XFree86 and many other X environments use what's known as the X Shape Extension. You can have an application display in an arbitrary shape, xeyes being the canonical example.
I think what he was trying to explain is not that it's difficult to use, but that it's less likely to abuse. All you need to to to be infected by MyDoom under Windows is double-click on the file after it's extracted. If this were under Linux, UNIX-based systems really, you'd need one more step - basically, you give permission to the file to be executable. It's only harder to accidentally run something. The same holds true in MacOS X for example, and there aren't that many MacOS X users complaining that it's difficult to use.
Here's an exercise for you. Right click on a file under Windows >=NT4. Somewhere in there (depending on the version) there's a setting to change permissions on your file, typical DOS-derived permissions give you Read-only, System and Hidden. It's something like that under UNIX as well, indeed under KDE (a UNIX desktop environment) it looks nearly the same. There's just an extra option called Exec(ute), which can be specified differently for users and groups of users.
That being said, it's quite possible for such a virus to spread under Linux/UNIX. The difference is that the security model of UNIX-like systems, especially users/admins giving applications permission to execute, makes it more difficult for these viruses to spread. Aside from market share and the ability to run on multiple platforms, that's why viruses under UNIX-like systems tend to wither and go away rather than propagating far and wide.
It's sort of like the OS equivalent of seat belts and air bags. Think about it: if Windows required users to give permission to an application before executing it, instead of automatically executing programs which masquerade as documents, odds are there would be far fewer viruses being spread in this manner.
Odds are you still had libs from 3.1 loaded, and the symbols were interfering with the 3.2 apps. Killing off the old processes before starting the new ones, and probably an ldconfig if you're on Linux, is probably all that was needed. An init 1 certainly works too.
Sure, but a Linux user with enough intelligence to install and use a cross compiler wouldn't be stupid enough to write something like MyDoom. Unless, of course, their intention was to support SCO's court case with more line noise.
There are certainly more Visual Studio users than Windows developers using cross compilers on Linux by several orders of magnitude. You're probably looking at Vegas odds here.
Odds are that it's someone with a copy of Visual Studio in a Windows environment. A Linux user probably wouldn't be wasting their time with this. Not to say that it couldn't be a Linux user, anything's possible, but Linux users have less to gain from MyDoom than SCO does.
FWIW, it's not like SCO's going to lose any business from their website being down. The bad faith they've generated with their wild claims took care of that for them, most Linux users understand this. The only result is another false leg for SCO to stand on in their PR department, which the article's author fell for.
Perhaps you're right about me not understanding you. There is not only one application in KDE, as certainly as there is not only one application in GNOME. They're both a set of smaller applications, which can be mixed and matched with each other. I think that's great, if I don't like part of one I can use the other. If there's only one thing in the other that I like, I can still use it. There is no non-trivial application that is the best for everyone, and I seriously doubt that it's possible to create such an application.
As far as a "one massive, bloated but...." consider emacs. It's open source, it's very large, is a set of applications written in lisp by its own admission and can do just about anything. Any vi (or derivative) user will tell you it's bloated. It's also 100% open source, GPL software originally written by Richard Stallman. There really was no reflection on KDE here, as it's a set of relatively lightweight apps for what they do. They're configurable and flexible, but in the end they just do what they do. Even konqueror is just a wrapper for various plugins, albeit a very flexible one.
Choice in its essence requires more than one out of which one can choose. That's what it means. What you're talking about is configurability. Don't misunderstand, I'm not trying to belittle your English. I'm merely stating my position, which you disagree with. The fact that you and I will not agree on this may say something about why there will probably never be one application that does what everyone wants, exactly as they want it to be done. If it truly is possible, then show me what it is that you believe I really want.
As for that last bit... I'm not sure how you got this impression, but the Linux kernel isn't all that big really. Most of the source that you download consists of drivers for the nearly infinite varieties of hardware that it supports. If Linux fails to meet my needs however, I'm free to boot another OS. Luckily this hasn't been an issue, but e.g. FreeBSD is always there if I someday need it. By the same token, if KDE somehow didn't meet my needs, I could use something like Windowmaker or Blackbox (both of which I occasionally use, esp. in low memory situations and in VNC.) XFCE is a great choice also, in fact I tried dozens of WMs and "desktop environments" before finally deciding that KDE was my preference. The beauty that I see in Open Source is that I have these choices. I don't see any beauty in censoring Linux in the manner you suggest, but do what you will.
It's not a fact, it's a question. Do we want a choice of applications, or do we want one massive, bloated but common set of applications that do everything that anyone might conceive that we want them to do? Like Windows and Office I suppose, but if that's all that you're looking for then a trip to Wal-Mart will fix that for you. I think that we have something better already.
A hammer might very well be able to break rocks, but I don't want to be the poor sap stuck doing that all day. By the same token, if an application isn't comfortable to me, then I'm going to pick another. The same can be said about desktops (the topic of this discussion.) People will argue all day about their preference of EMACS or vi, MacOS vs. Windows, etc. That's because they have a preference. Neither group would want their preferred application to become like the other.
What I like most about Linux is that I have a choice of just about anything without leaving the OS. That's what I want someone to see when they start using Linux. You're not stuck with what some guy in Redmond or Cupertino thinks you should use. If that's what this distribution is about, then it will come without the vitality and unique flexibility that is available with Linux. I can't and won't recommend that to anyone.
As a point of fact, both Qt and GTK can work without X on a framebuffer. Many commercial Linux-based PDAs use Qt on a framebuffer via the Qtopia environment. While that's neither here nor there, it does point out that the choice exists. In some cases, it simply makes more sense to use an alternative.
Whether or not you like choice, there are plenty of people who do. Many people can't even agree on pizza toppings, never mind what they want their desktop to work like. I personally don't think there's anything wrong with that, I'd rather people not dictate what desktop I can use. The concept of "best" is highly subjective, e.g. your vision of best practices in Open Source is different than mine.
You may remember that GNOME started out solely because KDE used a widget library (Qt) which at the time had a license with terms that some people didn't like. That was their choice, and they "wasted" a lot of effort coming up with a set of libraries and applications that did basically the same thing that KDE already did. In the end they have a desktop which suits their ideals, and that's fine. I prefer KDE for my desktop, but I support the ideals which lead GNOME to become what it is.
I do not support those who would state that I shouldn't have the choice to use KDE however. If a distribution decides that I should use GNOME and doesn't support KDE out of the box, then I'll simply not use it. Further, I won't recommend it to anyone else since I perceive it as being too restrictive.
To be honest, I don't find it very comforting that said distribution will tack on what users want as an afterthought either - they either will support these things, or they won't. If the latter is the case, then I don't believe they deserve to succeed. I didn't start using Linux to get another "One Microsoft Way", and I don't intend to go back now... I hope that this isn't the future for Linux.
Probably the latter. The ribbon is the connector which attaches the hard drive to the laptop. It's actually just an IDE drive, the power goes through the same connector as the data though. You can buy cables which plug into a standard HD power source and a 40 pin IDE port in a desktop PC for that matter. The enclosure of the drive can be opened on every Dell I've seen, from the Inspiron 3500 to the new Latitude D800 systems. The same is true for the IBM Thinkpads.
The hard drives in most laptops these days are identical. Apple uses IBM (now Hitachi) drives for that matter, I have a PB G4 open next to me rihgt now. (Getting the drive out of a PB G4 is a bit tricky though, heh.) I've actually taken a Mac drive and put it in a Dell laptop, it works fine. The reverse has also been the case as was noted here.
All this discussion about guilt is really silliness IMO. People should use the OS they are comfortable with, whatever that might be. I prefer Linux personally, but I know people who prefer OS X, Solaris, various BSD flavors. Even Windows. In the end, all that matters is that you have an environment which you enjoy working in. If it makes you happier to have gumdrop shaped widgets and a dead simple/nearly intuitive UI, then you owe it to yourself to get a Mac. If you want lots of current games, or need a lot of support from those around you, get a Windows box. If you want a genuine UNIX environment without all the extra fluff, pick a flavor of Linux, BSD or even Solaris that makes you happy. Just don't pick SCO;-)
> They might be switching from GNU/Linux systems to something a little bit more suited to enterprise environments.
I don't agree about the "more" suited bit, but I'll skip the flamebait. The real question is, why would they do a thing like that? They're hardly an enterprise environment outside of litigation. Given their current business model, Linux is overkill.
Outside of a DDOS, they're probably just as well with PWS running off some Windows 98 box. More likely they're just hiding behind a proxy service like Akamai, as MS did in a similar situation. Imagine that, SCO using tactics borrowed from MS.;-)
That's great, though I certainly wish the documentation had existed before we implemented our workarounds. When it was critical, it simply wasn't there without an NDA (which I'm not authorized to enter into.)
It's only a big lie if W2K is used for the server. There's no way in hades that I'm using a Windows box for a DC on my networks.
Back to the original point: Kerberos is a protocol which Microsoft has extended, placing their extensions under trade secret status to prevent others from benefiting from the work. That's the thing that they've actually done to irritate me.
We use a UNIX KDC, and won't be implementing a Windows DC. Since the Windows systems want to get their authorization info from an AD domain, based on Microsoft's extensions to the Kerberos protocol, we have to do some extra steps in order for our Windows systems to properly integrate into our network file system. It could have been much simpler had they not declared their extensions "trade secrets".
I'll hold you to that, for my part I won't let a Linux distro that doesn't have KDE touch my desktop/laptop anymore. I for one will admit to being a KDE zealot, Political Correctness be damned. I simply like it better.:-)
Those extra, proprietary, and incompatible Kerberos fields in W2k+ for one. I think they put those in just to make my life difficult. Of course, it wouldn't have mattered if they'd simply made their proprietary extensions open without having to sign some bizarre NDA...
That may well be, but the fact that Mandrake is coming out of bankruptcy at all says something about the company. Most companies in that position disappear well before it gets to that point.
SuSE is a nice product, but there's no downloadable ISO version. While that's given them a steady stream of revenue, it's also prevented more people from trying it and eventually becoming a customer. Mandrake's free ISO releases give them goodwill from their userbase, which tends to translate into brand loyalty over time. Keep in mind that there are plenty of people who like Windows simply because it came "free" with their computer, even if it does need to be wiped out and reinstalled every now and then.
SuSE is also significantly different from Redhat in many respects, whereas Mandrake was originally based on Redhat and has kept many of the filesystem conventions. This may seem insignificant, but it does simplify matters with e.g. binary RPMs for commercial software. My experience has shown Mandrake tends to take Redhat-targetted software with less effort than SuSE.
In any case, I don't buy the argument that there can be only one desktop-centric Linux distribution. The day that we're stuck with just one distribution is IMHO the day that Linux becomes another victim of Microsoft. I neither want nor believe that this will happen. I think that there's enough room for both SuSE and Mandrake, and hope that the competition between the vendors will continue to improve their quality and feature sets.
I don't see this as a problem. If a law enforcement official tomorrow were to confiscate your vehicle in the course of investigation because they suspected that you were involved in some sort of illegal practice, where would this leave you? Given that you only have one vehicle, and that you need the vehicle to work in order to live, wouldn't you say that you've suffered damages? For that matter, if you get the vehicle back in 2 years when they determine you're innocent, how useful is it to you then? Will you even have your job still?
Law enforcement needs to operate as efficiently as possible, but they also need to be certain that what they are doing is correct. If they aren't correct, then that should effect them in a negative manner so that they will make an effort to be correct. I still believe that people are innocent until proven guilty, and seizure of property without 100% certainty of wrongdoing is IMO ridiculous.
D'oh, never mind. After re-reading your comment, I noticed the "5 drives" and not "RAID 5 on 4". Still, level 5 isn't nearly as useful if you can't replace the drive quickly. I wonder how soon before we get hot swappable bays in a drive of this size?;-)
It's probably using RAID level 0 striping if it's RAID at all. RAID level 5 gives you (n-1)*disk capacity where n is the number of disks. Thus, if it was RAID level 5 and you had 4 250GB drives, you'd come up with (4-1)*250GB, or 750GB capacity, before filesystem formatting. Of course, if any of the drives failed, you could swap it out and put another in its place. This case is hardly appropriate for drive swapping though, so odds are it's either level 0 or just concatenated disks.
As an aside, I've seen software RAID over Firewire in action. It isn't as reliable as one might hope for, and you'd better hope noone accidentally pulls out one of those easily removed cables;-)
I got fed up with the noise from my Athlon XP 2000+ a few months ago. It was so loud that it could easily have been mistaken for a vacuum cleaner. I was using a generic 6-fan case (4 in the front, one on the side and one at the rear) and a Cooler Master 6700rpm heat sink to keep the heat down.
Having heard good things about the Antec Sonata, I went out and picked one up. After a bit of looking into what heat sinks were available at the local Fry's, and basically just wanting anything less noisy than the Cooler Master, I picked up a Zalman CNPS7000A-Cu.
Currently, the Zalman is spinning at 1430 RPM (fwiw the P/S fan is spinning at 1834 RPM.) It can be adjusted to suit your cooling needs with an included adjustment unit, but I was going for quiet. The computer now runs about 10c cooler than before on the CPU, and perhaps 15c cooler on the MB according to the onboard sensors. Not bad, considering that both fans are running at a fraction of the speed from the previous case. The only issue I could see someone having with the heatsink is that it's so large that it doesn't fit on all motherboards.
While the Zalman helped, the Sonata has really made a significant difference in both the area inside AND around my computer. Antec built this case with what is quite possibly the best configuration for hard drives I've ever seen in a PC case: they're mounted sideways on individual HD trays. The drives are mounted to the tray using rubber grommets, which make the two 7200RPM drives in my case nearly silent. The sound of the drives accessing, which before was a rather loud chatter, is now nearly inaudible. There are 4 internal 3.5" HD bays, so I didn't have to give anything up for the silence. Better yet, due to the design of the trays the HD cables don't block the airflow of the case since they go directly to the back of the case. The case even looks nice, with a glossy black finish on the metal parts of the case and matte black plastic on the front.
For someone who wants to throw together a system and get nearly all the parts in one box, the Sonata is an excellent case to go with. The Zalman complements the case nicely, though if you have a preferred quiet heatsink you might be better off using that - the Zalman is not compatible with all motherboards due to its large size. I'm quite happy with it on my system though (using an Asus A7V266-E).
Ok, I'll bite again. This may not convince you of anything, but I will maintain that I disagree.
If Redhat had merely eliminated updates for current releases to non-paying customers, I could buy that what they'd done is perfectly ethical. What they have done though, is eliminate all of their current products and replace them with different products requiring higher initial and continuing investment than their previous products. They also changed the entire pricing of their software lineup. In this light, the TCO argument which is constantly used for Redhat Linux wears thin.
Due to Redhat's policy change, any company which has based their product on an existing release of Redhat needs to either upgrade to a different Redhat version, provide updates themselves, do no updates, or switch to another distribution. In the projects I'm involved with, doing no updates is not a solution. Upgrading all of our systems immediately is even less of a solution. I'm not banging on the "openness" drum here, I'm talking about a stable platform being yanked out from under our feet. We may need to go the Progeny route until things are worked out on some of our systems, money which Redhat could be making had they not dropped the ball. Against this backdrop, it's laughable to think that they want to compete with the likes of Sun Microsystems. They seem to have neither a clue nor a desire to know what their customers' needs are.
That many people want people to be given things for free is to be expected. People in general seem to have no qualms about "borrowing" and installing a copy of licensed software on multiple systems from what I've seen. These people don't care about how libre the system is, only how gratis. They're not going anywhere, and plenty of them are users of proprietary systems. Still, open source is not a business model. It's a development strategy. It exists not because businesses are paying for it, but because people want something better (or at all.) Seriously. If it benefits businesses, so much the better. That's generally not the primary reason for development though.
If a software product survives only because a small number of businesses pay staff developers to write it, then the software probably isn't interesting enough to a large number of people to continue its development. If it is, someone will come along and do it. It may or may not be a business, but if a closed source product is attractive/useful enough then someone's going to want a good open source version. If someone has an itch to scratch and there's no product out there that does what they want, there's a strong possibility that they'll open source it given the collaborative development benefits. Bug fixes/additional development returned to the initial developers may outweigh their initial development costs. The price to end users is simply a happy coincidence.
I don't believe Redhat should provide support in perpetuity without some sort of remuneration, but they should provide some form of extended support for those who are willing to pay for it. A refusal to do that for the systems of current customers where significant investment has been made shows that Redhat clearly isn't ready for the enterprise.
Surely redhat has no legal obligation to provide support to customers who have paid. I'll grant you that. The fact that a large set of their customers are upset however, should be a clue that an ethical obligation has not been met.
Redhat initially started with the product "Redhat Commercial Linux". They touted their distribution as a commercially supported variant, as opposed to what was then currently available (Yggdrasil, Slackware, etc.) Under the GPL, their legal obligations were nothing. Their ethical obligation however, the stated purpose for their existence, was to support their distribution in ways that non-commercial entities would not be able to.
If Redhat continually modifies their support policies as they have, I'd say that ethically they're shirking their duties. Legally they can tell people to get lost, since there wasn't a contract stating otherwise. Don't confuse ethics with legal obligations though. Both are important to the perception of a company's value, but they are not the same.
Amongst other things, I work with networked embedded systems a good deal. Our investment in Redhat is now under significant question. 1 year is nothing to a system which may be developed over a period of potentially several years, and updating the distribution every year changes the systems in significant ways (hardware requirements, timings, etc.) Redhat was for a while touting their product as ideal for embedded systems, but this has certainly changed. Their product is simply no longer useful to us.
I understand where you're coming from, but I believe that a company's actions should be rewarded or punished appropriately by the consumer. Redhat's actions are simply not in line with those of an enterprise-class software provider. As someone who's in the position you're arguing against, I'll respond to a few of your points.
1) Redhat gave a perception of providing support, and many of people purchased their distributions based on this perception. Some would call this a bait-and-switch. It's causing mass confusion where I work.
2) Do you really use vi? Perhaps it's actually vim, or elvis, or another of the many vi clones. Likely vim, the default in most distros. Certainly not a hard-to-reach person like Bill Joy. Whether or not the person who wrote vim is on the list though, the packagers of the OS where the copy of vim that was installed are. They're fully capable of patching it if something goes wrong, and probably have the resources to send a patch to the developer.
3) Similar to your first response really, so I'll expand on this. When someone purchased a copy of Redhat, there was a perception that there would be support for a number of years afterward. That was part of what made it an attractive product. By making these changes, Redhat has made it obvious that they are no longer providing the product as perceived by the greater amount of their customers. Whether or not this is unethical, it promotes a perception of instability along with a considerable loss of good will. That's likely to make their market smaller.
6) This is true. This perception was not discouraged by Redhat until recently though. As for those people who made the mistake of believing that RH 7.3-9 would be supported for some time, I doubt they'll make this mistake again. In fact, they're likely to seek another vendor. I for one don't blame them.
In the end, Redhat's clarified their position, and so has Mandrake. I'm personally happier with where Mandrake ended up, and would be more likely to recommend their product at this point.
Slashdot Mirror
XFree86 and many other X environments use what's known as the X Shape Extension. You can have an application display in an arbitrary shape, xeyes being the canonical example.
I think what he was trying to explain is not that it's difficult to use, but that it's less likely to abuse. All you need to to to be infected by MyDoom under Windows is double-click on the file after it's extracted. If this were under Linux, UNIX-based systems really, you'd need one more step - basically, you give permission to the file to be executable. It's only harder to accidentally run something. The same holds true in MacOS X for example, and there aren't that many MacOS X users complaining that it's difficult to use.
Here's an exercise for you. Right click on a file under Windows >=NT4. Somewhere in there (depending on the version) there's a setting to change permissions on your file, typical DOS-derived permissions give you Read-only, System and Hidden. It's something like that under UNIX as well, indeed under KDE (a UNIX desktop environment) it looks nearly the same. There's just an extra option called Exec(ute), which can be specified differently for users and groups of users.
That being said, it's quite possible for such a virus to spread under Linux/UNIX. The difference is that the security model of UNIX-like systems, especially users/admins giving applications permission to execute, makes it more difficult for these viruses to spread. Aside from market share and the ability to run on multiple platforms, that's why viruses under UNIX-like systems tend to wither and go away rather than propagating far and wide.
It's sort of like the OS equivalent of seat belts and air bags. Think about it: if Windows required users to give permission to an application before executing it, instead of automatically executing programs which masquerade as documents, odds are there would be far fewer viruses being spread in this manner.
Odds are you still had libs from 3.1 loaded, and the symbols were interfering with the 3.2 apps. Killing off the old processes before starting the new ones, and probably an ldconfig if you're on Linux, is probably all that was needed. An init 1 certainly works too.
;-)
Nice nick BTW.
Sure, but a Linux user with enough intelligence to install and use a cross compiler wouldn't be stupid enough to write something like MyDoom. Unless, of course, their intention was to support SCO's court case with more line noise.
There are certainly more Visual Studio users than Windows developers using cross compilers on Linux by several orders of magnitude. You're probably looking at Vegas odds here.
Odds are that it's someone with a copy of Visual Studio in a Windows environment. A Linux user probably wouldn't be wasting their time with this. Not to say that it couldn't be a Linux user, anything's possible, but Linux users have less to gain from MyDoom than SCO does.
FWIW, it's not like SCO's going to lose any business from their website being down. The bad faith they've generated with their wild claims took care of that for them, most Linux users understand this. The only result is another false leg for SCO to stand on in their PR department, which the article's author fell for.
Perhaps you're right about me not understanding you. There is not only one application in KDE, as certainly as there is not only one application in GNOME. They're both a set of smaller applications, which can be mixed and matched with each other. I think that's great, if I don't like part of one I can use the other. If there's only one thing in the other that I like, I can still use it. There is no non-trivial application that is the best for everyone, and I seriously doubt that it's possible to create such an application.
...." consider emacs. It's open source, it's very large, is a set of applications written in lisp by its own admission and can do just about anything. Any vi (or derivative) user will tell you it's bloated. It's also 100% open source, GPL software originally written by Richard Stallman. There really was no reflection on KDE here, as it's a set of relatively lightweight apps for what they do. They're configurable and flexible, but in the end they just do what they do. Even konqueror is just a wrapper for various plugins, albeit a very flexible one.
;-)
As far as a "one massive, bloated but
Choice in its essence requires more than one out of which one can choose. That's what it means. What you're talking about is configurability. Don't misunderstand, I'm not trying to belittle your English. I'm merely stating my position, which you disagree with. The fact that you and I will not agree on this may say something about why there will probably never be one application that does what everyone wants, exactly as they want it to be done. If it truly is possible, then show me what it is that you believe I really want.
As for that last bit... I'm not sure how you got this impression, but the Linux kernel isn't all that big really. Most of the source that you download consists of drivers for the nearly infinite varieties of hardware that it supports. If Linux fails to meet my needs however, I'm free to boot another OS. Luckily this hasn't been an issue, but e.g. FreeBSD is always there if I someday need it. By the same token, if KDE somehow didn't meet my needs, I could use something like Windowmaker or Blackbox (both of which I occasionally use, esp. in low memory situations and in VNC.) XFCE is a great choice also, in fact I tried dozens of WMs and "desktop environments" before finally deciding that KDE was my preference. The beauty that I see in Open Source is that I have these choices. I don't see any beauty in censoring Linux in the manner you suggest, but do what you will.
Vive la difference!
It's not a fact, it's a question. Do we want a choice of applications, or do we want one massive, bloated but common set of applications that do everything that anyone might conceive that we want them to do? Like Windows and Office I suppose, but if that's all that you're looking for then a trip to Wal-Mart will fix that for you. I think that we have something better already.
A hammer might very well be able to break rocks, but I don't want to be the poor sap stuck doing that all day. By the same token, if an application isn't comfortable to me, then I'm going to pick another. The same can be said about desktops (the topic of this discussion.) People will argue all day about their preference of EMACS or vi, MacOS vs. Windows, etc. That's because they have a preference. Neither group would want their preferred application to become like the other.
What I like most about Linux is that I have a choice of just about anything without leaving the OS. That's what I want someone to see when they start using Linux. You're not stuck with what some guy in Redmond or Cupertino thinks you should use. If that's what this distribution is about, then it will come without the vitality and unique flexibility that is available with Linux. I can't and won't recommend that to anyone.
As a point of fact, both Qt and GTK can work without X on a framebuffer. Many commercial Linux-based PDAs use Qt on a framebuffer via the Qtopia environment. While that's neither here nor there, it does point out that the choice exists. In some cases, it simply makes more sense to use an alternative.
Whether or not you like choice, there are plenty of people who do. Many people can't even agree on pizza toppings, never mind what they want their desktop to work like. I personally don't think there's anything wrong with that, I'd rather people not dictate what desktop I can use. The concept of "best" is highly subjective, e.g. your vision of best practices in Open Source is different than mine.
You may remember that GNOME started out solely because KDE used a widget library (Qt) which at the time had a license with terms that some people didn't like. That was their choice, and they "wasted" a lot of effort coming up with a set of libraries and applications that did basically the same thing that KDE already did. In the end they have a desktop which suits their ideals, and that's fine. I prefer KDE for my desktop, but I support the ideals which lead GNOME to become what it is.
I do not support those who would state that I shouldn't have the choice to use KDE however. If a distribution decides that I should use GNOME and doesn't support KDE out of the box, then I'll simply not use it. Further, I won't recommend it to anyone else since I perceive it as being too restrictive.
To be honest, I don't find it very comforting that said distribution will tack on what users want as an afterthought either - they either will support these things, or they won't. If the latter is the case, then I don't believe they deserve to succeed. I didn't start using Linux to get another "One Microsoft Way", and I don't intend to go back now... I hope that this isn't the future for Linux.
Probably the latter. The ribbon is the connector which attaches the hard drive to the laptop. It's actually just an IDE drive, the power goes through the same connector as the data though. You can buy cables which plug into a standard HD power source and a 40 pin IDE port in a desktop PC for that matter. The enclosure of the drive can be opened on every Dell I've seen, from the Inspiron 3500 to the new Latitude D800 systems. The same is true for the IBM Thinkpads.
;-)
The hard drives in most laptops these days are identical. Apple uses IBM (now Hitachi) drives for that matter, I have a PB G4 open next to me rihgt now. (Getting the drive out of a PB G4 is a bit tricky though, heh.) I've actually taken a Mac drive and put it in a Dell laptop, it works fine. The reverse has also been the case as was noted here.
All this discussion about guilt is really silliness IMO. People should use the OS they are comfortable with, whatever that might be. I prefer Linux personally, but I know people who prefer OS X, Solaris, various BSD flavors. Even Windows. In the end, all that matters is that you have an environment which you enjoy working in. If it makes you happier to have gumdrop shaped widgets and a dead simple/nearly intuitive UI, then you owe it to yourself to get a Mac. If you want lots of current games, or need a lot of support from those around you, get a Windows box. If you want a genuine UNIX environment without all the extra fluff, pick a flavor of Linux, BSD or even Solaris that makes you happy. Just don't pick SCO
> They might be switching from GNU/Linux systems to something a little bit more suited to enterprise environments.
;-)
I don't agree about the "more" suited bit, but I'll skip the flamebait. The real question is, why would they do a thing like that? They're hardly an enterprise environment outside of litigation. Given their current business model, Linux is overkill.
Outside of a DDOS, they're probably just as well with PWS running off some Windows 98 box. More likely they're just hiding behind a proxy service like Akamai, as MS did in a similar situation. Imagine that, SCO using tactics borrowed from MS.
That's great, though I certainly wish the documentation had existed before we implemented our workarounds. When it was critical, it simply wasn't there without an NDA (which I'm not authorized to enter into.)
It's only a big lie if W2K is used for the server. There's no way in hades that I'm using a Windows box for a DC on my networks.
Back to the original point: Kerberos is a protocol which Microsoft has extended, placing their extensions under trade secret status to prevent others from benefiting from the work. That's the thing that they've actually done to irritate me.
We use a UNIX KDC, and won't be implementing a Windows DC. Since the Windows systems want to get their authorization info from an AD domain, based on Microsoft's extensions to the Kerberos protocol, we have to do some extra steps in order for our Windows systems to properly integrate into our network file system. It could have been much simpler had they not declared their extensions "trade secrets".
I'll hold you to that, for my part I won't let a Linux distro that doesn't have KDE touch my desktop/laptop anymore. I for one will admit to being a KDE zealot, Political Correctness be damned. I simply like it better. :-)
Those extra, proprietary, and incompatible Kerberos fields in W2k+ for one. I think they put those in just to make my life difficult. Of course, it wouldn't have mattered if they'd simply made their proprietary extensions open without having to sign some bizarre NDA...
That may well be, but the fact that Mandrake is coming out of bankruptcy at all says something about the company. Most companies in that position disappear well before it gets to that point.
SuSE is a nice product, but there's no downloadable ISO version. While that's given them a steady stream of revenue, it's also prevented more people from trying it and eventually becoming a customer. Mandrake's free ISO releases give them goodwill from their userbase, which tends to translate into brand loyalty over time. Keep in mind that there are plenty of people who like Windows simply because it came "free" with their computer, even if it does need to be wiped out and reinstalled every now and then.
SuSE is also significantly different from Redhat in many respects, whereas Mandrake was originally based on Redhat and has kept many of the filesystem conventions. This may seem insignificant, but it does simplify matters with e.g. binary RPMs for commercial software. My experience has shown Mandrake tends to take Redhat-targetted software with less effort than SuSE.
In any case, I don't buy the argument that there can be only one desktop-centric Linux distribution. The day that we're stuck with just one distribution is IMHO the day that Linux becomes another victim of Microsoft. I neither want nor believe that this will happen. I think that there's enough room for both SuSE and Mandrake, and hope that the competition between the vendors will continue to improve their quality and feature sets.
I don't see this as a problem. If a law enforcement official tomorrow were to confiscate your vehicle in the course of investigation because they suspected that you were involved in some sort of illegal practice, where would this leave you? Given that you only have one vehicle, and that you need the vehicle to work in order to live, wouldn't you say that you've suffered damages? For that matter, if you get the vehicle back in 2 years when they determine you're innocent, how useful is it to you then? Will you even have your job still?
Law enforcement needs to operate as efficiently as possible, but they also need to be certain that what they are doing is correct. If they aren't correct, then that should effect them in a negative manner so that they will make an effort to be correct. I still believe that people are innocent until proven guilty, and seizure of property without 100% certainty of wrongdoing is IMO ridiculous.
D'oh, never mind. After re-reading your comment, I noticed the "5 drives" and not "RAID 5 on 4". Still, level 5 isn't nearly as useful if you can't replace the drive quickly. I wonder how soon before we get hot swappable bays in a drive of this size? ;-)
It's probably using RAID level 0 striping if it's RAID at all. RAID level 5 gives you (n-1)*disk capacity where n is the number of disks. Thus, if it was RAID level 5 and you had 4 250GB drives, you'd come up with (4-1)*250GB, or 750GB capacity, before filesystem formatting. Of course, if any of the drives failed, you could swap it out and put another in its place. This case is hardly appropriate for drive swapping though, so odds are it's either level 0 or just concatenated disks.
;-)
As an aside, I've seen software RAID over Firewire in action. It isn't as reliable as one might hope for, and you'd better hope noone accidentally pulls out one of those easily removed cables
Why? Owning the rights to UNIX has been poison to the last few companies that had it. ;-)
I got fed up with the noise from my Athlon XP 2000+ a few months ago. It was so loud that it could easily have been mistaken for a vacuum cleaner. I was using a generic 6-fan case (4 in the front, one on the side and one at the rear) and a Cooler Master 6700rpm heat sink to keep the heat down.
Having heard good things about the Antec Sonata, I went out and picked one up. After a bit of looking into what heat sinks were available at the local Fry's, and basically just wanting anything less noisy than the Cooler Master, I picked up a Zalman CNPS7000A-Cu.
Currently, the Zalman is spinning at 1430 RPM (fwiw the P/S fan is spinning at 1834 RPM.) It can be adjusted to suit your cooling needs with an included adjustment unit, but I was going for quiet. The computer now runs about 10c cooler than before on the CPU, and perhaps 15c cooler on the MB according to the onboard sensors. Not bad, considering that both fans are running at a fraction of the speed from the previous case. The only issue I could see someone having with the heatsink is that it's so large that it doesn't fit on all motherboards.
While the Zalman helped, the Sonata has really made a significant difference in both the area inside AND around my computer. Antec built this case with what is quite possibly the best configuration for hard drives I've ever seen in a PC case: they're mounted sideways on individual HD trays. The drives are mounted to the tray using rubber grommets, which make the two 7200RPM drives in my case nearly silent. The sound of the drives accessing, which before was a rather loud chatter, is now nearly inaudible. There are 4 internal 3.5" HD bays, so I didn't have to give anything up for the silence. Better yet, due to the design of the trays the HD cables don't block the airflow of the case since they go directly to the back of the case. The case even looks nice, with a glossy black finish on the metal parts of the case and matte black plastic on the front.
For someone who wants to throw together a system and get nearly all the parts in one box, the Sonata is an excellent case to go with. The Zalman complements the case nicely, though if you have a preferred quiet heatsink you might be better off using that - the Zalman is not compatible with all motherboards due to its large size. I'm quite happy with it on my system though (using an Asus A7V266-E).
Ok, I'll bite again. This may not convince you of anything, but I will maintain that I disagree.
If Redhat had merely eliminated updates for current releases to non-paying customers, I could buy that what they'd done is perfectly ethical. What they have done though, is eliminate all of their current products and replace them with different products requiring higher initial and continuing investment than their previous products. They also changed the entire pricing of their software lineup. In this light, the TCO argument which is constantly used for Redhat Linux wears thin.
Due to Redhat's policy change, any company which has based their product on an existing release of Redhat needs to either upgrade to a different Redhat version, provide updates themselves, do no updates, or switch to another distribution. In the projects I'm involved with, doing no updates is not a solution. Upgrading all of our systems immediately is even less of a solution. I'm not banging on the "openness" drum here, I'm talking about a stable platform being yanked out from under our feet. We may need to go the Progeny route until things are worked out on some of our systems, money which Redhat could be making had they not dropped the ball. Against this backdrop, it's laughable to think that they want to compete with the likes of Sun Microsystems. They seem to have neither a clue nor a desire to know what their customers' needs are.
That many people want people to be given things for free is to be expected. People in general seem to have no qualms about "borrowing" and installing a copy of licensed software on multiple systems from what I've seen. These people don't care about how libre the system is, only how gratis. They're not going anywhere, and plenty of them are users of proprietary systems. Still, open source is not a business model. It's a development strategy. It exists not because businesses are paying for it, but because people want something better (or at all.) Seriously. If it benefits businesses, so much the better. That's generally not the primary reason for development though.
If a software product survives only because a small number of businesses pay staff developers to write it, then the software probably isn't interesting enough to a large number of people to continue its development. If it is, someone will come along and do it. It may or may not be a business, but if a closed source product is attractive/useful enough then someone's going to want a good open source version. If someone has an itch to scratch and there's no product out there that does what they want, there's a strong possibility that they'll open source it given the collaborative development benefits. Bug fixes/additional development returned to the initial developers may outweigh their initial development costs. The price to end users is simply a happy coincidence.
I don't believe Redhat should provide support in perpetuity without some sort of remuneration, but they should provide some form of extended support for those who are willing to pay for it. A refusal to do that for the systems of current customers where significant investment has been made shows that Redhat clearly isn't ready for the enterprise.
Surely redhat has no legal obligation to provide support to customers who have paid. I'll grant you that. The fact that a large set of their customers are upset however, should be a clue that an ethical obligation has not been met.
Redhat initially started with the product "Redhat Commercial Linux". They touted their distribution as a commercially supported variant, as opposed to what was then currently available (Yggdrasil, Slackware, etc.) Under the GPL, their legal obligations were nothing. Their ethical obligation however, the stated purpose for their existence, was to support their distribution in ways that non-commercial entities would not be able to.
If Redhat continually modifies their support policies as they have, I'd say that ethically they're shirking their duties. Legally they can tell people to get lost, since there wasn't a contract stating otherwise. Don't confuse ethics with legal obligations though. Both are important to the perception of a company's value, but they are not the same.
Amongst other things, I work with networked embedded systems a good deal. Our investment in Redhat is now under significant question. 1 year is nothing to a system which may be developed over a period of potentially several years, and updating the distribution every year changes the systems in significant ways (hardware requirements, timings, etc.) Redhat was for a while touting their product as ideal for embedded systems, but this has certainly changed. Their product is simply no longer useful to us.
I understand where you're coming from, but I believe that a company's actions should be rewarded or punished appropriately by the consumer. Redhat's actions are simply not in line with those of an enterprise-class software provider. As someone who's in the position you're arguing against, I'll respond to a few of your points.
1) Redhat gave a perception of providing support, and many of people purchased their distributions based on this perception. Some would call this a bait-and-switch. It's causing mass confusion where I work.
2) Do you really use vi? Perhaps it's actually vim, or elvis, or another of the many vi clones. Likely vim, the default in most distros. Certainly not a hard-to-reach person like Bill Joy. Whether or not the person who wrote vim is on the list though, the packagers of the OS where the copy of vim that was installed are. They're fully capable of patching it if something goes wrong, and probably have the resources to send a patch to the developer.
3) Similar to your first response really, so I'll expand on this. When someone purchased a copy of Redhat, there was a perception that there would be support for a number of years afterward. That was part of what made it an attractive product. By making these changes, Redhat has made it obvious that they are no longer providing the product as perceived by the greater amount of their customers. Whether or not this is unethical, it promotes a perception of instability along with a considerable loss of good will. That's likely to make their market smaller.
6) This is true. This perception was not discouraged by Redhat until recently though. As for those people who made the mistake of believing that RH 7.3-9 would be supported for some time, I doubt they'll make this mistake again. In fact, they're likely to seek another vendor. I for one don't blame them.
In the end, Redhat's clarified their position, and so has Mandrake. I'm personally happier with where Mandrake ended up, and would be more likely to recommend their product at this point.
Apologies to those who don't get the joke.