We both know for fact that kernel updates break binary only drivers. Now what kind of "hard facts" are you asking me to show? My "anecdotal evidence" that this really happens in the real world apparently is not enough.
Let me state that at the moment Linux is the best OS there is. I like it. That is why I have put it to friends machines.
But I am open to the problems it has and I would like to see them fixed. I am not going to pretend there are no problems. I am not going to defend it for the faults it has. I am not going to fight over semantics, syntax, "proving", or like.
You can consider yourself a "winner of this fight" as I am not going to continue it nor put more "evidence".
See, you do sound like a fanboy - to me it is clear there cannot be any evidence which would make you admit that there are FOSS drivers which do break in minor-minor updates and it is a big PITA.
Every out of the kernel tree FOSS driver dies on every kernel update.
TV cards, wifi, webcams, you name it. Once one ethernet driver was a such and the machine was used by a person incapable of recompiling the driver - and as you can deduce there is no way I can do it remotely.
Therefore on every installation I have done my advice has been: "do not install *any* updates, ever". Hardly the best way to deal with security updates? Unfortunately the only way!
I hope you can "help me out" and kill this utterly ridiculous concept of purposefully killing binary compatibility on minor-minor updates. I do not care about major, I can handle once a year, but almost once a month[1] is just too often.
[1] Approximately the number of kernel updates in Ubuntu
The problem with the "design" is that the design is Just Bad. Not only does it kill proprietary drivers it more often kills FOSS drivers. Besides most users do not give a shit whether some driver is FOSS or not.
They just want their machine to work. And especially, they do not want their machine to die on every kernel update. I know people who do not install even security patches anymore.
There are numerous examples of user space compatibility issues caused by kernel changes, e.g. USB, audio and some wifi cards.
2. The technical support is sometimes much worse. "Paying someone to fix" is not a good option.
3. Once in a while the feature X is a deal breaker. Really.
4. For example there is no stable Linux kernel, hasn't been for a while and apparently won't be, ever. The kernel today and kernel year ago are not compatible.
5. This is probably bullshit, but impossible to prove so. Especially as lately Firefox has had same number of exploits as IE.
6. "You are incompetent" is not very good selling point.
7. There are huge amount of software for which this is true. Then, again, there are software (e.g. OOo) for which it is likely not.
Weekly reports are no problem. Quite contrary, we want to know how others progress, what problems pop-up. We want to know deadlines.
Sure, if we could do whatever we like there would be champagne and strawberries... but seriously, although there are people who will try to do as little reporting as possible most/smartest are not that stubborn/stupid.
If there are no status reports (whether written or oral in a group meeting) how the hell are the OTHER programmers going to know what you are doing or how far have you progressed?
Besides, the messages can be sent by a nurse who can then read the replies - no need for the doctor to hold the phone, he might have something else to do with his hands...
Why do you need a privilege escalation? Why does it need to be remote?
Most of current Windows viruses are "britneynude.exe" type e-mails. Sure, Windows usually run as admin, but the virus really does not need any admin rights, though admin rights makes things like hiding ("rootkits") easier. But being admin is by no means necessary for sending spam, snooping keyboard for passwords, etc.
You are right, it is not up to me. But it certainly is up to the ISP.
I cannot see why ISP should (or would) not limit high users more than low users. Therefore those who use it for VoIP should not even notice if there is congestion, after all VoIP takes very little bandwidth.
Your buffet analogy is good - I am certain the owner does think those who eat little and fast are better customers than those who eat several hours and a lot. As I am certain ISP thinks about those who turn on their ADSL routers only rarely and use it very little.
It is interesting to notice the "capitalism" point, because I personally do not fully believe in it. That could (therefore would) lead to the situation where one VoD company would get good bandwidth (on some ISP) and the others would get shitty.
Slashdot Mirror
Linux is installed in new machines, which makes the rest of your rambling complete bullshit.
Besides I really do not care if Linux is not the best OS for the rest of the world, it is for me.
We both know for fact that kernel updates break binary only drivers. Now what kind of "hard facts" are you asking me to show? My "anecdotal evidence" that this really happens in the real world apparently is not enough.
Let me state that at the moment Linux is the best OS there is. I like it. That is why I have put it to friends machines.
But I am open to the problems it has and I would like to see them fixed. I am not going to pretend there are no problems. I am not going to defend it for the faults it has. I am not going to fight over semantics, syntax, "proving", or like.
You can consider yourself a "winner of this fight" as I am not going to continue it nor put more "evidence".
See, you do sound like a fanboy - to me it is clear there cannot be any evidence which would make you admit that there are FOSS drivers which do break in minor-minor updates and it is a big PITA.
Every out of the kernel tree FOSS driver dies on every kernel update.
TV cards, wifi, webcams, you name it. Once one ethernet driver was a such and the machine was used by a person incapable of recompiling the driver - and as you can deduce there is no way I can do it remotely.
Therefore on every installation I have done my advice has been: "do not install *any* updates, ever". Hardly the best way to deal with security updates? Unfortunately the only way!
I hope you can "help me out" and kill this utterly ridiculous concept of purposefully killing binary compatibility on minor-minor updates. I do not care about major, I can handle once a year, but almost once a month[1] is just too often.
[1] Approximately the number of kernel updates in Ubuntu
There are a lot of out-of-kernel FOSS drivers there.
However, the kernel.org gatekeepers *do* care about FOSS drivers.
No, they do not! And I have already proven it: huge amount of FOSS drivers die in every minor-minor (e.g. security) kernel update. Unnecessarily.
I know the theory behind the policy. Unfortunately it works in practice exactly shoot-the-foot way.
Not having to worry too much about back-compat makes their job easier.
Sure. And it makes my job much harder.
But after this you'll never claim the kernel to be "stable", will you?
Kernel security updates? Really?
P.S. I know you cannot.
If he tested all 3 OSes on the exact same hardware configuration and compared those results, then the tests results are valid.
Kindly disagree. There are billion different ways to make or tune benchmarks so that your favourite (OS, language, whatever) will look better.
The results shown are 100% meaningless.
Ah! Forgot, sorry!
I cannot consider a kernel to be "stable" in any sense if you cannot install security patches without the system dying.
The problem with the "design" is that the design is Just Bad. Not only does it kill proprietary drivers it more often kills FOSS drivers. Besides most users do not give a shit whether some driver is FOSS or not.
They just want their machine to work. And especially, they do not want their machine to die on every kernel update. I know people who do not install even security patches anymore.
There are numerous examples of user space compatibility issues caused by kernel changes, e.g. USB, audio and some wifi cards.
Unfortunately they still are true to some extent.
1. The price is quite often immaterial.
2. The technical support is sometimes much worse. "Paying someone to fix" is not a good option.
3. Once in a while the feature X is a deal breaker. Really.
4. For example there is no stable Linux kernel, hasn't been for a while and apparently won't be, ever. The kernel today and kernel year ago are not compatible.
5. This is probably bullshit, but impossible to prove so. Especially as lately Firefox has had same number of exploits as IE.
6. "You are incompetent" is not very good selling point.
7. There are huge amount of software for which this is true. Then, again, there are software (e.g. OOo) for which it is likely not.
Not only that, no sane ftp server puts the execute bit on an uploaded file.
I think you can get a few bucks for the iPod.
Maybe even enough to buy a decent mp3 player!
Sure, Felt did the right thing.
Unfortunately there is nobody left to do it now.
http://danielwebb.us/software/vnc/vncviewer.html
No we don't.
We resist silly reports, insane schedules, etc.
Weekly reports are no problem. Quite contrary, we want to know how others progress, what problems pop-up. We want to know deadlines.
Sure, if we could do whatever we like there would be champagne and strawberries ... but seriously, although there are people who will try to do as little reporting as possible most/smartest are not that stubborn/stupid.
If there are no status reports (whether written or oral in a group meeting) how the hell are the OTHER programmers going to know what you are doing or how far have you progressed?
I am disappointed - no one mentioned Krazy Kat!
wouldnt' particularly recommend more modern and high-level languages on purpose: they hide too much of Linux
This is exactly why I would recommend something other than C/C++.
But then, the question itself was a flamebait.
it will have viruses
When, or rather if, that day arrives I have had at least twenty, yes 20, years of virus, anti-virus and worry free surfing.
And counting.
Besides, the messages can be sent by a nurse who can then read the replies - no need for the doctor to hold the phone, he might have something else to do with his hands ...
Last I checked Notes, Rose, etc. were not free.
You cannot argue closed source software does have some advantages: you can have bigger price on it.
Open source has its own advantages, it is cheaper to the customer, and the "customer is the king".
Why do you need a privilege escalation?
Why does it need to be remote?
Most of current Windows viruses are "britneynude.exe" type e-mails. Sure, Windows usually run as admin, but the virus really does not need any admin rights, though admin rights makes things like hiding ("rootkits") easier. But being admin is by no means necessary for sending spam, snooping keyboard for passwords, etc.
You are right, it is not up to me. But it certainly is up to the ISP.
I cannot see why ISP should (or would) not limit high users more than low users. Therefore those who use it for VoIP should not even notice if there is congestion, after all VoIP takes very little bandwidth.
Your buffet analogy is good - I am certain the owner does think those who eat little and fast are better customers than those who eat several hours and a lot. As I am certain ISP thinks about those who turn on their ADSL routers only rarely and use it very little.
It is interesting to notice the "capitalism" point, because I personally do not fully believe in it. That could (therefore would) lead to the situation where one VoD company would get good bandwidth (on some ISP) and the others would get shitty.
Why should your bits take priority over mine?
Emergency call. Actually any call, if your data is ftp, http, torrent, or similar bulk data.
Or if I pay more - this is capitalism.
And, perhaps, because I use less (average) bandwidth (i.e. am a better customer).
And yes, I want the utility company to limit it, as it is obvious you will not (use QoS).
Very strange, as getting a German model (apparently not similar to swiss) to be sent to Finland is a piece of cake.
No, I cannot help, sorry.
What???
www.preis.de, www.amazon.co.uk, ... you name it.
I live in northern europe and getting Linux version is as easy as getting XP version. And cheaper.
BTW, I think 901 is overpriced compared to 900, but it just me.