This is not necessarily true. Laptop drivers for their power control and especially pointing devices are often painful to integrate into both Windows and Linux operating systems. By pre-bundling the OS, we know that Lenovo has worked out those issues in advance and included whatever modified drivers or software components are necessary.
I've run into these issues with high-end RAID controllers in the server world, and lots of us who've provided Linux systemns for our workplaces have run into it with NVidia or older video chipsets in the graphics world.
Not Windows, Vista. The support costs of Vista are so high, and it remains so fragile in customer environments, that SuSE may well be a fiscal benefit for Lenovo to provide instead, irrelevant of the greater retail cost of Vista.
This is certainly the case right now for Windows XP and Vista, as numerous laptop and desktop retailers have learned to their dismay.
That's exactly the sort of approach I'd recommend. One might, *might* add a read-only client or guest for permitting access from other local servers, or for developers to download and use the database locally, or for system monitoring.. One might also sanely add a master/master server in parallel, for failover operations.
I *meant* that the database of a public server, for security and performance reasons, should not be SQL. One might be forced to do so for contractual or "corporate policy" reasons, but it's generally a pretty bad idea for the sort of reason we're seeing right now.
The money on laptops is leverage money: it reduces the costs of books, printing documents, planning, and communicating with the students and their families. And you had better believe those laptops will be *family* resources for information access for the whole family, with information those families could otherwise find incredibly difficult to obtain about local weather, crop prices, available markets, jobs, and even literature. And it does it in a modular way, one that does require infrastructure but far less than a new set of books.
No, they're not. Both Intel and Microsoft have a long, documented history of stealing work from other companies, then using the stolen goods to drive the other company out of business. And they double teamed DEC, with Intel stealing the Alpha technologies and Microsoft stealing the VMS technologies. They've both got far more of a history of this than their competitors.
Welcome to the NVidia drivers, which cannot be pre-installed because of this and are installed exactly how you describe. Then their installer screws up your local system: they can't use a standard.deb. or.rpm installer because they insist on users manually signing their software license, and you can't automate that without violating their license agreements.
As soon as you provide closed source modules for the kernel, such as the NVidia modules and some the McAfee Linux components, you "taint" the kernel. That means you cannot pre-compile the kernel and distribute it with your models pre-installed, they "taint" the kernel. You have to provide your components separately.
This drives the closed source designers, like McAfee and NVidia, nuts, because they can't pre-include their tools. It's worse because most of their installers suck really, really hard, and actually imperil the systems they are run on even more than a typical security risk does. And it drives we open source developers nuts because people want to know why their tools don't run automatically or well, or why they're not already built in. And we have to waste our time fixing the mess caused by their components for which we have no authority, no control, and no access to address the failures.
Well, the "theft" of our work is a loaded phrase in legal terms and confuses the lines between moral and legal obligations. Please use the word "freeloader" or another term, because the use of open source tools to build closed source products is permitted by some licenses, such as BSD. This is precisely why I think the BSD license is a freeloader on the GPL license, which protects against exactly the freeloading you describe, and provides a lot more territory for the BSD closed source authors to operate.
Moreover, clear or not, it's been repeatedly tested in court, in lots of interesting ways. Take a look at http://boycottnovell.com/2007/11/23/gpl-court-test/ for references, and a Google search reveals far more links with good citations.
With cause. Their Linux operating system products certainly include kernel modules or customized kernels: the Linux kernel is GPLv2. This means that installing closed source, proprietary security modules "taints" the kernel and means the tainted components cannot be included as a built-in part of the kernel. This is vaguely similar to the problem NVidia has with their proprietary drivers and OpenGL libraries, and the absolutely awful procedures for installing such modules that various closed source vendors use. I assume that the problem also occurs with their other components, but I can directly report the kernel module issue.
The "oh, no, they'll see our code" concern is a basic security through obscurity claim, and is common to a lot of commercial security vendors. It's countered by the difficulty of testing and integrating such closed code into lighter weight, better featured or better performing tools.
We also need to keep an eye on these security companies and their desire to integrate DRM support that could certainly be useful for authenticating software installations and protecting user's files, but can easily be used to enforce "you can only read this file with this software" features that for business reasons, they will need to support as so-called "Trusted Computing" is supported in more CPU's and is deeply integrated into Vista and the next Microsoft release. They need to protect those features for obvious business reasons, and not allow the publication of open source code to access the DRM protected features. The GPLv3 is aimed directly at this problem, and McAfee surely is thinking about how to protect that market.
Oh, it happens in the Linux world too. There are packges that install new users and mis-set them in the open source world, or overwrite system files and make recovery a nightmare. NVidia drivers are a classic example, as are developers who randomly pull in PHP or Perl modules without packaging them and create awful dependencies that break other packages. Kernels, gcc, and Perl modules are classic examples.
Well, maybe. But it's a backup: you're supposed to be able to restore reliable from it, not have to hand massage it. Doing a "sync" operation before doing the suspend might help. If you're backing up something big, like a database, and someone executes operations right in the middle of it, then your files may wind up corrupted. And that metadata *itself* may be corrupted, depending on which backup technology you use. If you can, it might be worth locking a database temporarily and then running sync while you do the snapshot.
All in all, I'd rather be safe than sorry with backup data. And having a text dump of something like a MySQL database means that you can also restore to another version or another system, rather than relying on having the exact same version of MySQL.
Such a sense of humor is not polite to your staff, and can be avoided by simply calling the VMWare instance on your desktop//desktop-vmware, so htey know whom to go looking for.
And the kernel of that hypervisor is.... Linux 2.4. So its ability to deal with contemporary hardware relies on backporting the drivers from 2.6. This way lies a fascinating level of difficulty, and the need to run ESX on rather limited hardware.
They're incredibly useful for QA andn build environments, though. Being able to whip up a new OS from a base image in a matter of minutes is a pearl of great price, and helps eliminate the 3 hours of downloads and updataes for all sorts of awkward environments.
"suspend" is not enough for a database, like MySQL, or files that were being written and are still resident only in RAM. You need to actually do a shutdown, or make sure that sensitive databases were backed up. If you don't believe me about such issues, try it and then run "fsck" against the partitions of the virtual guest.
I've heard of such, but they tend to be very, very nasty, and prone to causing more service interruptions than they cause. Lightweight userland failover systems, such as creating MySQL databases in master/master configurations and using IP failover tools like wackamole seem to work more gracefully, and without imperiling your kernel and hardware support.
For lightweight systems, you can use rsync and database backup tools to store the database, then rsync the filesystems to a repository. For full systems, you can do a very fast shutdown of most virtual OS's, use LVM or another partition based snapshot technique, restore the virtual server, and backup the snapshot.
Theh combinations can get even more interesting, but you may begin to see the possibilities.
Which GPL? GPLv3 includes changes aimed very specifically at preventing another of these patent protection deals, and putting a spike in Microsoft attempts to encumber open source projects with patent agreements. The GPL needed to evolve to prevent this and some of the DRM craziness being attempted, and has. But will Novell follow suit by publishing its own software or changes under GPLv3?
Slashdot Mirror
This is not necessarily true. Laptop drivers for their power control and especially pointing devices are often painful to integrate into both Windows and Linux operating systems. By pre-bundling the OS, we know that Lenovo has worked out those issues in advance and included whatever modified drivers or software components are necessary. I've run into these issues with high-end RAID controllers in the server world, and lots of us who've provided Linux systemns for our workplaces have run into it with NVidia or older video chipsets in the graphics world.
Not Windows, Vista. The support costs of Vista are so high, and it remains so fragile in customer environments, that SuSE may well be a fiscal benefit for Lenovo to provide instead, irrelevant of the greater retail cost of Vista.
This is certainly the case right now for Windows XP and Vista, as numerous laptop and desktop retailers have learned to their dismay.
That's exactly the sort of approach I'd recommend. One might, *might* add a read-only client or guest for permitting access from other local servers, or for developers to download and use the database locally, or for system monitoring.. One might also sanely add a master/master server in parallel, for failover operations.
I *meant* that the database of a public server, for security and performance reasons, should not be SQL. One might be forced to do so for contractual or "corporate policy" reasons, but it's generally a pretty bad idea for the sort of reason we're seeing right now.
Given the authors, perhaps they could share their insight on how studying Ada is damaging to your employment prospects and your career?
That's right. Just make sure it has a GPL or possibly an Apache license, and your security status will improve quite a bit.
For performance and security reasons, does anyone with real developers run SQL on an externally exposed website?
The money on laptops is leverage money: it reduces the costs of books, printing documents, planning, and communicating with the students and their families. And you had better believe those laptops will be *family* resources for information access for the whole family, with information those families could otherwise find incredibly difficult to obtain about local weather, crop prices, available markets, jobs, and even literature. And it does it in a modular way, one that does require infrastructure but far less than a new set of books.
Go ahead, name a single African culture that has "no concept of the future".
The Classmate also doesn't have a Linux BIOS, which provides a lot of flexibility and a far, far, faster and battery saving boot and shutdown time.
No, they're not. Both Intel and Microsoft have a long, documented history of stealing work from other companies, then using the stolen goods to drive the other company out of business. And they double teamed DEC, with Intel stealing the Alpha technologies and Microsoft stealing the VMS technologies. They've both got far more of a history of this than their competitors.
Welcome to the NVidia drivers, which cannot be pre-installed because of this and are installed exactly how you describe. Then their installer screws up your local system: they can't use a standard .deb. or .rpm installer because they insist on users manually signing their software license, and you can't automate that without violating their license agreements.
As soon as you provide closed source modules for the kernel, such as the NVidia modules and some the McAfee Linux components, you "taint" the kernel. That means you cannot pre-compile the kernel and distribute it with your models pre-installed, they "taint" the kernel. You have to provide your components separately.
This drives the closed source designers, like McAfee and NVidia, nuts, because they can't pre-include their tools. It's worse because most of their installers suck really, really hard, and actually imperil the systems they are run on even more than a typical security risk does. And it drives we open source developers nuts because people want to know why their tools don't run automatically or well, or why they're not already built in. And we have to waste our time fixing the mess caused by their components for which we have no authority, no control, and no access to address the failures.
Well, the "theft" of our work is a loaded phrase in legal terms and confuses the lines between moral and legal obligations. Please use the word "freeloader" or another term, because the use of open source tools to build closed source products is permitted by some licenses, such as BSD. This is precisely why I think the BSD license is a freeloader on the GPL license, which protects against exactly the freeloading you describe, and provides a lot more territory for the BSD closed source authors to operate.
Moreover, clear or not, it's been repeatedly tested in court, in lots of interesting ways. Take a look at http://boycottnovell.com/2007/11/23/gpl-court-test/ for references, and a Google search reveals far more links with good citations.
With cause. Their Linux operating system products certainly include kernel modules or customized kernels: the Linux kernel is GPLv2. This means that installing closed source, proprietary security modules "taints" the kernel and means the tainted components cannot be included as a built-in part of the kernel. This is vaguely similar to the problem NVidia has with their proprietary drivers and OpenGL libraries, and the absolutely awful procedures for installing such modules that various closed source vendors use. I assume that the problem also occurs with their other components, but I can directly report the kernel module issue.
The "oh, no, they'll see our code" concern is a basic security through obscurity claim, and is common to a lot of commercial security vendors. It's countered by the difficulty of testing and integrating such closed code into lighter weight, better featured or better performing tools.
We also need to keep an eye on these security companies and their desire to integrate DRM support that could certainly be useful for authenticating software installations and protecting user's files, but can easily be used to enforce "you can only read this file with this software" features that for business reasons, they will need to support as so-called "Trusted Computing" is supported in more CPU's and is deeply integrated into Vista and the next Microsoft release. They need to protect those features for obvious business reasons, and not allow the publication of open source code to access the DRM protected features. The GPLv3 is aimed directly at this problem, and McAfee surely is thinking about how to protect that market.
Oh, it happens in the Linux world too. There are packges that install new users and mis-set them in the open source world, or overwrite system files and make recovery a nightmare. NVidia drivers are a classic example, as are developers who randomly pull in PHP or Perl modules without packaging them and create awful dependencies that break other packages. Kernels, gcc, and Perl modules are classic examples.
Well, maybe. But it's a backup: you're supposed to be able to restore reliable from it, not have to hand massage it. Doing a "sync" operation before doing the suspend might help. If you're backing up something big, like a database, and someone executes operations right in the middle of it, then your files may wind up corrupted. And that metadata *itself* may be corrupted, depending on which backup technology you use. If you can, it might be worth locking a database temporarily and then running sync while you do the snapshot.
All in all, I'd rather be safe than sorry with backup data. And having a text dump of something like a MySQL database means that you can also restore to another version or another system, rather than relying on having the exact same version of MySQL.
Such a sense of humor is not polite to your staff, and can be avoided by simply calling the VMWare instance on your desktop //desktop-vmware, so htey know whom to go looking for.
And the kernel of that hypervisor is.... Linux 2.4. So its ability to deal with contemporary hardware relies on backporting the drivers from 2.6. This way lies a fascinating level of difficulty, and the need to run ESX on rather limited hardware.
They're incredibly useful for QA andn build environments, though. Being able to whip up a new OS from a base image in a matter of minutes is a pearl of great price, and helps eliminate the 3 hours of downloads and updataes for all sorts of awkward environments.
"suspend" is not enough for a database, like MySQL, or files that were being written and are still resident only in RAM. You need to actually do a shutdown, or make sure that sensitive databases were backed up. If you don't believe me about such issues, try it and then run "fsck" against the partitions of the virtual guest.
I've heard of such, but they tend to be very, very nasty, and prone to causing more service interruptions than they cause. Lightweight userland failover systems, such as creating MySQL databases in master/master configurations and using IP failover tools like wackamole seem to work more gracefully, and without imperiling your kernel and hardware support.
For lightweight systems, you can use rsync and database backup tools to store the database, then rsync the filesystems to a repository. For full systems, you can do a very fast shutdown of most virtual OS's, use LVM or another partition based snapshot technique, restore the virtual server, and backup the snapshot.
Theh combinations can get even more interesting, but you may begin to see the possibilities.
Which GPL? GPLv3 includes changes aimed very specifically at preventing another of these patent protection deals, and putting a spike in Microsoft attempts to encumber open source projects with patent agreements. The GPL needed to evolve to prevent this and some of the DRM craziness being attempted, and has. But will Novell follow suit by publishing its own software or changes under GPLv3?
I think not.