 GMT AndyS
Re: "Does the chip vendor publish enough to let someone write a driver?"
> And why should they? It's their IP, not yours
You're an idiot, right? The man was complaining that his wifi doesn't work on Linux. The response is that, the vast majority of time there are missing drivers, it's because the manufacturer hasn't published specs, making it impossible to write drivers. And you ask "why should they?"
So, are you on the side of the guy who wants his computer to work with Linux? Or the side of the people saying it's difficult to do without manufacturer support? At the moment it seems like you're just... stupid.
-
Monday 6th March 2017 10:29 GMT Anonymous Coward
Re: "Does the chip vendor publish enough to let someone write a driver?"
No, you're an idiot - because you like to offend people who have a different opinion than yours - or better, Stallman ideology you all worship.
Yes, the OP was complaining about the not working wifi, and the answer was complaining of the lack of published specs to write open source drivers. So my answer stands: why they should publish their IP for free? Because of course even binary drivers are evil in Stallman "paradise", right? So, who looks stupid? Maybe those who prefer not to have drivers because they're not open code based on open specs?
MS is able to get companies write drivers for Windows. Ask yourself why Linux is unable to obtain the same. Maybe less ideology and better collaboration would help? Even if Stallman moooooohs?
-
Monday 6th March 2017 10:46 GMT Doctor Syntax
Re: "Does the chip vendor publish enough to let someone write a driver?"
"So my answer stands: why they should publish their IP for free? "
And so does mine. Drivers exist to sell hardware, not vice versa.
Unless, of course, you think they should go a step further and not publish Windows drivers either. That would really keep their IP under wraps, wouldn't it?
-
Monday 6th March 2017 10:56 GMT WatAWorld
Re: "Does the chip vendor publish enough to let someone write a driver?"
" why they should publish their IP for free? Because of course even binary drivers are evil in Stallman "paradise", right? So, who looks stupid? Maybe those who prefer not to have drivers because they're not open code based on open specs?"
I don't think the specs are written in open code, I think they're written in English with graphics.
Are you thinking of example code that you can clone?
Marketability is the jurisdiction of the sales department. That the product is marketable requires usability. The sales department at the device manufacturer should ensure that either drivers are published or the specifications for drivers are published, for all the important operating systems the device will be marketed for.
And sale departments have their limits:
1. They can't override human resources and force employees to work with abusive contacts.
2. They often won't care about salability to tiny markets.
Inability to work with others is a big problem in Linux.
I mean, look at all the problems with Windows and Apple (the Apple problems are hidden from the public, but you're professionals so you know).
All those problems and Linux can't give away its product when the alternatives both cost money and are so bad.
And it isn't only home users who avoid Linux. It is the professionals at banks, governments, electronics manufacturers, engineering companies. They use Linux, but sparingly, only when they have to.
Ask yourself, why do those in the industry with so much varied experience in so many industries not go with Linux all the time? Is everyone else stupid and you're the only smart person? Or are your needs different than theirs? Or do you not understand your needs?
-
Monday 6th March 2017 11:35 GMT nijam
Re: "Does the chip vendor publish enough to let someone write a driver?"
> MS is able to get companies write drivers for Windows.
Sometimes they pay those companies, if the device is one that MS sees a strategic or market significance in. Sometimes the company writes a driver because ... oh, you know, the device is useless without it. And there you have it - a device with no Linux driver is useless to anyone who is using Linux.
-
Tuesday 7th March 2017 08:45 GMT jbuk1
Re: "Does the chip vendor publish enough to let someone write a driver?"
"No, you're an idiot - because you like to offend people who have a different opinion than yours - or better, Stallman ideology you all worship.
Yes, the OP was complaining about the not working wifi, and the answer was complaining of the lack of published specs to write open source drivers. So my answer stands: why they should publish their IP for free? Because of course even binary drivers are evil in Stallman "paradise", right? So, who looks stupid? Maybe those who prefer not to have drivers because they're not open code based on open specs?"
No, it's definitely still you who looks stupid.
-
-
-
Monday 6th March 2017 08:56 GMT Paul Crawford
Re: "Does the chip vendor publish enough to let someone write a driver?"
"And why should they? It's their IP, not yours."
To make it work?
The IP is in the chip, not in the API. Unless of course its a bug-riddled pile of sh*t that has many workarounds in the driver code and they don't what that available without a NDA?
-
Monday 6th March 2017 10:35 GMT Anonymous Coward
Re: "Does the chip vendor publish enough to let someone write a driver?"
The APIs may tell a lot about the architecture of the chip itself, and you may need to know a lot of the chip architecture to use the low-level APIs proficiently. Thus, a company that has invested not a little sum of money to design that chip, may not want to give that IP away for free.
I know, it's always far easy to be munificent with someone else money.... and in my experience, those same people are the most greed when it comes to their own IP - if ever they have something valuable.
-
Monday 6th March 2017 11:18 GMT Doctor Syntax
Re: "Does the chip vendor publish enough to let someone write a driver?"
"I know, it's always far easy to be munificent with someone else money.... and in my experience, those same people are the most greed when it comes to their own IP - if ever they have something valuable."
Who are those "same people"? The ones who donate their IP to projects such as the Linux kernel, the Gnu toolset that surrounds it, KDE, Gnome, LibreOffice...?
-
Monday 6th March 2017 11:37 GMT nijam
Re: "Does the chip vendor publish enough to let someone write a driver?"
> The APIs may tell a lot about the architecture of the chip itself, and you may need to know a lot of the chip architecture to use the low-level APIs proficiently.
That is just (very) bad design. If the device does not have a clearly-specified function that can be encapsulated as a clean API, the implementation is probably just as shit inside as the API looks from the outside.
-
Tuesday 7th March 2017 14:37 GMT BinkyTheMagicPaperclip
Re: "Does the chip vendor publish enough to let someone write a driver?"
It is not poor design - have you ever looked at chipset specifications? It's frequently of the form 'send this request, wait at least this time, then send this. Don't send this when this other thing has been sent. This can be used only up to a maximum of n. etc.' This all helps reveal what the hardware is capable of.
There are many devices that can be permanently broken if incorrectly programmed. This is not rare.
Also, it does not matter. What matters is if it works. There are plenty of cases where a supposedly technically superior solution has failed against an architecturally inferior but better implemented solution. Obviously this may be an issue if the way the hardware works maps more closely on to one operating system than another..
-
-
-
-
Monday 6th March 2017 09:06 GMT Adam 52
Re: "Does the chip vendor publish enough to let someone write a driver?"
"Good device drivers is mainly the hardware oems' responsibility"
"Does the chip vendor publish enough to let someone write a driver?"
"if they supply drivers as binary ones, the GNUs will moooooooohhhh"
It seems to me that these are, if not mutually inconsistent, then at least confusing. In Windows land if you don't supply a driver, or fit with one of Microsoft's, then your product won't sell.
With Linux someone might produce a half-baked driver that's just about good enough by reverse engineering. Even if you do make a driver some distributions will reject it. So Linux ends up with inconsistent and poor quality drivers leading to a poor quality user experience unless all device manufacturers give away their code for free and take the reputation/support hit risk of Joe random developer breaking it.
You can see why hardware manufacturers don't want to take the risk.
-
-
Monday 6th March 2017 10:45 GMT Anonymous Coward
"Because what they sell isn't the driver, it's the device it drives"
Sure - but they probably make a decision about how much they could earn from Linux users, and how much will cost them to make their IP public - and any code under a dangerous license like the GPL. The result is often it's better to leave the few Linux desktop users without drivers.
Even server devices may have proprietary drivers available for distros like RedHat (which care far less about the "purity" of the drivers), and not Debian and its purists.
-
Monday 6th March 2017 11:04 GMT Doctor Syntax
Re: "Because what they sell isn't the driver, it's the device it drives"
"not Debian and its purists."
Searches package descriptions for "firmware". Finds 116 packages.
OK, some of them aren't firmware themselves, such as b43-fwcutter, Utility for extracting Broadcom 43xx firmware, but others are, some explicitly non-free.
All told I've got 14 of those packages installed, several of which provide binaries. Do I care they contain binaries? No.
-
-
-
Monday 6th March 2017 13:02 GMT dajames
Re: "Does the chip vendor publish enough to let someone write a driver?"
And why should they? It's their IP, not yours.
The clue is in the word "vendor". The chipmaker wants to sell the chips they make ... which they will find easier to do if they publish the interface specs so that people can actually use them -- that's everybody, not just those who can make use of any binary drivers the chipmaker may deign to provide.
The IP can be protected by copyright (so others can't clone the chip) but still published so that people can write software that makes use of it,
-
-
-
Monday 6th March 2017 10:29 GMT Anonymous Coward
Hstubbe: maybe you can share the name of the manufacturer of your laptop to see if we can grill them long enough to provide anything useful for getting your wi-fi going.
In the mean time, reading through your other postings, what happened to the idea of buying a Mac and running VMs in there to end your troubles?
-
Tuesday 7th March 2017 11:40 GMT Lars
Last time my wifi didn't work the problem was between the chair and the keyboard as I was too lazy to put my glasses on and got that "wifi code" wrong. Could it be that you have misunderstood the "out-of-the-box" regarding wifi. Just trying to help and as for those who talk about a wifi card a five year old laptop should have an inbuilt wifi and could be broken and does that wifi of yours work with some other device and how old is your Linux distro, perhaps you need some help.
-
-
-
Monday 6th March 2017 10:32 GMT Doctor Syntax
Re: NEWSFLASH!
"Man known for shouty & sweary outbursts....has a shouty & sweary outburst."
You really should have read the article before commenting.
Shouting? There's scarcely an expletive to report in Torvalds' post, which is rather tame compared to past missives.
Even better, read the announcement the article links: http://lkml.iu.edu/hypermail/linux/kernel/1703.0/03031.html
Go on, read it now.
What was that you were saying?
-
-
-
Monday 6th March 2017 10:07 GMT Robert Sneddon
Testing Testing Testing
"How can anyone working towards something as important as the Linux kernel not do at least at much?"
Nobody has created comprehensive code testing processes for Linux. Nobody is continuously improving those testing processes. Nobody is working to ensure all code submitted gets tested to a high standard. Nobody is paying them to test their own code. Nobody is standing over them ordering to test their code. Nobody gets their code and tests it independently. Nobody reviews their code. Nobody provides the sort of expensive hardware systems to run exhaustive tests on their code to ensure a high level of hardware compatibility over multiple platforms (not just two or three).
Linux is Open Source written in the main by folks who are doing it as a hobby. I have a vision of Mr. Torvalds standing up on a stage screaming "Testers! Testers! Testers!" while throwing chairs.
-
-
Monday 6th March 2017 10:55 GMT Anonymous Coward
Re: Testing Testing Testing
Oooh, it will be a hard discovery for all those anarchist basement Linux users who truly believe they are fighting "the system" and the evil megacorps using and promoting Linux... believing Linux is written by anointed druids of open source, wearing white robes, living of just pure water, happy to share for nothing, and never touched by the evil of money....
The megacorps say "thank you"....
-
Monday 6th March 2017 12:16 GMT LionelB
Re: Testing Testing Testing
"Oooh, it will be a hard discovery for all those anarchist basement Linux users..."
Eh? Never personally met any Linux users like that. Most, like myself, use it at work because it does the job better than the alternatives, and at home because it fulfils their home computing needs better than the alternatives.
-
Monday 6th March 2017 18:07 GMT Orv
Re: Testing Testing Testing
"...believing Linux is written by anointed druids of open source, wearing white robes, living of just pure water, happy to share for nothing, and never touched by the evil of money...."
I think you're confusing Linux with OpenBSD. The politics over in that camp make Linux's look tame.
-
Tuesday 7th March 2017 14:48 GMT BinkyTheMagicPaperclip
Re: Testing Testing Testing
OpenBSD has a very clear philosophy, a fairly unforgiving community, and a willingness to plough their own furrow. The size of the developer base affects things, too.
It may at times be spectacularly inconvenient not to be able to use any firewire or bluetooth devices, or any Nvidia card newer than 2009, but that's because they're sticking to their principles, or there are insufficient people available to engineer a suitable solution.
-
-
-
-
Monday 6th March 2017 15:50 GMT Chemist
Re: No excuse
"I'm getting a Raspberry Pi soon so I can also test on that."
As a little footnote to all the people claiming as usual that LInux is flaky, difficult, cli driven - I've just set a Pi up as a temp. measure after my wife's laptop died after she spilt water all over it . She wanted access to the laserprinter which lives on my file-server. So I installed the GUI printer controller from the (GUI) synaptic software installer and when I clicked on it it had found the printer and was all set to go. That was even easier than on the laptop (OpenSUSE 13.2)
-
-
Monday 6th March 2017 18:15 GMT Chemist
Re: No excuse
"Have you managed to get stable WiFi on your Pi? I have a couple old A-model ones here and I have yet to find a USB WiFi adapter that will stay up for more than a few days at a time."
Well most are Pi3s with built-in WiFi which have no problems but 1 Pi2 has a USB Ralink Technology, Corp. RT5370 Wireless Adapter which as been running continuously for ~ 150 days without a prob.
(Just had one of those phone calls, funny accent from "Windows Repair Centre". "Please press Windows key and .... well I wasn't listening really just grunting and saying 'Yeh' etc. Described the desktop - which was mostly blank - he eventually twigged it wasn't a Windows machine so accused me of lying about the Windows key 'because you have a Mac'. Needless to say it was the PI I'd setup for my wife. Well it amused my and wasted 10 minutes of his time...)
-
Tuesday 7th March 2017 10:06 GMT Tom 7
Re: No excuse Orv
Have you tried sudo apt-get dist-upgrade? There were a few problems with the earlier sw ( or rather I have some wifi problems but they went away with software updates. Just got a Pi-ZeroW and there are no problems at all with that - after two hours that is. But for ~£20, a TV and leads and keyboards from around the place I'd challenge anyone to find a better value for money piece of hardware anywhere.
-
-
-
-
Monday 6th March 2017 10:01 GMT WatAWorld
Charity case Torvalds gets what he he pays for.
If Torvalds were to treat his workers like human beings and pay them like professionals and he'd attract better quality developers.
Instead he's mostly got amateurs working for the glory of being associated with their spiritual leader / demi-god.
Aside for the amateurs, he's got donations from other companies, from people who work for the good of other companies, people who work to fulfill the needs of other companies, who are loyal to those other companies, and who follow the procedures, rules and leaders of those other companies.
Torvalds, you're getting what you pay for.
Be grateful for the charity. (And do you even hand out charitable donation receipts so people and companies can claim their donations to you on their income tax?)
-
Monday 6th March 2017 10:25 GMT Doctor Syntax
Re: Charity case Torvalds gets what he he pays for.
http://www.zdnet.com/article/whos-writing-linux-today/
"Leading the way were Intel employees with 10.5 percent of Linux code to their credit. Following Intel was Red Hat, 8.4 percent; Linaro, 5.6 percent; Samsung, 4.4 percent; IBM 3.2 percent; and SUSE, 3 percent."
Maybe it's Intel & co who are the charity cases. Torvalds is managing their product development for them.
-
Monday 6th March 2017 12:48 GMT Doctor Syntax
Re: Charity case Torvalds gets what he he pays for.
"Maybe it's Intel & co who are the charity cases. Torvalds is managing their product development for them."
OK, all this rhetoric is fun but it's obscuring the real situation. Let's get down to reality.
For most of its existence Linux has been supported by the Linux Foundation. The Foundation supports and manages kernel.org which is where the work is done. That, AIUI, includes paying Torvalds' salary.
The Foundation is supported by those businesses I included as "Intel & co" - and, yes, they include Microsoft. So Linux is, and has been for years, their joint product, collectively developed. And, we have to suppose, developed for what they think they can get out of it.
Organisationally it's an interesting situation. Collectively they contribute to pay Linus and separately they pay most of the individual developers. Yet Linus has no place the the line management of those developers, can't hire or fire them, doesn't perform their annual reviews or otherwise discipline them and can't even dictate what they work on. The only influence he has is the content and tone of feedback. I'm not sure what most of us would do in that position.
-
-
Monday 6th March 2017 16:29 GMT HieronymusBloggs
Re: Charity case Torvalds gets what he he pays for.
"If Torvalds were to treat his workers like human beings and pay them like professionals and he'd attract better quality developers."
You're sounding like a stuck record. I'd venture to suggest that reading some of the links you've been pointed to might unstick it.
-
-
Monday 6th March 2017 10:11 GMT kain preacher
WiFI seems to be hit or miss. I've seen intel wifi do weird things even under windows that required and update that was buried deep in the web site. But it's not limited to just wi fi. I had a gigabyte mother board with an on board NIC that hated Ubuntu. It would disconnect and reconnect to the network every 30 seconds. Worked fine on windows 7 and PCBSD. The real problem is these OEM folks that write drivers that are just good enough for windows and all other OS be screwed. Hell the drivers are flaky on windows and you want other OS support. I'm looking at you hapauge . Your USB products are shit.
-
Monday 6th March 2017 15:52 GMT Down not across
I had a gigabyte mother board with an on board NIC that hated Ubuntu. It would disconnect and reconnect to the network every 30 seconds.
Sounds like what the usual Realtek crap ones do if you actually try to push any real traffic through them. I've given up on them and always install intel NIC. They just work. The Realtek is fine for emergency ssh, bit of SNMP,etc but trying any heavier traffic locks (at least some of) the RTL NICs up.
Not to mention dual/quad Intel NICs are dirt cheap on tat bazaar so it is not worth fighting with the Realteks.
-
Monday 6th March 2017 20:08 GMT kain preacher
It was a low cost broadcom. Back then broadcom was having weird issue. You might have to broadcom chips that look identical but the issue was they had different drivers. There were also different version of the drivers depending on the version of the chip.
I wound up going to a electronic scrap yard and score a dual intel nic that was also a TCP accelerator.
-
-
Wednesday 8th March 2017 21:25 GMT Ken Hagan
"WiFI seems to be hit or miss. [...] But it's not limited to just wi fi."
WiFi is particularly prone to flakiness, however, because it depends on the quality of the WiFi at the other end and that's usually some "new" (at least when the router was built) chip, chosen for cheapness and driven by equally new (version 1.0) driver code and never updated since.
-
-
Monday 6th March 2017 23:43 GMT Anonymous Coward
You have to be careful picking hardware for Linux. I bought a cheap USB 3.0 card that didn't have drivers. I found another cheap with a good chipset, and no problems. Video cards might be a problem, but I suspect Linux drivers don't get such frequent updates as for Windows: most of the changelists I have seen seem to be focused on settings for specific Windows 10 games. The version numbers may also be imperfect for comparisons. Last year, a Windows user told me the version I was using was a beta, It was the code then being released with Ubuntu.
There don't seem to be any big problems with video drivers. NVIDIA release fairly up-to-date Linux drivers, the latest is v375.39, released last month for Linux, while the Windows drivers have reached v378.xx.
I'm not much into games, though Kerbal Space Program works fine. I'd say if you're looking for a specific game you might have a reason to choose Windows. I'd worry more about Amazon's Kindle reader. What do you worry about?