We now have MS's monthly patches evaluated, tested and rolled out in under a week
Apart from the question of the difference between evaluation and testing, how exactly do you do this? The descriptions are generally to vague to know what differences to test for, MS prefer to patch vulns with no available exploit code. What exactly do you test? Simply that it installs without errors?
I am not trolling, I'm just a Solaris guy, and have no idea what is happening when I patch my wife's Windows PC
There's no need to be petty just because somebody has pointed out that other platforms provide better solutions to low-level booting than your own preferred platform.
Personally, I use x86 daily as a laptop, but far prefer Sun's OBP as a boot environment. Being able to set "auto-boot? = true" (or false), "diag-level", set a list of boot devices to try, in order, and failthrough until one works, set the tty settings of the parallel port (or, these days, the Lights-Out-Management subsystem) for remote console access, and countless other features.
An earlier post from myself showed that I managed to get an x86 system (with a patched GRUB) to boot Windows/Linux/Solaris on demand, remotely, but I can't find a vendor which supports this, I used a patched GRUB. The x86 BIOS has no concept of control of the system, from low-level testing, network probing, SCSI probing, etc, etc.
Don't knock what you don't know.
Old kernel, new options (e.g.,/boot/vmlinuz-2.6.10 single)? Copy-n-paste the old title/kernel/initrd entries in c:\menu.lst, change the title, and add the new options.
Does this method allow for changing the boot options at boot time, or require rebooting and editing a file? LILO and GRUB allow for such mods at boot-time.
Most importantly for me, they allow for one-off changes (though the GRUB patch doesn't seem to have made its way to the mainstream distros - I had to patch grub myself to get a config working in the office which boots Windows, Solaris and Linux, in a way in which I can remotely reboot into my desired OS - Admittedly, I have to boot into Linux first (as the others can't read/write reiserfs AFAIK), but from home, I can boot a build server to Linux, build the software, reboot it into Solaris, build the Solaris package, reboot into Linux, reboot into Windows, build that package (courtesty of VNC), reboot back to Linux to get back to where I started from. That is useful.
POSIX has never been free (see an early Linus post asking for the POSIX standards... go use google) and is a STANDARD, not an IMPLEMENTATION.
Sorry for shouting, but it sometimes helps numbskulls
(1) correcting someone's erroneous statement that gcc was the only free compiler available for x86? It was not.
Name them. Note: they need to run on Minix, as that was what Torvalds had available to him (and, presumably, MSDOS)
Most of those examples would have also spread around from the original authors
Not without the excellent communication channels provided by the Roman empire.
"empires rised and falled throughout history, Rome simply produced the first major one across Europe"
Empires rose and fell, surely. "Rised and falled"?? Clearly we have an academic genious lecturing us here.
Rome simply produced the largest empire in human history, ever - not "the first major one across Europe" - Yes, sorry to tell you, even bigger than the current American empire, by any coherent measure.
As another poster has already said, lack of availability doesn't force anybody to "pirate" anything. However, there is some merit in the point, and it should be pointed out that the same dilemma exists for fans of movies and music which is no longer available; albums get deleted from record companies' catalogues all the time.
The options are to spend weeks going round second-hand record shops, go without, or get it illegally.
Most times, I just accept that I have to go without.
"The doctor, Philip Mach, had a license to practice medicine in New Jersey but he provided prescriptions to people throughout the United States without ever evaluating them, both of which are big no-no's"
I can see that providing scripts without "evaluating" them is a no-no, but apparently both of these things are "no-no's" - having a license to practice medicine in New Jersey is illegal??
I had a boss who was an ex-techie, moved into management, and he was actually really good. He understood that his role was to interface between the techies and the board. His knowledge wasn't quite up to date, but he understood what we said, and could translate it to non-techies. Similarly, he could translate non-techie stuff to us.
That is a good manager.
(From talking to other staff after I left, apparently he got too "into" the director stuff, and lost touch with the techies, and lost his credibility with it)
Where I am at the moment (and have been for many years), the Director is an ex-techie; he hired a "manager" to do the management stuff, but the "manager" knows neither IT nor how to manage. Both are money-grubbing bastards, focussed solely on how to make the most money, not at all interested in getting work done (which strikes me as the obvious way of making money!).
Between them, they've got no idea about how to manage staff, from the HR perspective in particular. Although they've been pretty cool about working hours (do the hours, prove it, don't have to do 9-5), I'm looking forward to going to a "proper" company in a month, with decent management.
It must be something to do with Unicode, and international keyboards. Possibly keyboard rates. Maybe the letter "o" shows a greater disposition towards repetition (maybe their browser interpets "o" as "0" and decides "oh, zero's are cheap, let's send two, that'll make them happy")
On second thoughts, no, I think most people are too damn stupid.
If a company hires people to sift through data garbage, it makes sense that there will be people in the news who will refuse to grant interviews with them.
But this is reporting on the person who hires people / machines to sift through garbage. That's Google's job, is to sift through all the garbage on the web and try to find useful info amongst it. It's turning the tables, and suddenly Schmidt doesn't like people using his service for the purpose it was designed for.
Bemused by any notion that Linux is poorly supported in the enterprise, Robertson says he has never had any trouble in finding support for De Bortoli's open-source systems, and says that
those IT managers having trouble simply aren't looking hard enough.
I think it's pretty clear that it is a direct quote from TFA.
I believe that as long as the code is GPL and derived from Linux code, he doesn't care what it ultimately ends up being called, as long as someone isn't producing something that ISN'T Linux code and calling it Linux.
The entire point of this is to prevent people from turning out something that is NOT LINUX CODE and calling it Linux.
No, it's not. It's about people taking *anything* and registering a trademark for it which includes the Linux trademark.
If I take Linux and distribute it as SteveOS, trademarked as SteveOS, that's fine, just as if I developed my own OS and called it SteveOS.
However, if I take Linux (or my own code) and trademark it as SteveLinux, that's using the Linux trademark.
It's not to do with the code, it's a simple thing about the name.
Kudos for reminding me of Archie, but the rest is crap.
It's not about Schmidt complaining about stuff stored on google.com (whose robots.txt he has control over), it's about C|Net using Google to find other stuff which has been reported about him by third parties.
What does he want? That I can search for "Britney Spears" and find information about her, but I can't search for "Eric Schmidt" and find information about him?
The app involved does not include GUI elements, which makes it easier; the Windows development cycle seems to concentrate on the GUI 90%, the code 10%. I'm not in that game.
2 months to test a small single-platform (i.e., Windows) product? Are you serious?
I work on a small cross-platform product which runs on Linux/Solaris (SPARC/x86)/Windows/MacOSX. Building the bundle takes well under an hour; testing (assuming we're talking about sane testing... what has changed => Test it... what should work => Test it) would be difficult to justify more than 3 days across all 5 platforms.
If it takes 2 months to test your product, your test regime has serious problems.
In fact, why bother with Linux kernel at all - there's nothing here relevant to Linux, it's just about "some kind of OS I would like to see" - let Linux get on with being Unixy and doing a good job at it.
Feel free to design a system (with or without a Linux kernel) which has your own bizarre filesystem config, because there's no way on earth that you'll get buy-in for some freakish filesystem layout for mainstream use.
Design your own distro with this config - feel free, you'll end up doing all your own packaging, and have even less developer support than Linspire get due to your weird and proprietary (not to mention fucking obscure and unusable) methods.
What is so evil about/bin,/usr, etc? How is this worse than "C:\Program Files\"?
As for the supposed fix for library support - everyone bundles every library with every application... there are still a *LOT* of people on dialup, and you totally gloss over the security implications.
So, let's review: An OS with random libraries installed all over the place, with no central control, no way of auditing the security, random config possibly based on some undefined "registry", I think you've just achieved the impossible - proposing a worse OS than Windows on Slashdot and actually getting it onto the front page!
I think he's complaining that "I don't understand the plain shell", which sounds, to me, on a Friday afternoon, pretty close to saying "I'm not capable of calling myself a sysadmin, so why am I even trying to go beyond my abilities?"
But maybe I'm being harsh,
They use sh standard for user accounts, csh for root.
That's flamebait! See Csh Considered Harmful. Solaris uses Bourne shell for root and user accounts, by default. Bash is installed if you install the full (SUNWCXall) package cluster.
Solaris scales to more than 64 CPUs - certainly 96, presumably higher in principle (though the biggest Sun server, AFAIK, goes to 96 CPUs)
Do USAians actually understand what poverty means? A huge number (I don't have the figures to hand) earn less than USD10 per month.
In fact, the Make Poverty History have a poster (which unfortunately does not appear to be online) quoting a statistic that a London (UK) parking meter earns more in an hour than something like 75% of the world's population earns in a month.
Slashdot Mirror
Apart from the question of the difference between evaluation and testing, how exactly do you do this? The descriptions are generally to vague to know what differences to test for, MS prefer to patch vulns with no available exploit code. What exactly do you test? Simply that it installs without errors?
I am not trolling, I'm just a Solaris guy, and have no idea what is happening when I patch my wife's Windows PC
They gave an MCSE an *office*?!!!
There's no need to be petty just because somebody has pointed out that other platforms provide better solutions to low-level booting than your own preferred platform.
Personally, I use x86 daily as a laptop, but far prefer Sun's OBP as a boot environment. Being able to set "auto-boot? = true" (or false), "diag-level", set a list of boot devices to try, in order, and failthrough until one works, set the tty settings of the parallel port (or, these days, the Lights-Out-Management subsystem) for remote console access, and countless other features.
An earlier post from myself showed that I managed to get an x86 system (with a patched GRUB) to boot Windows/Linux/Solaris on demand, remotely, but I can't find a vendor which supports this, I used a patched GRUB. The x86 BIOS has no concept of control of the system, from low-level testing, network probing, SCSI probing, etc, etc.
Don't knock what you don't know.
Does this method allow for changing the boot options at boot time, or require rebooting and editing a file? LILO and GRUB allow for such mods at boot-time.
Most importantly for me, they allow for one-off changes (though the GRUB patch doesn't seem to have made its way to the mainstream distros - I had to patch grub myself to get a config working in the office which boots Windows, Solaris and Linux, in a way in which I can remotely reboot into my desired OS - Admittedly, I have to boot into Linux first (as the others can't read/write reiserfs AFAIK), but from home, I can boot a build server to Linux, build the software, reboot it into Solaris, build the Solaris package, reboot into Linux, reboot into Windows, build that package (courtesty of VNC), reboot back to Linux to get back to where I started from. That is useful.
POSIX has never been free (see an early Linus post asking for the POSIX standards... go use google) and is a STANDARD, not an IMPLEMENTATION.
Sorry for shouting, but it sometimes helps numbskulls
(1) correcting someone's erroneous statement that gcc was the only free compiler available for x86? It was not.
Name them. Note: they need to run on Minix, as that was what Torvalds had available to him (and, presumably, MSDOS)
Not without the excellent communication channels provided by the Roman empire.
"empires rised and falled throughout history, Rome simply produced the first major one across Europe"
Empires rose and fell, surely. "Rised and falled"?? Clearly we have an academic genious lecturing us here.
Rome simply produced the largest empire in human history, ever - not "the first major one across Europe" - Yes, sorry to tell you, even bigger than the current American empire, by any coherent measure.
As another poster has already said, lack of availability doesn't force anybody to "pirate" anything. However, there is some merit in the point, and it should be pointed out that the same dilemma exists for fans of movies and music which is no longer available; albums get deleted from record companies' catalogues all the time.
The options are to spend weeks going round second-hand record shops, go without, or get it illegally.
Most times, I just accept that I have to go without.
I can see that providing scripts without "evaluating" them is a no-no, but apparently both of these things are "no-no's" - having a license to practice medicine in New Jersey is illegal??
Can I work for you?!!
I had a boss who was an ex-techie, moved into management, and he was actually really good. He understood that his role was to interface between the techies and the board. His knowledge wasn't quite up to date, but he understood what we said, and could translate it to non-techies. Similarly, he could translate non-techie stuff to us. That is a good manager. (From talking to other staff after I left, apparently he got too "into" the director stuff, and lost touch with the techies, and lost his credibility with it) Where I am at the moment (and have been for many years), the Director is an ex-techie; he hired a "manager" to do the management stuff, but the "manager" knows neither IT nor how to manage. Both are money-grubbing bastards, focussed solely on how to make the most money, not at all interested in getting work done (which strikes me as the obvious way of making money!). Between them, they've got no idea about how to manage staff, from the HR perspective in particular. Although they've been pretty cool about working hours (do the hours, prove it, don't have to do 9-5), I'm looking forward to going to a "proper" company in a month, with decent management.
It must be something to do with Unicode, and international keyboards. Possibly keyboard rates. Maybe the letter "o" shows a greater disposition towards repetition (maybe their browser interpets "o" as "0" and decides "oh, zero's are cheap, let's send two, that'll make them happy") On second thoughts, no, I think most people are too damn stupid.
But this is reporting on the person who hires people / machines to sift through garbage. That's Google's job, is to sift through all the garbage on the web and try to find useful info amongst it. It's turning the tables, and suddenly Schmidt doesn't like people using his service for the purpose it was designed for.
It's not about Schmidt complaining about stuff stored on google.com (whose robots.txt he has control over), it's about C|Net using Google to find other stuff which has been reported about him by third parties.
What does he want? That I can search for "Britney Spears" and find information about her, but I can't search for "Eric Schmidt" and find information about him?
Google have fucked up this time.
The app involved does not include GUI elements, which makes it easier; the Windows development cycle seems to concentrate on the GUI 90%, the code 10%. I'm not in that game.
I work on a small cross-platform product which runs on Linux/Solaris (SPARC/x86)/Windows/MacOSX. Building the bundle takes well under an hour; testing (assuming we're talking about sane testing... what has changed => Test it ... what should work => Test it) would be difficult to justify more than 3 days across all 5 platforms.
If it takes 2 months to test your product, your test regime has serious problems.
Are you sure?
The Q is "what do we all need to know", not "what do I need to know"
This is a 20-year retrograde step.
Why not look into ACLs, etc.
In fact, why bother with Linux kernel at all - there's nothing here relevant to Linux, it's just about "some kind of OS I would like to see" - let Linux get on with being Unixy and doing a good job at it. /bin, /usr, etc? How is this worse than "C:\Program Files\"?
Feel free to design a system (with or without a Linux kernel) which has your own bizarre filesystem config, because there's no way on earth that you'll get buy-in for some freakish filesystem layout for mainstream use.
Design your own distro with this config - feel free, you'll end up doing all your own packaging, and have even less developer support than Linspire get due to your weird and proprietary (not to mention fucking obscure and unusable) methods.
What is so evil about
As for the supposed fix for library support - everyone bundles every library with every application... there are still a *LOT* of people on dialup, and you totally gloss over the security implications.
So, let's review: An OS with random libraries installed all over the place, with no central control, no way of auditing the security, random config possibly based on some undefined "registry", I think you've just achieved the impossible - proposing a worse OS than Windows on Slashdot and actually getting it onto the front page!
I think he's complaining that "I don't understand the plain shell", which sounds, to me, on a Friday afternoon, pretty close to saying "I'm not capable of calling myself a sysadmin, so why am I even trying to go beyond my abilities?"
But maybe I'm being harsh,
They use sh standard for user accounts, csh for root.
That's flamebait! See Csh Considered Harmful. Solaris uses Bourne shell for root and user accounts, by default. Bash is installed if you install the full (SUNWCXall) package cluster.
Solaris scales to more than 64 CPUs - certainly 96, presumably higher in principle (though the biggest Sun server, AFAIK, goes to 96 CPUs)
I don't know where you get your "sex offender" data from, but are there really no sex offenders with white skin?
Or are they just not recorded?
Those so poor they can afford a $500 PC
Oh, my heart bleeds.
Do USAians actually understand what poverty means? A huge number (I don't have the figures to hand) earn less than USD10 per month.
In fact, the Make Poverty History have a poster (which unfortunately does not appear to be online) quoting a statistic that a London (UK) parking meter earns more in an hour than something like 75% of the world's population earns in a month.
Please, the http://www.makepovertyhistory.org/ campaign, put this stuff up on the web, not just on dead trees!
We do realise that the G8 summit is upon us, and that huge international protests against international poverty are due to coincide with it? ... Don't we?
Or is this just some sheltered young white well-to-do middle-class ... oh, just remembered where I am.
Go, Dubya!