The key problem here is not so much a flawed OS structure but a flawed hardware structure. The whole PC hardware architecture was never designed around fault tolerance, that's why the PC hardware is so much cheaper than a zSeries mainframe or even a Sun E-series. There's no guarantees with PC hardware, that's why PC fault tolerance always involves failover boxes. That aside, there are things the OS can do to handle hardware failure but I think of it like an airbag in a car, it may save you,though not always, but your car is gonna look like hell afterwards anyways.
How big are these "spoons" of salt? Two tablespoons of salt will taste like hell but it shouldn't kill you. You're talking about ~10g per tablespoon while the fatal oral dose of table salt is somewhere in the ~100g range. Now, 10 heaping tablespoons of salt, that'll f--- you up real nice and proper.
Let me get this straight, patient monitoring systems are plugged into the same LAN in which doctors, admins, and what-not are free to plug in their laptops? I don't work in a hospital but even we have DMZ subnets for more sensitive parts of our network. I can't (or rather don't want to) believe that hospitals don't segment their networks the same way.
I would hope hospitals do not run critical systems a) on Microsoft software but especially b) on a LAN with any access to the internet. It's sheer lunacy if they do and could be used as grounds for a lawsuit. On the otherhand, they can do whatever they want with their accounting, cafeteria, and parking meter systems since a lawyer wouldn't pounce on that kind of... wait... I'm probably underestimating now.
I think they're complaining about all the dependencies for compiling the app. The majority of the installs I've done are runtime only environments. Setting up a full development environment is overkill for the vast majority of not-so-techie users.
<rant>
That said, my beef with GnuCash is the same as my beef with KDE apps. Tying an application to a desktop is a carry over from the Win-word that I simply do not agree with. Code to GTK or QT unless your app is an integral part of the Gnome or KDE desktop.
</rant>
I do think GnuCash is a good effort and these guys have obviously poured their hearts into it. However, I switch between IceWM, TWM, Gnome, and KDE depending on the machine, day, task, etc. Because of this, I avoid apps that are tied to desktops for this very reason. I know they sometimes work but only by starting up a whole rarely even used subsystem. Maybe somebody here can enlighten me by giving me a good reason why the GnuCash team ported the original sources from Motif to Gnome instead of plain old GTK. I don't buy the "familiarity" excuse since happened in 1999 (Gnome 1.0 had *just* arrived), not 2003.
Good point! Are any other of the pissant Canopy group companies in the Fortune 500? What better way to weasel out a press release than by having your sister company buy a worthless license from you.
We're seeing a steady upward trend in 135 reqs too. Much worse from our backup ISP than our primary. We've got our firewalls flicking these off at the doorstep but then again they were never allowed in in the first place.
There's a flaw in your argument. You can purchase private security other than your local police department. That doesn't make your local police any less of a public good. Likewise, you could choose not to run free software.
My personal take on it is that the government should cut back on purchasing proprietary software and use free software whenever possible. Those savings could then be used to cut taxes and give taxpayers back more of their paycheck. I don't like the idea of the government handing cash to free software developers any more than to proprietary software developers. If the government does contribute to free software it should be though man-hours that advance features the government itself requires.
I'll bet it took a lot longer than 3 months to put that together too. I have material collected for a tech book (Apache/JBoss/Tomcat material) but the sheer scope of compiling 200+ pages of notes into something coherent just doesn't seem worth the potential reward of a few thousand bucks. I think that's why we see a lot of tech authors who release one book and just give up on making revised editions.
A certain database software company I worked for when I got out of university had the same "get 'em off the phone quick" mentality. It wasn't what you fixed, or your satisfaction rating, but rather how many you could turnover in a day that mattered most. The name of the game was to shunt somebody off with a bulletin and a "reopen this if you still have problems after you've read that". Not so ironically, it's been a while since I worked there.
I'm totally with you here, GRUB is miles ahead of LILO. I'll admit I didn't like GRUB at first but that boot command line feature is something I simply would not do without now, especially when I'm torking with kernel modules.
Even better, it's against clasue #3 of the GPL to not distribute the source code or at least post a notice offering the source code. Chaulk up another actionable violation.
Absolutely!!! I have paid for several pieces of commercial software for Linux namely Win4Lin, Codeweavers, Netlock, and StarOffice. I especially like the StarOffice model with an open source base product and the value added "finished" product for a fee.
It's more like SCO is a chihuahua that was nipping and yipping at the IBM grizzly bear. The grizzly wasn't too annoyed at first but eventually, when the chihuahua just wouldn't stop, the grizzly spun around, and grabbed the chihuahua firmly in it's jaws. The point we are at now is where we just wait for the crunching and splattering of blood.
I use lprng explicitly over CUPS which is the default on SuSE. The reason is simple, I could never get CUPS to network printers consistently. I'll say that this impression dates from a while back but I've never had a problem with lprng so I've never looked back. Also, some of propreitary drivers of a certain printer company only work with lprng.
I was looking for a post like this. My biggest annoyance is definitely X11. Don't get me wrong, I think X11 is an awesome piece of work but getting it configured "just right" really tests the patience. I keep multiple copies of my XF86Configs for all my machines locked away in CVS to avoid having to re-live the configuration experience. It's also the one thing that simply cannot be done remotely.
My second worst experience is related to nForce chipsets. I know this is a new chipset and I expected getting Linux up and going to be a challenge but getting 2.4.21-pre3+ (initial support for nForce) took hour upon hour of tweaking with the kernel boot params.
All in all though, these two irritations are well on par with the driver hell experiences I've had on Windows. Minded this was pre XP, I've never used XP.
Adobe helped to develop WINE to run their Photoshop.
Where did you get this information from? The only contribution I've seen from Adobe to WINE is with regards to font metrics, and even that was indirectly. Codeweavers did most of the work for getting Photoshop to run on WINE and that was primarily sponsored by Disney not Adobe. You might well be correct but this is this would be the first I've heard of it since I can't even find mention of it on the WINE mailing lists.
There isn't any. However, SCO going after the government should negate most of the influence that SCO had via Orrin Hatch who's son is one of their lawyers. Since this is an election year for Senator Hatch, I'm quite sure he'd rather keep a low profile over ties to a company looking to extort public money, especially after allegations of his being "bought" by the pharmaceutical giants.
Your example is wrong. Here's the proper scenario, you publish your code under the GPL, I have access to and the opportunity to steal some of your code for my own use. I then proceed to accuse you of copyright infringement. You ask to see the code and I tell you I want you to sign an NDA that would prevent you from ever working on your GPL'd code again. That's the SCO case in a nutshell. Even that is only one possibility.
With regards to the GPL and copyrights, there is a Copyright notice in the Linux kernel code from Caldera. Remember, Linux is the "program/work" released under the GPL, each individual file need not bear a copyright notice. Also, a quick grep reveals that there are many contributions from "sco.com", "caldera.com", and "caldera.de" email addresses. That said, SCO has stated that the 2.2 kernel is not infringing, this means that the infringing code was inserted in the 2.4 timeframe which dates to January 30, 2001. The oldest development release of 2.4 I could locate was dated August 11, 2000. Not so ironically, Caldera purchased SCO on August 2, 2000, so code donated by Caldera from the SCO code base to 2.4.0 would fall under the GPL. All that time between August and January was available for Caldera (ie. ex-SCO) developers to donate SCO code to Linux.
Sure one can counter sue. However, you have to counter sure over something specific. Exactly what do you think SCO will counter sue over? They have no contract with Red Hat other than the GPL'd code for which they have yet to show any proof publicly and for which no court has shown they even own in the first place. SCO will bark and grumble but there's nothing they can sue Red Hat for that would not involve disclosing the GPLed code and I doubt SCO would risk their new Linux FUD and licensing business over what is essentially a gag order.
So will the next South Park movie feature Darl McBride as Satan's new love interest? I can see it now, Darl's flapping head spouting, "Aw, come on, guy. Pinch my nipples while I torture this little partridge..."
I found this link which has more details, looks like it is EAL2 after all. I also found that Red Hat and Oracle are planning on going after EAL4 for the latest RHAS so the W2K advantage might be short lived.
Slashdot Mirror
The key problem here is not so much a flawed OS structure but a flawed hardware structure. The whole PC hardware architecture was never designed around fault tolerance, that's why the PC hardware is so much cheaper than a zSeries mainframe or even a Sun E-series. There's no guarantees with PC hardware, that's why PC fault tolerance always involves failover boxes. That aside, there are things the OS can do to handle hardware failure but I think of it like an airbag in a car, it may save you,though not always, but your car is gonna look like hell afterwards anyways.
How big are these "spoons" of salt? Two tablespoons of salt will taste like hell but it shouldn't kill you. You're talking about ~10g per tablespoon while the fatal oral dose of table salt is somewhere in the ~100g range. Now, 10 heaping tablespoons of salt, that'll f--- you up real nice and proper.
Tack on platform optimizations and compiler options and the success rate would drop even further in a "source code included" world.
Let me get this straight, patient monitoring systems are plugged into the same LAN in which doctors, admins, and what-not are free to plug in their laptops? I don't work in a hospital but even we have DMZ subnets for more sensitive parts of our network. I can't (or rather don't want to) believe that hospitals don't segment their networks the same way.
I would hope hospitals do not run critical systems a) on Microsoft software but especially b) on a LAN with any access to the internet. It's sheer lunacy if they do and could be used as grounds for a lawsuit. On the otherhand, they can do whatever they want with their accounting, cafeteria, and parking meter systems since a lawyer wouldn't pounce on that kind of ... wait ... I'm probably underestimating now.
I think they're complaining about all the dependencies for compiling the app. The majority of the installs I've done are runtime only environments. Setting up a full development environment is overkill for the vast majority of not-so-techie users.
<rant>
That said, my beef with GnuCash is the same as my beef with KDE apps. Tying an application to a desktop is a carry over from the Win-word that I simply do not agree with. Code to GTK or QT unless your app is an integral part of the Gnome or KDE desktop.
</rant>
I do think GnuCash is a good effort and these guys have obviously poured their hearts into it. However, I switch between IceWM, TWM, Gnome, and KDE depending on the machine, day, task, etc. Because of this, I avoid apps that are tied to desktops for this very reason. I know they sometimes work but only by starting up a whole rarely even used subsystem. Maybe somebody here can enlighten me by giving me a good reason why the GnuCash team ported the original sources from Motif to Gnome instead of plain old GTK. I don't buy the "familiarity" excuse since happened in 1999 (Gnome 1.0 had *just* arrived), not 2003.
Good point! Are any other of the pissant Canopy group companies in the Fortune 500? What better way to weasel out a press release than by having your sister company buy a worthless license from you.
Yep, I wouldn't bet against you on this one.
We're seeing a steady upward trend in 135 reqs too. Much worse from our backup ISP than our primary. We've got our firewalls flicking these off at the doorstep but then again they were never allowed in in the first place.
There's a flaw in your argument. You can purchase private security other than your local police department. That doesn't make your local police any less of a public good. Likewise, you could choose not to run free software.
My personal take on it is that the government should cut back on purchasing proprietary software and use free software whenever possible. Those savings could then be used to cut taxes and give taxpayers back more of their paycheck. I don't like the idea of the government handing cash to free software developers any more than to proprietary software developers. If the government does contribute to free software it should be though man-hours that advance features the government itself requires.
I'll bet it took a lot longer than 3 months to put that together too. I have material collected for a tech book (Apache/JBoss/Tomcat material) but the sheer scope of compiling 200+ pages of notes into something coherent just doesn't seem worth the potential reward of a few thousand bucks. I think that's why we see a lot of tech authors who release one book and just give up on making revised editions.
A certain database software company I worked for when I got out of university had the same "get 'em off the phone quick" mentality. It wasn't what you fixed, or your satisfaction rating, but rather how many you could turnover in a day that mattered most. The name of the game was to shunt somebody off with a bulletin and a "reopen this if you still have problems after you've read that". Not so ironically, it's been a while since I worked there.
I'm totally with you here, GRUB is miles ahead of LILO. I'll admit I didn't like GRUB at first but that boot command line feature is something I simply would not do without now, especially when I'm torking with kernel modules.
Even better, it's against clasue #3 of the GPL to not distribute the source code or at least post a notice offering the source code. Chaulk up another actionable violation.
commercial and open source software can co-exist
Absolutely!!! I have paid for several pieces of commercial software for Linux namely Win4Lin, Codeweavers, Netlock, and StarOffice. I especially like the StarOffice model with an open source base product and the value added "finished" product for a fee.
It's more like SCO is a chihuahua that was nipping and yipping at the IBM grizzly bear. The grizzly wasn't too annoyed at first but eventually, when the chihuahua just wouldn't stop, the grizzly spun around, and grabbed the chihuahua firmly in it's jaws. The point we are at now is where we just wait for the crunching and splattering of blood.
I use lprng explicitly over CUPS which is the default on SuSE. The reason is simple, I could never get CUPS to network printers consistently. I'll say that this impression dates from a while back but I've never had a problem with lprng so I've never looked back. Also, some of propreitary drivers of a certain printer company only work with lprng.
I was looking for a post like this. My biggest annoyance is definitely X11. Don't get me wrong, I think X11 is an awesome piece of work but getting it configured "just right" really tests the patience. I keep multiple copies of my XF86Configs for all my machines locked away in CVS to avoid having to re-live the configuration experience. It's also the one thing that simply cannot be done remotely.
My second worst experience is related to nForce chipsets. I know this is a new chipset and I expected getting Linux up and going to be a challenge but getting 2.4.21-pre3+ (initial support for nForce) took hour upon hour of tweaking with the kernel boot params.
All in all though, these two irritations are well on par with the driver hell experiences I've had on Windows. Minded this was pre XP, I've never used XP.
Adobe helped to develop WINE to run their Photoshop.
Where did you get this information from? The only contribution I've seen from Adobe to WINE is with regards to font metrics, and even that was indirectly. Codeweavers did most of the work for getting Photoshop to run on WINE and that was primarily sponsored by Disney not Adobe. You might well be correct but this is this would be the first I've heard of it since I can't even find mention of it on the WINE mailing lists.
There isn't any. However, SCO going after the government should negate most of the influence that SCO had via Orrin Hatch who's son is one of their lawyers. Since this is an election year for Senator Hatch, I'm quite sure he'd rather keep a low profile over ties to a company looking to extort public money, especially after allegations of his being "bought" by the pharmaceutical giants.
Your example is wrong. Here's the proper scenario, you publish your code under the GPL, I have access to and the opportunity to steal some of your code for my own use. I then proceed to accuse you of copyright infringement. You ask to see the code and I tell you I want you to sign an NDA that would prevent you from ever working on your GPL'd code again. That's the SCO case in a nutshell. Even that is only one possibility.
With regards to the GPL and copyrights, there is a Copyright notice in the Linux kernel code from Caldera. Remember, Linux is the "program/work" released under the GPL, each individual file need not bear a copyright notice. Also, a quick grep reveals that there are many contributions from "sco.com", "caldera.com", and "caldera.de" email addresses. That said, SCO has stated that the 2.2 kernel is not infringing, this means that the infringing code was inserted in the 2.4 timeframe which dates to January 30, 2001. The oldest development release of 2.4 I could locate was dated August 11, 2000. Not so ironically, Caldera purchased SCO on August 2, 2000, so code donated by Caldera from the SCO code base to 2.4.0 would fall under the GPL. All that time between August and January was available for Caldera (ie. ex-SCO) developers to donate SCO code to Linux.
SCO will now get to counter sue
Sure one can counter sue. However, you have to counter sure over something specific. Exactly what do you think SCO will counter sue over? They have no contract with Red Hat other than the GPL'd code for which they have yet to show any proof publicly and for which no court has shown they even own in the first place. SCO will bark and grumble but there's nothing they can sue Red Hat for that would not involve disclosing the GPLed code and I doubt SCO would risk their new Linux FUD and licensing business over what is essentially a gag order.
So will the next South Park movie feature Darl McBride as Satan's new love interest? I can see it now, Darl's flapping head spouting, "Aw, come on, guy. Pinch my nipples while I torture this little partridge..."
You've got it wrong, that's a typo in Darl's letter, here run this to correct it:
s/Linux/our lawsuits against Linux
I found this link which has more details, looks like it is EAL2 after all. I also found that Red Hat and Oracle are planning on going after EAL4 for the latest RHAS so the W2K advantage might be short lived.