But if I were to try to make a decision about whether drillpress "A" is technically superior to drillpress "B", I'm simply out of my league.
Here's the thing: People who need to evaluate drillpresses generally take the time to do some research on the topic. In fact, they tend to actually be the "hard core hobbyists and professional drillpress operators" that you mention.
Schneir even makes the point that if he, as a security professional, can't make an informed decision on certain classes of products, how can the layman be expected to do better?
That is absolutely a valid point. In general though, I find that with computer topics people give up far too soon and suffer from their ignorance. Gamers who can't be bothered to figure out the difference between a "GeForce 8600 GT" and a "GeForce 8800 GTO". Managers who assume that they need to implement Microsoft Outlook and Microsoft Exchange just for basic email. Security Programmers who make comments like "RSA is more secure than AES because it can use a larger key size".
Right. Back before that work, writing cross platform GUI C applications was harder. But now - because of that effort - we have cross platform Gtk which everyone and their dog can use to write nice cross-platform GUI apps in C.
I was thinking "It has no password, therefore there is no password you can type that will let you log in" rather than "it has a null password, so you can get in without typing a password".
It's easy to be informed about your field, less so to be adequately informed about, say, 5 fields external to yours, and impossible to be informed about every field, even to some minimum degree.
This is true, but it's also an overused lame excuse for being ignorant.
Sure, you can't know everything - but that doesn't mean that you shouldn't know the difference between a bit and a byte, a chainsaw and a drill press, a cell and an organ, an oak tree and a pine tree, limestone and granite, diesel fuel and gasoline, an assault rifle and a machine gun, etc.
I think BitTorrent is a great idea, but in nearly every case it's slower than using the regular source. And I don't feel too guilty given Shuttleworth's wealth;-)
The speed of a bittorrent download is dependent on the speed of your connection. If your connection doesn't suck, it will be faster. If you're using 6 meg down / 128 k up DSL, bittorrent will be slow.
Don't look at it as "Ubuntu is the best thing since sliced bread, way better than Mandriva". Look at it as "Ubuntu is really nice, as is Mandriva. Ubuntu is just more popular because they've made their free-as-in-cost business model clearer."
How do you even start to generate stats like that? Market share for what?
You could generate meaningful stats for questions like "Market share of desktop PCs sold at Best Buy". Trying to generate stats for the whole "computer" market at once is probably a waste of time.
It is strongly recommended to not simply alter sources.list and do a dist-upgrade.
The supported upgrade method is to simply use Upgrade Manager (you know, the GUI tool). If you're upgrading a headless server, there are upgrade instructions at https://help.ubuntu.com/community/.
Of course it is not a flaw... software installation will always be a task that requires administrator privileges.
I'm going to have to agree that it is a flaw - just as much on Ubuntu as it is on Windows.
Some applications shouldn't need to be able to run arbitrary code as root to install - take a Tetris program for example, no need for it to do anything but copy its binary into/usr/games. Heck, it could even just drop its binary into/home/yourname/bin if there were a install-for-me-only mechanism.
Mac OS X actually allows this sort of thing with its drag & drop package mechanism - at least the single-user-installs part. I'm no Mac fan, but that feature is damn nice.
By default, there is no root password on Ubuntu. If you run the "passwd" command as root, you are setting a password, but there is no old password to replace.
Sorry to nitpick, but I don't want anyone thinking that Ubuntu has a default root password. That would be an ugly security issue.
The general concept of installing a package on a multi-user system involves running arbitrary third party code as root. This is true on Windows, Mac OS X, every GNU/Linux distro, and every Unix. You're right that this is dangerous from a security perspective - if you can get a user to install something, you've got root on their system. Some packages shouldn't need this, but many actually do - I think that Mac OS X is the only system that provides a built in mechanism to install packages that don't need root access without granting it.
Both Microsoft and Debian/Ubuntu have settled on the same "solution" to this problem: Cryptographically sign install packages. This means that the user has to trust everyone who can sign packages. For Windows, signing packages requires getting a developer signing key that costs a couple thousand dollars. For Debian/Ubuntu, the packages are signed when they are put in the official repositories - so there's a little bit more control since packages are stored centrally.
Either way, you can still install unsigned packages. but that's obviously a risk - you have to figure out if you trust the source enough to run arbitrary code as root on your own.
Ubuntu as a distribution fits into the software ecosystem as a parasite. It gives back very litte in terms of original new code or projects and weakens the push for Free software.
I can see picking on Red Hat or SuSE for this, but Ubuntu?
The only non-free software that Ubuntu packages by default are drivers. Proprietary drivers are extremely obnoxious, but they're not as bad from a ideological perspective as other OS components or applications because they don't create lock-in. If you have an Intel wireless card it'll work under Ubuntu with a warning. Later, you can switch to a Realtek card and you'll have precisely the same functionality with free drivers.
Even Richard Stallman agrees that you can't eliminate all proprietary software in a single step, as indicated by the existence of the LGPL. As Ubuntu becomes more popular through the "it works if it can, but complains about non-free drivers", the Ubuntu project will be able to put more pressure on device manufacturers to actually release their specs. This is good for everyone.
All of the things that work with free drivers in other distributions also work with free drivers in Ubuntu. Ubuntu provides the additional feature that computers with poorly chosen hardware can also be used for a while before you replace them with better supported systems.
You have to remember that one of the main goals of the Ubuntu project is to allow poor people to use recycled computers. Sometimes those recycled computers have hardware that can be best supported by a binary blob - in those cases the Ubuntu model of using the needed blob but alerting the user that it cannot be supported actually introduces people to the Free Software philosophy and allows them to use a Free Software system in a case where they would get no computer or Windows 98 otherwise.
Yes, you could say that it would be better if the Ubuntu project tried to pressure all the hardware vendors into releasing hardware specs / free drivers right now. They're doing that to some extent with the proprietary driver popup warnings. But... an OS needs popularity before it can exert pressure, and the Ubuntu developers have decided that they're going to try to maximize their impact later by compromising slightly now.
As for the inclusion of Java, it's in the Multiverse repository - equivalent to Debian's non-free. It's defiantly no more interesting than having a package for Macromedia Flash Player or Realplayer - and most distros do. Multiverse isn't enabled by default.
I'm actually something of a Java fan, but the cross-platform aspect is way over emphasized. It's reasonably straightforward to write programs that can be built anywhere, and then to provide platform-specific binaries for everyone. Look at applications like Gaim and The Gimp - they work wonderfully on Windows and any sort of Unix-ish environment.
Yea, it'll be annoying to use some desktop apps with a stylus / virtual keyboard interface, but I think you're overemphasizing the platform difference. We're talking about the sort of screen resolution that many Debian apps were designed for (800x480), and the stylus is basically the same as a mouse. Yea - the keyboard thing is annoying, but many GUI apps are designed to allow mouse-only operation - especially those that aren't specifically intended for textual input.
If the 6600 is really available for $40 now, that's not bad. I wouldn't recommend that for this year's games, but for years-old games like Doom III it'd be fine. As for your 8800 prices, I was aware of those. I still doubt you'll find a "GeForce 8800 GTX" (the card I mentioned) for less than $500 new.
"Media producers" want control over "their" video content. Most web browsers already do support [object src="foo.mp4"] tag syntax, but people won't use it because it would allow users to download the video content and save it.
I have no idea why you are hyping your point of view so strongly.
For the same reason I don't lend my Season 2 DVDs of a TV show to someone before I've lent them Season 1. For the same reason I won't let someone watch the Lord of the Rings movies for the first time on a 13" TV if they have access to a 60" TV with decent speakers.
Sure, for some games the visual presentation isn't the point. If you want to play Counterstrike: Source on a $150 Dell, that's not my problem - you're not going to miss anything that you can't experience later if you decide to upgrade. For some games though, the ability to create atmosphere is a key element to the game, and you can only get the full effect if you play through it the first time on a decent system.
My best example is probably Doom III. Lots of people said that Doom III sucked, the flashlight mechanic was lame, and the game play was boring and repetitive. Later, they would reveal that they tried to play it in a well-lit room with glare on their monitor, with shadows turned off so they could get a decent frame rate, and with frequent distractions so they couldn't get immersed. The whole point of the game was the horror movie atmosphere it created, obviously the game was going to suck if you didn't let it do that.
Even for Doom III, the difference between playing it on a $90 GeForce 6600 and a $500 GeForce 8800 GTX today isn't that big a deal. Either is enough to display the graphical quality the game needs to create its atmosphere. But if you try to play a game like that on a card that *can't* display shadows and "Medium" textures, you're not just wasting the money the game cost, you're wasting your opportunity to ever play that game for the first time the way the developers intended.
I have no idea why people on Slashdot hype your point of view so strongly. It might be to make up for the fact that you really want to get a decent graphics card, but you just can't handle the idea of spending $200 "just" on entertainment.
Yes, most games probably will run on a cheap graphics card. At least at low resolution, with most of the graphics options turned off. All of these games will run better on a nicer card. In order to see games as the developers intended (i.e. shadows on in Doom III, HDR lighting in Oblivion) you really need to have a card that would have been in the $200+ bracket when the game came out.
I'm sure that you'll respond disagreeing with me - whatever, you can have crappy looking games if you want. For me, games are more enjoyable with the visual experience that the developers spent so much effort creating.
For anyone who is actually considering buying a card - see if you can find a way to see the difference on real computers with real games before you buy. Based on my experience actually using cards at the different levels for many years, you'll see a *huge* jump from a sub-$100 card to a $150 card, a significant jump from a $150 card to a $250 card, and a minor jump from a $250 card to a $400 card (unless you're playing on a silly 30" display - then it's a major jump again).
Mepis and Pioneer are minor players. There are a lot of minor players, and many of them fall in the "having a bunch of players in the same niche is redundant" category. I'm not going to try to judge individual minor distros, but I will say that for most users a major distro is probably a better choice.
OpenSuSE (like Fedora) is a "demo version" of SuSE Linux. I see no reason to consider it separately from SuSE Linux - it doesn't stand alone the way something like Debian does, and it doesn't allow seamless transitions between payed commercial support and community support the way Ubuntu does.
CentOS is really interesting. It's a reaction to the fact that some applications are targetted at Red Hat Enterprise Server - and it's 100% appropriate for running those applications in cases where Red Hat paid support is not needed. It's still only interesting as an alternative to RHES, and so I see no advantage to considering it as a separate distribution.
Simple desktop distros: Ubuntu and Mandriva.
Enterprise-priced server support: Red Hat and SuSE.
Community supported for techies: Debian and Gentoo.
Localized in Chinese: Red Flag.
I wouldn't really consider any of the other distros to be "major" (ignoring non general-purpose PC platforms).
Mandriva is a legitimate zero cost commercially supported desktop Linux distribution. There is only one other distribution in that category: Ubuntu. Having a bunch of distros in the same niche would be redundant, but having two is a good thing. Mandriva is definatly one of the major players, and they have been for a very long time.
Slashdot Mirror
Here's the thing: People who need to evaluate drillpresses generally take the time to do some research on the topic. In fact, they tend to actually be the "hard core hobbyists and professional drillpress operators" that you mention.
That is absolutely a valid point. In general though, I find that with computer topics people give up far too soon and suffer from their ignorance. Gamers who can't be bothered to figure out the difference between a "GeForce 8600 GT" and a "GeForce 8800 GTO". Managers who assume that they need to implement Microsoft Outlook and Microsoft Exchange just for basic email. Security Programmers who make comments like "RSA is more secure than AES because it can use a larger key size".
Right. Back before that work, writing cross platform GUI C applications was harder. But now - because of that effort - we have cross platform Gtk which everyone and their dog can use to write nice cross-platform GUI apps in C.
That's a better way of phrasing it.
I was thinking "It has no password, therefore there is no password you can type that will let you log in" rather than "it has a null password, so you can get in without typing a password".
This is true, but it's also an overused lame excuse for being ignorant.
Sure, you can't know everything - but that doesn't mean that you shouldn't know the difference between a bit and a byte, a chainsaw and a drill press, a cell and an organ, an oak tree and a pine tree, limestone and granite, diesel fuel and gasoline, an assault rifle and a machine gun, etc.
The speed of a bittorrent download is dependent on the speed of your connection. If your connection doesn't suck, it will be faster. If you're using 6 meg down / 128 k up DSL, bittorrent will be slow.
Don't look at it as "Ubuntu is the best thing since sliced bread, way better than Mandriva". Look at it as "Ubuntu is really nice, as is Mandriva. Ubuntu is just more popular because they've made their free-as-in-cost business model clearer."
How do you even start to generate stats like that? Market share for what?
You could generate meaningful stats for questions like "Market share of desktop PCs sold at Best Buy". Trying to generate stats for the whole "computer" market at once is probably a waste of time.
It is strongly recommended to not simply alter sources.list and do a dist-upgrade.
The supported upgrade method is to simply use Upgrade Manager (you know, the GUI tool). If you're upgrading a headless server, there are upgrade instructions at https://help.ubuntu.com/community/.
Unfortunately, there's a kernel update, so Ubuntu will want to reboot after the update from Edgy to Feisty.
I'm going to have to agree that it is a flaw - just as much on Ubuntu as it is on Windows.
Some applications shouldn't need to be able to run arbitrary code as root to install - take a Tetris program for example, no need for it to do anything but copy its binary into /usr/games. Heck, it could even just drop its binary into /home/yourname/bin if there were a install-for-me-only mechanism.
Mac OS X actually allows this sort of thing with its drag & drop package mechanism - at least the single-user-installs part. I'm no Mac fan, but that feature is damn nice.
By default, there is no root password on Ubuntu. If you run the "passwd" command as root, you are setting a password, but there is no old password to replace.
Sorry to nitpick, but I don't want anyone thinking that Ubuntu has a default root password. That would be an ugly security issue.
The general concept of installing a package on a multi-user system involves running arbitrary third party code as root. This is true on Windows, Mac OS X, every GNU/Linux distro, and every Unix. You're right that this is dangerous from a security perspective - if you can get a user to install something, you've got root on their system. Some packages shouldn't need this, but many actually do - I think that Mac OS X is the only system that provides a built in mechanism to install packages that don't need root access without granting it.
Both Microsoft and Debian/Ubuntu have settled on the same "solution" to this problem: Cryptographically sign install packages. This means that the user has to trust everyone who can sign packages. For Windows, signing packages requires getting a developer signing key that costs a couple thousand dollars. For Debian/Ubuntu, the packages are signed when they are put in the official repositories - so there's a little bit more control since packages are stored centrally.
Either way, you can still install unsigned packages. but that's obviously a risk - you have to figure out if you trust the source enough to run arbitrary code as root on your own.
I can see picking on Red Hat or SuSE for this, but Ubuntu?
The only non-free software that Ubuntu packages by default are drivers. Proprietary drivers are extremely obnoxious, but they're not as bad from a ideological perspective as other OS components or applications because they don't create lock-in. If you have an Intel wireless card it'll work under Ubuntu with a warning. Later, you can switch to a Realtek card and you'll have precisely the same functionality with free drivers.
Even Richard Stallman agrees that you can't eliminate all proprietary software in a single step, as indicated by the existence of the LGPL. As Ubuntu becomes more popular through the "it works if it can, but complains about non-free drivers", the Ubuntu project will be able to put more pressure on device manufacturers to actually release their specs. This is good for everyone.
All of the things that work with free drivers in other distributions also work with free drivers in Ubuntu. Ubuntu provides the additional feature that computers with poorly chosen hardware can also be used for a while before you replace them with better supported systems.
You have to remember that one of the main goals of the Ubuntu project is to allow poor people to use recycled computers. Sometimes those recycled computers have hardware that can be best supported by a binary blob - in those cases the Ubuntu model of using the needed blob but alerting the user that it cannot be supported actually introduces people to the Free Software philosophy and allows them to use a Free Software system in a case where they would get no computer or Windows 98 otherwise.
Yes, you could say that it would be better if the Ubuntu project tried to pressure all the hardware vendors into releasing hardware specs / free drivers right now. They're doing that to some extent with the proprietary driver popup warnings. But... an OS needs popularity before it can exert pressure, and the Ubuntu developers have decided that they're going to try to maximize their impact later by compromising slightly now.
As for the inclusion of Java, it's in the Multiverse repository - equivalent to Debian's non-free. It's defiantly no more interesting than having a package for Macromedia Flash Player or Realplayer - and most distros do. Multiverse isn't enabled by default.
I'm actually something of a Java fan, but the cross-platform aspect is way over emphasized. It's reasonably straightforward to write programs that can be built anywhere, and then to provide platform-specific binaries for everyone. Look at applications like Gaim and The Gimp - they work wonderfully on Windows and any sort of Unix-ish environment.
Yea, it'll be annoying to use some desktop apps with a stylus / virtual keyboard interface, but I think you're overemphasizing the platform difference. We're talking about the sort of screen resolution that many Debian apps were designed for (800x480), and the stylus is basically the same as a mouse. Yea - the keyboard thing is annoying, but many GUI apps are designed to allow mouse-only operation - especially those that aren't specifically intended for textual input.
You're right. There is no comparison. ARM is a fully supported Debian platform, and there are definately more Debian apps than Windows Mobile apps.
If the 6600 is really available for $40 now, that's not bad. I wouldn't recommend that for this year's games, but for years-old games like Doom III it'd be fine. As for your 8800 prices, I was aware of those. I still doubt you'll find a "GeForce 8800 GTX" (the card I mentioned) for less than $500 new.
"Media producers" want control over "their" video content. Most web browsers already do support [object src="foo.mp4"] tag syntax, but people won't use it because it would allow users to download the video content and save it.
For the same reason I don't lend my Season 2 DVDs of a TV show to someone before I've lent them Season 1. For the same reason I won't let someone watch the Lord of the Rings movies for the first time on a 13" TV if they have access to a 60" TV with decent speakers.
Sure, for some games the visual presentation isn't the point. If you want to play Counterstrike: Source on a $150 Dell, that's not my problem - you're not going to miss anything that you can't experience later if you decide to upgrade. For some games though, the ability to create atmosphere is a key element to the game, and you can only get the full effect if you play through it the first time on a decent system.
My best example is probably Doom III. Lots of people said that Doom III sucked, the flashlight mechanic was lame, and the game play was boring and repetitive. Later, they would reveal that they tried to play it in a well-lit room with glare on their monitor, with shadows turned off so they could get a decent frame rate, and with frequent distractions so they couldn't get immersed. The whole point of the game was the horror movie atmosphere it created, obviously the game was going to suck if you didn't let it do that.
Even for Doom III, the difference between playing it on a $90 GeForce 6600 and a $500 GeForce 8800 GTX today isn't that big a deal. Either is enough to display the graphical quality the game needs to create its atmosphere. But if you try to play a game like that on a card that *can't* display shadows and "Medium" textures, you're not just wasting the money the game cost, you're wasting your opportunity to ever play that game for the first time the way the developers intended.
I have no idea why people on Slashdot hype your point of view so strongly. It might be to make up for the fact that you really want to get a decent graphics card, but you just can't handle the idea of spending $200 "just" on entertainment.
Yes, most games probably will run on a cheap graphics card. At least at low resolution, with most of the graphics options turned off. All of these games will run better on a nicer card. In order to see games as the developers intended (i.e. shadows on in Doom III, HDR lighting in Oblivion) you really need to have a card that would have been in the $200+ bracket when the game came out.
I'm sure that you'll respond disagreeing with me - whatever, you can have crappy looking games if you want. For me, games are more enjoyable with the visual experience that the developers spent so much effort creating.
For anyone who is actually considering buying a card - see if you can find a way to see the difference on real computers with real games before you buy. Based on my experience actually using cards at the different levels for many years, you'll see a *huge* jump from a sub-$100 card to a $150 card, a significant jump from a $150 card to a $250 card, and a minor jump from a $250 card to a $400 card (unless you're playing on a silly 30" display - then it's a major jump again).
Slackware is extremely popular among Slackware users, but I'm not sure that there's really any reason for a new user to adopt Slackware.
What advantage does Slackware really have over one of FreeBSD, Debian, or Gentoo?
Mepis and Pioneer are minor players. There are a lot of minor players, and many of them fall in the "having a bunch of players in the same niche is redundant" category. I'm not going to try to judge individual minor distros, but I will say that for most users a major distro is probably a better choice.
OpenSuSE (like Fedora) is a "demo version" of SuSE Linux. I see no reason to consider it separately from SuSE Linux - it doesn't stand alone the way something like Debian does, and it doesn't allow seamless transitions between payed commercial support and community support the way Ubuntu does.
CentOS is really interesting. It's a reaction to the fact that some applications are targetted at Red Hat Enterprise Server - and it's 100% appropriate for running those applications in cases where Red Hat paid support is not needed. It's still only interesting as an alternative to RHES, and so I see no advantage to considering it as a separate distribution.
Simple desktop distros: Ubuntu and Mandriva.
Enterprise-priced server support: Red Hat and SuSE.
Community supported for techies: Debian and Gentoo.
Localized in Chinese: Red Flag.
I wouldn't really consider any of the other distros to be "major" (ignoring non general-purpose PC platforms).
Mandriva is a legitimate zero cost commercially supported desktop Linux distribution. There is only one other distribution in that category: Ubuntu. Having a bunch of distros in the same niche would be redundant, but having two is a good thing. Mandriva is definatly one of the major players, and they have been for a very long time.