Actually, you can thank the grand ol' GPL for screwing up the nVidia drivers for Linux. Until they discovered a method that would allow them to follow all the third party software that nVidia uses in their drivers, they couldn't release drivers for Linux. It was until they discovered it was possible to create a binary module for Linux and still follow their contractual obligations that they even created the drivers.
And the nVidia drivers come with 90% of the performance they have under Windows. nVidia's Linux drivers are some of the best 3D/2D drivers available for Linux, and the cards all work basically without a hitch. So while you can whine about them not opening the source to their drivers, I'll continue to use my GeForce2 under XFree4 and continue to play QuakeIII and Unreal Memory Leak - I mean, Unreal Tournament - under Linux.
Bash them for being closed all you want, but you have to realize that unlike the people making XFree and the people making Apache, some nVidia technology is licensed from other companies. nVidia can't release the driver source, as it isn't all theirs. (Honestly, though, I doubt they would anyway - I know I wouldn't.)
And as far as I know, every current nVidia card is currently supported via their drivers for Linux. (Which, admittedly, are for x86 only - one of the disadvantages of binary only, but there isn't much to be done about it.)
So Linux support may not be as far fetched as you seem to believe.
When did the telephone become truly standard? When did the telephone move from a line for a town to a line per house? When did the telephone move from 34K to broadband?
If you're going to use the telephone as an example, than you should realize that the telephone really does exemplify what Dvorak's saying. When the telephone started, it was a line between two locations. In it's first useful implementation, there were actual human operators who used switchboards to connect a caller from one line to another. The telephone had to go a long way from to reach the 34K standard - and it took a while for it to go just like it'll take a while for broadband to come into common usage.
In 1877, there were no switchboards in use - there was practically no telephone infrastructure (they were building the infrastructure then, as the first public telephone service started in January 1878). If you wanted to talk to someone a ways away, you used telegraphs or the postal service. While Decemeber 1877 stands on the start of the telephone revolution, the telephone had a slow start - the first "modern" telephones became common around the 1950s - when rotary dialing and the dial tone finally became a common standard (a variety of other methods were used before then, including operators, existing rotary dialing, etc.) - but it was in the 1950s that it became possible to dial for even long distance calls. Although it wasn't until the 1960s before even that became universal.
Which means that telephones as almost all of us think about them didn't really come into existance until the 1960s - and those were rotary phones. Touch tone didn't come into play until around the 1980s, about the same time as cellphones came around.
Which means that although the telephone may have drastically changed people's lives, it didn't come into widespread usage until after the Great Depression (during which the President got a telephone) and didn't even reach what we'd consider "normal" until around the 1960s. So if you want to think of broadband as a new standard, and base it on the telephone, then Dvorak's right - standard broadband (dial tone / rotary dialing) is still a couple of decades off, while most people stick with the human operators or crank switches.
You don't need a fancy application for this stuff; IRC, FTP, HTTP...
Or, come to think of it, SSL. Since SSL is a Socket Layer and wraps around the actual protocol that goes through encrypted, no ISP could ever block it - most users would be completely pissed if they discovered that their ISP had broken every e-commerce site under the sun. And I think users would find out when they go to use that fancy new Internet bill paying service their bank is hyping and when they try to log on it doesn't work. After all, I currently pay my cellphone bill online, and I know that a lot of non-Slashdot types are getting used to the idea of doing secure transactions online. And almost all of those involves HTTPS - or SSL.
And using SSL, you could run ANY protocol under it and your ISP wouldn't be able to tell what you were doing.
Actually, you can link to it!
Here's the link. It's the URL of the article (http://slashdot.org/articles/01/05/05/1459212.sht ml in this case) followed by the comment number (57, as it turns out) used as a inner page reference (or whatever the W3C calls it).
When Slashdot creates it's HTML, it creates little anchors at the start of every single comment. You can actually do that in unarchived stories too... Strangly enough, when Slash archives posts, it DOESN'T change the little (#Comment ID) links to use this. I should submit a bug report about that...
None of which will be pleasing to the MS loyalists
Why would they care? Guess what - when an OS has 95%+ of the market or whatever the figure is (that sounds high), you have a statistically higher chance of being hit. Besides, I'll bet you that Linux would do just as poorly in the hands of the same people who had their machines zombified - how many people are capable of properly securing their machine? And how many people do you think would have been running and old, out of date copy of the OS anyway?
In fact, what he's saying is that by making Win2000 more like UNIX that Microsoft is making the entire Internet less secure. And in this case (making spoofing packets easier), I believe him.
To quote the article, "bullshit." Slashdot may be followed by many people, but most people I know take everything they read on Slashdot with a grain of salt. Or more.
Seriously, I almost never read just the article and assume that the little Slashdot writeup is correct. Assuming I have time, I read the comments. Generally speaking, if Slashdot makes a goof, the high-ranking comments will explain it. Something most industry 'zines can't do, since it's hard to add comments on dead trees in realtime.
So while there may be more people reading Slashdot, most are intelligent enough to think for themselves and know enough to question what they read. I don't believe everything I read on Slashdot, and I try and check out the facts as best I can. It's a skill that most "nerds" have, being able to recognize poor arguments and arguments that are lacking in facts. Any college student had better be able to structure an argument based on facts, and be able to recognize when an argument doesn't have the facts to back it up.
(For example, my whole argument here doesn't have any facts to back it up, hehe... so you should think for yourself in this case - does your experience match what I'm suggesting? Can you think of times when Slashdot posted wrong information? Can you think of times when the comments have corrected information in the article, and the article has later been updated based on comments?)
The bottom line is that most people who read Slashdot have learned not to blindly accept everything written on the site. For that respect, this review is "ok" - it's obviously detailing only Michael's experience, and it tells me that should I ever look into getting Myst III I'll need to be careful to ensure that my system will run the game - I got burned in the same way with Black & White, it wouldn't run on my machine without several patches to Win2K. And B&W's protection scheme also effects my ability to play - it takes literally around two minutes to decide the CD is valid on my machine.
So Michael's article, which should obviously be mostly Michael blowing off steam, is acceptable. This wouldn't be an acceptable print article, but in the context of Slashdot, where it's being used to promote discussion, it's fine.
The real problem with Linux desktops is that as far as I know, there is no really good desktop distribution. The best that I'm aware of is Mandrake but with the recent Mandrake shakeup, I'd be worried about going with them. I've used SuSE and while it's OK, it still doesn't seem as polished as Win2K. What would it take to make a nice desktop distribution, though? I'll give my thoughts on that below. These are just my opinions based on developing applications for both Windows and GNOME, and on using both Windows and GNOME. I haven't used KDE (recently, I used it on a SuSE box a year ago, so I'm a bit out of date with it's current state).
First, there needs to be a winner in the desktop war. Right now, neither KDE or GNOME can win the desktop war - KDE is out because it's centered around a GPLed library which removes commerical concerns (which is why there is an official GTK+ AOL Instant Messenger client but not a KDE one). GNOME is gaining wider commerical support since it's core LGPL libraries allow commerical apps to be built using them. And that is important for a desktop - trying to create a Free Software only desktop is an admirable idea, but if "world domination" is your thing, you have to be prepared to allow people to license on their own terms, something that the LGPL allows third-party application developers while the GPL does not. Not that it matters, since, having tried to write a GNOME application, I now know that the GNOME libraries suck. I'd assume that KDE is much easier to develop for, simply because Trolltech presumably ensures that Qt is easy to develop on, while GTK+ seems to be a case of the Open Source community refusing to rewrite anything, regardless of the need. Which means that a real Desktop Distribution would probably need to create a third desktop system to use to try and get around the problems of the existing systems.
Based on this, the second problem is a lack of a definition of an "application" and what goes where. Under Windows, your application installs into the "Program Files" directory which is nicely hidden from the user by default in Win98+ and Win2000+. There are known hooks to integrate an application with the shell - which is why WinZip can add a "Add to Zip" option to every folder shown via Windows Explorer. The Linux desktop really needs a graphical shell that does things right - Konquerer is really impressive in that respect (first Linux file browser I've seen that looks right, sorry gmc and Nautilus) but unfortunately runs into the third-party closed source app problem thanks to the viral GPL. Under Linux, though, applications often put the binary in/usr/bin, miscellaneous libraries in/usr/lib, and random data in/usr/share. Uninstalling an application becomes a nightmare because little things can hide themselves in miscellaneous directories.
The third thing that really needs to go is XFree86. While there is no question anymore that any new graphical environment must be usable over a network, XFree86 still has major issues with creating new drivers. Windows has tackled this problem, but there is still too much "server" code in what is supposed to be a driver. For a good example, the Render extension wasn't supported until recently on the nVidia drivers. Why should the video driver effect that? Maybe the video driver can offer acceleration of certain features, but it has to offer them natively? The other problem with XFree86 is that you can't easily configure video options while it's running. While Quake III demonstrates that it's apparently possible to make XFree86 set a specific resolution, it seems to be impossible to force a specific color depth without restarting the server. A good desktop would be able to change the color depth without requiring the user to close all open applications.
The final problem is really minor cosmetic problems starting the Linux kernel. Although based on Mandrake it is possible to create a splash screen while the Linux kernel is loading, that is a requirement for a desktop. Most modern PCs BIOS screens contain almost no useful information simply because it confuses the average user. Users are annoyingly inquisitive, they want to know what's happening when the text flys up the screen. Telling them "it's just loading, ignore it" doesn't stop them from wondering. Hiding it puts the user at ease.
These are just my thoughts on creating a good Linux desktop, based on having used Windows and GNOME. Yes, I've got an anti-KDE slant based on licensing issues - if you think I'm wrong about that, please argue from the point of view accepting that there will always be companies who want to develop closed source applications and do it cheaply. I'm aware that you can license Qt from Trolltech for cash, I'm not entirely sure about the rest of the core KDE libraries though.
Hmm, it's Sunday, so I suppose it would be a slow news day. But wow, was that a non-story. The story itself contained one sentance about the X-Box crashing, and apparently the X-Box sample that crashed wasn't the same hardware configuration as the final X-Box, namely lacking memory. So the game probably tried to address non-existant memory and crashed. Wowee.
Seriously though, this is really, really weak. Yeah, you could do a "Ha-ha" towards them, but this is really just purely immature. Wanna watch my Linux box crash if I tell the kernel it has 256MB of memory and only have 128MB (and then try and run Mozilla:-P)? Making fun of a pre-release verison crashing because it's not the same hardware is really foolish and immature.
I suppose the Nelson reference is appropriate - this article is about as mature as Nelson.
2) Michael, you do realize other languages exist in the US, right? They may not be official
I remember reading somewhere that there was no offical United States language, but I then decided I should check that fact and can't find any evidence for it...
However, I did find this webpage that lists languages spoken in the United States. From the site:
The number of languages listed for USA is 213. Of those, 176 are living languages, 2 are second languages without mother tongue speakers, and 35 are extinct.
Also of interest is this specific entry which lists about 1 million French Cajun speakers. So I guess that he hasn't really forgotten that he's in the United States after all...
Most of NASA's "work" is contracted out to industry, though. In fact, most government science and research is contracted out to either industrial groups or universities.
Of the research that government agencies actually do, most of it has to do with helping industry as opposed to raw science - and recent political and cultural trends are suggesting that government-sponsered science is about to stop, as more and more people question why the government should do science the private sector can do cheaply with more "pratical" gains.
I know it's easy to predict the death of one technology when another comes along, and (for example) it's clear that TV hasn't killed radio yet.
Not killed, maybe, but radically altered. Before TV, there were these radio shows on radio - they told stories, they were like audio-only TV shows. With the advent of the TV in many households, radio shows started to die (remember the song "Video Killed the Radio Star?"), and be replaced with something else radio did better than TV - music. Anything on the radio these days is basically some form of music - the exceptions are basically NPR, news radio stations, and radio call-in shows.
Radio is mostly used now as something in the background to listen to. Families used to spend the evenings together in front of their radio listening to stories - now, TV has replaced that use.
So new technologies may not "kill" an old technology, but they will radically alter them.
Yeah - that's the first thing I tried to do when my Ethernet adapter didn't work. Well, at least, I tried to surf the web... didn't work too well. (I'm actually being a little sarcastic - the card worked as expected under Win98 but under this little OS called "Linux" it wouldn't work - until I got the updated kernel module (or installed 2.4).)
A large part of the problems people call computer tech support for prevent them from getting on the web and searching USENET. If you're trying to find out why you can't seem to connect to your ISP, and they've determined that it's a problem with your computer and not their end, you're stuck with either calling the OEM's tech support or asking that geeky friend of yours.
Just because you're capable of trouble-shooting issues doesn't mean Suzy Secretary who's trying to do work at home can troubleshoot her modem and then go online to ask for help on USENET for this annoying problem with Word crashing two seconds after it loads. In order to get help on USENET, you're going to need to have an idea as to what the problem is - yeah, I've always been able to solve stuff by looking around the web a little or just playing with the broken software/hardware. It's unfair to expect your Average User to know enough to troubleshoot their own computer though. In theory, that's what tech support is there for.
The other thing I'd like to see done for Linux is more people porting the desktop apps to Windows so that those people currently in the Windows world can try out the applications that they'd be using in the Linux world.
Most people I know who are thinking about switching to Linux always start by dual-booting between Windows and Linux. And since they're so used to doing things the Windows way, they more often than not continue to use Windows as opposed to using Linux, and therefore never learn the Linux apps. Since more often then not they wind up going back to Windows to do some important task (like, say, write a paper for a college course, balance a checkbook, what have you) they then often remain in Windows for the duration of the day since rebooting really is a pain. (Unless Windows crashes and forces them to reboot:))
Before someone suggests it, WINE is not the solution since using WINE is hardly intuitive (in Windows, you double-click on the executable - being forced to wine --managed/mnt/winC/Program\ Files/Microsoft\ Office\word.exe in a terminal is not intuitive - unless the various File Browser programs can run Windows apps out-of-the-box via WINE).
If, on the other hand, you could use these free software alternatives in Windows, you could then try them out while running in Windows. You could learn them while keeping those applications you Absolutely Need. I think this could help people learn what they'd need to know in order to use the Linux desktops. They might be starting out on Windows, but when they learn all that apps they need to do their daily work, and eventually learn enough to just switch into Linux full-time.
I believe in this case he works a standard 40-hour week as well as being on-call. So he gets an extra 8 hours for having the cell-phone, working a "virtual" if you will 48-hour week. If we works more, he gets more pay, above the standard 40 hours.
At least, that's how I interpretted the comments. Doesn't sound like a "money grab" to me - sounds like a way to earn more money by being on call for the duration of a week.
On an OT note, what's with the deluge of mod points that seems to have hit/. users ?
I don't know - but I have a guess. First of all, I'd guess that all the new users that Slashdot has been generating (over 300,000 (!!) new accounts since I created mine - and that was a year and a half ago!!) - have suddenly become elligible to moderate. And so we have a flood of mod points.
Plus, it's possible Taco screwed with the moderation settings since I was a moderator on Thursday and I come back from the weekend to discover I'm a moderator today - maybe Taco's upped the frequency people get mod points.
When I installed Nautilus 1.0, I a) requested it not to install Medusa (it did anyway, thanks Eazel) and b) found that what it qualified as "idle" and I qualified as "idle" didn't match AT ALL.
I'd be writing some code or browsing the web or something, and then after about a minute of inactivity (size a page long enough and it can easily take over a minute to read all the text on it) medusa would start thrashing my harddrive. More annoying is that it would take about 20-30 seconds for medusa to stop once I did something again. And I never was able to figure out how to tell medusa not to do idle searches but only use the 4am cron job (at least, I think it installed a 4am cron job).
To be semi-ontopic, I've got Mandrake 8.0 running now (ReiserFS now:)) and medusa isn't installed on their default install (or at least, ps aux | grep medusa comes up with the grep medusa process) and I've found that Nautilus works just fine without since I honestly don't use Find File all that often - and if I did, it would be usually just be recursing $HOME and the entire directory structure!
Of course, YMMV with medusa but I agree - I found it to just be annoying.
I can only think of one way that doing that becomes legal under US law - the FBI was attacking a non-US citizen. Since the Russian was not a US citizen, he has no US rights under US law.
That explanation nicely ignores the fact that this happened in Russia and would therefore fall under Russian law. Which means that if Russian law allows the FBI to snoop on crackers, the FBI is in the clear, I guess... except I thought that US law said the FBI only had athority inside the US...
I kind of hope this becomes an incident simply because I can't think of any way that this is legal.
And this would be different from normal FBI policy how? Seems to me that the FBI always wants to be the ones with the power to do nasty things while preventing civilians from these rights - for example, Carnivore.
I'm pretty sure that civilians sniffing e-mail would be considered cyber-terrorism, but the gov't doing it isn't, since they have our best interests in mind, of course (sarcasm). This isn't the first time that law enforcement is doing dubiously ethical things in the name of protecting the people that they would prosecute private civilians for - just try wire tapping or bugging private property. You'll wind up in jail why the FBI is simply "doing their job."
In another way, though, this isn't necessarily a bad thing - if I knew I could trust every FBI member to use the authority in a way to ensure justice, I wouldn't mind knowing that law enforcement could randomly check up on various suspects. There are many powers the government has which they in theory have because they won't abuse them - why else do we allow the government to have a military in peace time? As long as this trust is never abused, these extra powers are generally a good thing - but I get the feeling that trust in the government is getting heavily abused these days. I guess I'm just cynical.
Actually, if you're releaseing binary-only modules, it does reak havoc since your 2.2.17 module won't insmod on a 2.2.18 kernel - meaning you have to supply binary versions for every possible kernel combo to keep insmod happy.
At least - that's my understanding of how modules work. As far as I know, they contain specific symbols that link them to one kernel compile at a time - which is why the NVIDIA kernel module is a small C module to interface with the actual binary module. You compile the small C module which just links in at compile time the actual NVIDIA binary-only module. That solves the kernel versioning issues - although there may be ways to use out-dated modules in newer builds (like a 2.2.17 module in 2.2.18, but definately not 2.2.17 in 2.4.3). But if there is, I don't know what it is.
Of course, in the perfect Open Source world, you can recompile the modules every new kernel, but if people want vendor Linux driver support, some sacrafices must be made...
No, this is much closer to the Plymouth Arco - a large building meant to hold many people. The Launch Arco isn't scheduled until about 2050, and besides, it requires those discs with the blue lines ("thrusters" or something, that according to SimCity 2000, stabilize the thing during earthquakes).
Unfortunately, you never were able to launch your Launch Arcos into space to colonize the moon...
Slashdot Mirror
And the nVidia drivers come with 90% of the performance they have under Windows. nVidia's Linux drivers are some of the best 3D/2D drivers available for Linux, and the cards all work basically without a hitch. So while you can whine about them not opening the source to their drivers, I'll continue to use my GeForce2 under XFree4 and continue to play QuakeIII and Unreal Memory Leak - I mean, Unreal Tournament - under Linux.
Bash them for being closed all you want, but you have to realize that unlike the people making XFree and the people making Apache, some nVidia technology is licensed from other companies. nVidia can't release the driver source, as it isn't all theirs. (Honestly, though, I doubt they would anyway - I know I wouldn't.)
And as far as I know, every current nVidia card is currently supported via their drivers for Linux. (Which, admittedly, are for x86 only - one of the disadvantages of binary only, but there isn't much to be done about it.)
So Linux support may not be as far fetched as you seem to believe.
When did the telephone become truly standard? When did the telephone move from a line for a town to a line per house? When did the telephone move from 34K to broadband?
If you're going to use the telephone as an example, than you should realize that the telephone really does exemplify what Dvorak's saying. When the telephone started, it was a line between two locations. In it's first useful implementation, there were actual human operators who used switchboards to connect a caller from one line to another. The telephone had to go a long way from to reach the 34K standard - and it took a while for it to go just like it'll take a while for broadband to come into common usage.
In 1877, there were no switchboards in use - there was practically no telephone infrastructure (they were building the infrastructure then, as the first public telephone service started in January 1878). If you wanted to talk to someone a ways away, you used telegraphs or the postal service. While Decemeber 1877 stands on the start of the telephone revolution, the telephone had a slow start - the first "modern" telephones became common around the 1950s - when rotary dialing and the dial tone finally became a common standard (a variety of other methods were used before then, including operators, existing rotary dialing, etc.) - but it was in the 1950s that it became possible to dial for even long distance calls. Although it wasn't until the 1960s before even that became universal.
Which means that telephones as almost all of us think about them didn't really come into existance until the 1960s - and those were rotary phones. Touch tone didn't come into play until around the 1980s, about the same time as cellphones came around.
Which means that although the telephone may have drastically changed people's lives, it didn't come into widespread usage until after the Great Depression (during which the President got a telephone) and didn't even reach what we'd consider "normal" until around the 1960s. So if you want to think of broadband as a new standard, and base it on the telephone, then Dvorak's right - standard broadband (dial tone / rotary dialing) is still a couple of decades off, while most people stick with the human operators or crank switches.
Nice solution... but you must have fun going to LAN parties :)
Or, come to think of it, SSL. Since SSL is a Socket Layer and wraps around the actual protocol that goes through encrypted, no ISP could ever block it - most users would be completely pissed if they discovered that their ISP had broken every e-commerce site under the sun. And I think users would find out when they go to use that fancy new Internet bill paying service their bank is hyping and when they try to log on it doesn't work. After all, I currently pay my cellphone bill online, and I know that a lot of non-Slashdot types are getting used to the idea of doing secure transactions online. And almost all of those involves HTTPS - or SSL.
And using SSL, you could run ANY protocol under it and your ISP wouldn't be able to tell what you were doing.
[Sig:] Before you ask, I already have a boyfriend and he's more of a man than you'll ever be.
Hmm.... I think I have to ask...
Actually, you can link to it! Here's the link. It's the URL of the article (http://slashdot.org/articles/01/05/05/1459212.sht ml in this case) followed by the comment number (57, as it turns out) used as a inner page reference (or whatever the W3C calls it).
When Slashdot creates it's HTML, it creates little anchors at the start of every single comment. You can actually do that in unarchived stories too... Strangly enough, when Slash archives posts, it DOESN'T change the little (#Comment ID) links to use this. I should submit a bug report about that...
Why would they care? Guess what - when an OS has 95%+ of the market or whatever the figure is (that sounds high), you have a statistically higher chance of being hit. Besides, I'll bet you that Linux would do just as poorly in the hands of the same people who had their machines zombified - how many people are capable of properly securing their machine? And how many people do you think would have been running and old, out of date copy of the OS anyway?
In fact, what he's saying is that by making Win2000 more like UNIX that Microsoft is making the entire Internet less secure. And in this case (making spoofing packets easier), I believe him.
Seriously, I almost never read just the article and assume that the little Slashdot writeup is correct. Assuming I have time, I read the comments. Generally speaking, if Slashdot makes a goof, the high-ranking comments will explain it. Something most industry 'zines can't do, since it's hard to add comments on dead trees in realtime.
So while there may be more people reading Slashdot, most are intelligent enough to think for themselves and know enough to question what they read. I don't believe everything I read on Slashdot, and I try and check out the facts as best I can. It's a skill that most "nerds" have, being able to recognize poor arguments and arguments that are lacking in facts. Any college student had better be able to structure an argument based on facts, and be able to recognize when an argument doesn't have the facts to back it up.
(For example, my whole argument here doesn't have any facts to back it up, hehe... so you should think for yourself in this case - does your experience match what I'm suggesting? Can you think of times when Slashdot posted wrong information? Can you think of times when the comments have corrected information in the article, and the article has later been updated based on comments?)
The bottom line is that most people who read Slashdot have learned not to blindly accept everything written on the site. For that respect, this review is "ok" - it's obviously detailing only Michael's experience, and it tells me that should I ever look into getting Myst III I'll need to be careful to ensure that my system will run the game - I got burned in the same way with Black & White, it wouldn't run on my machine without several patches to Win2K. And B&W's protection scheme also effects my ability to play - it takes literally around two minutes to decide the CD is valid on my machine.
So Michael's article, which should obviously be mostly Michael blowing off steam, is acceptable. This wouldn't be an acceptable print article, but in the context of Slashdot, where it's being used to promote discussion, it's fine.
First, there needs to be a winner in the desktop war. Right now, neither KDE or GNOME can win the desktop war - KDE is out because it's centered around a GPLed library which removes commerical concerns (which is why there is an official GTK+ AOL Instant Messenger client but not a KDE one). GNOME is gaining wider commerical support since it's core LGPL libraries allow commerical apps to be built using them. And that is important for a desktop - trying to create a Free Software only desktop is an admirable idea, but if "world domination" is your thing, you have to be prepared to allow people to license on their own terms, something that the LGPL allows third-party application developers while the GPL does not. Not that it matters, since, having tried to write a GNOME application, I now know that the GNOME libraries suck. I'd assume that KDE is much easier to develop for, simply because Trolltech presumably ensures that Qt is easy to develop on, while GTK+ seems to be a case of the Open Source community refusing to rewrite anything, regardless of the need. Which means that a real Desktop Distribution would probably need to create a third desktop system to use to try and get around the problems of the existing systems.
Based on this, the second problem is a lack of a definition of an "application" and what goes where. Under Windows, your application installs into the "Program Files" directory which is nicely hidden from the user by default in Win98+ and Win2000+. There are known hooks to integrate an application with the shell - which is why WinZip can add a "Add to Zip" option to every folder shown via Windows Explorer. The Linux desktop really needs a graphical shell that does things right - Konquerer is really impressive in that respect (first Linux file browser I've seen that looks right, sorry gmc and Nautilus) but unfortunately runs into the third-party closed source app problem thanks to the viral GPL. Under Linux, though, applications often put the binary in /usr/bin, miscellaneous libraries in /usr/lib, and random data in /usr/share. Uninstalling an application becomes a nightmare because little things can hide themselves in miscellaneous directories.
The third thing that really needs to go is XFree86. While there is no question anymore that any new graphical environment must be usable over a network, XFree86 still has major issues with creating new drivers. Windows has tackled this problem, but there is still too much "server" code in what is supposed to be a driver. For a good example, the Render extension wasn't supported until recently on the nVidia drivers. Why should the video driver effect that? Maybe the video driver can offer acceleration of certain features, but it has to offer them natively? The other problem with XFree86 is that you can't easily configure video options while it's running. While Quake III demonstrates that it's apparently possible to make XFree86 set a specific resolution, it seems to be impossible to force a specific color depth without restarting the server. A good desktop would be able to change the color depth without requiring the user to close all open applications.
The final problem is really minor cosmetic problems starting the Linux kernel. Although based on Mandrake it is possible to create a splash screen while the Linux kernel is loading, that is a requirement for a desktop. Most modern PCs BIOS screens contain almost no useful information simply because it confuses the average user. Users are annoyingly inquisitive, they want to know what's happening when the text flys up the screen. Telling them "it's just loading, ignore it" doesn't stop them from wondering. Hiding it puts the user at ease.
These are just my thoughts on creating a good Linux desktop, based on having used Windows and GNOME. Yes, I've got an anti-KDE slant based on licensing issues - if you think I'm wrong about that, please argue from the point of view accepting that there will always be companies who want to develop closed source applications and do it cheaply. I'm aware that you can license Qt from Trolltech for cash, I'm not entirely sure about the rest of the core KDE libraries though.
Seriously though, this is really, really weak. Yeah, you could do a "Ha-ha" towards them, but this is really just purely immature. Wanna watch my Linux box crash if I tell the kernel it has 256MB of memory and only have 128MB (and then try and run Mozilla :-P)? Making fun of a pre-release verison crashing because it's not the same hardware is really foolish and immature.
I suppose the Nelson reference is appropriate - this article is about as mature as Nelson.
I remember reading somewhere that there was no offical United States language, but I then decided I should check that fact and can't find any evidence for it...
However, I did find this webpage that lists languages spoken in the United States. From the site:
Also of interest is this specific entry which lists about 1 million French Cajun speakers. So I guess that he hasn't really forgotten that he's in the United States after all...
Of the research that government agencies actually do, most of it has to do with helping industry as opposed to raw science - and recent political and cultural trends are suggesting that government-sponsered science is about to stop, as more and more people question why the government should do science the private sector can do cheaply with more "pratical" gains.
Not killed, maybe, but radically altered. Before TV, there were these radio shows on radio - they told stories, they were like audio-only TV shows. With the advent of the TV in many households, radio shows started to die (remember the song "Video Killed the Radio Star?"), and be replaced with something else radio did better than TV - music. Anything on the radio these days is basically some form of music - the exceptions are basically NPR, news radio stations, and radio call-in shows.
Radio is mostly used now as something in the background to listen to. Families used to spend the evenings together in front of their radio listening to stories - now, TV has replaced that use.
So new technologies may not "kill" an old technology, but they will radically alter them.
A large part of the problems people call computer tech support for prevent them from getting on the web and searching USENET. If you're trying to find out why you can't seem to connect to your ISP, and they've determined that it's a problem with your computer and not their end, you're stuck with either calling the OEM's tech support or asking that geeky friend of yours.
Just because you're capable of trouble-shooting issues doesn't mean Suzy Secretary who's trying to do work at home can troubleshoot her modem and then go online to ask for help on USENET for this annoying problem with Word crashing two seconds after it loads. In order to get help on USENET, you're going to need to have an idea as to what the problem is - yeah, I've always been able to solve stuff by looking around the web a little or just playing with the broken software/hardware. It's unfair to expect your Average User to know enough to troubleshoot their own computer though. In theory, that's what tech support is there for.
Most people I know who are thinking about switching to Linux always start by dual-booting between Windows and Linux. And since they're so used to doing things the Windows way, they more often than not continue to use Windows as opposed to using Linux, and therefore never learn the Linux apps. Since more often then not they wind up going back to Windows to do some important task (like, say, write a paper for a college course, balance a checkbook, what have you) they then often remain in Windows for the duration of the day since rebooting really is a pain. (Unless Windows crashes and forces them to reboot :))
Before someone suggests it, WINE is not the solution since using WINE is hardly intuitive (in Windows, you double-click on the executable - being forced to wine --managed /mnt/winC/Program\ Files/Microsoft\ Office\word.exe in a terminal is not intuitive - unless the various File Browser programs can run Windows apps out-of-the-box via WINE).
If, on the other hand, you could use these free software alternatives in Windows, you could then try them out while running in Windows. You could learn them while keeping those applications you Absolutely Need. I think this could help people learn what they'd need to know in order to use the Linux desktops. They might be starting out on Windows, but when they learn all that apps they need to do their daily work, and eventually learn enough to just switch into Linux full-time.
At least, that's how I interpretted the comments. Doesn't sound like a "money grab" to me - sounds like a way to earn more money by being on call for the duration of a week.
I don't know - but I have a guess. First of all, I'd guess that all the new users that Slashdot has been generating (over 300,000 (!!) new accounts since I created mine - and that was a year and a half ago!!) - have suddenly become elligible to moderate. And so we have a flood of mod points.
Plus, it's possible Taco screwed with the moderation settings since I was a moderator on Thursday and I come back from the weekend to discover I'm a moderator today - maybe Taco's upped the frequency people get mod points.
You forgot Windows XP :-)
I'd be writing some code or browsing the web or something, and then after about a minute of inactivity (size a page long enough and it can easily take over a minute to read all the text on it) medusa would start thrashing my harddrive. More annoying is that it would take about 20-30 seconds for medusa to stop once I did something again. And I never was able to figure out how to tell medusa not to do idle searches but only use the 4am cron job (at least, I think it installed a 4am cron job).
To be semi-ontopic, I've got Mandrake 8.0 running now (ReiserFS now :)) and medusa isn't installed on their default install (or at least, ps aux | grep medusa comes up with the grep medusa process) and I've found that Nautilus works just fine without since I honestly don't use Find File all that often - and if I did, it would be usually just be recursing $HOME and the entire directory structure!
Of course, YMMV with medusa but I agree - I found it to just be annoying.
That explanation nicely ignores the fact that this happened in Russia and would therefore fall under Russian law. Which means that if Russian law allows the FBI to snoop on crackers, the FBI is in the clear, I guess... except I thought that US law said the FBI only had athority inside the US...
I kind of hope this becomes an incident simply because I can't think of any way that this is legal.
I'm pretty sure that civilians sniffing e-mail would be considered cyber-terrorism, but the gov't doing it isn't, since they have our best interests in mind, of course (sarcasm). This isn't the first time that law enforcement is doing dubiously ethical things in the name of protecting the people that they would prosecute private civilians for - just try wire tapping or bugging private property. You'll wind up in jail why the FBI is simply "doing their job."
In another way, though, this isn't necessarily a bad thing - if I knew I could trust every FBI member to use the authority in a way to ensure justice, I wouldn't mind knowing that law enforcement could randomly check up on various suspects. There are many powers the government has which they in theory have because they won't abuse them - why else do we allow the government to have a military in peace time? As long as this trust is never abused, these extra powers are generally a good thing - but I get the feeling that trust in the government is getting heavily abused these days. I guess I'm just cynical.
No one get any ideas now...
At least - that's my understanding of how modules work. As far as I know, they contain specific symbols that link them to one kernel compile at a time - which is why the NVIDIA kernel module is a small C module to interface with the actual binary module. You compile the small C module which just links in at compile time the actual NVIDIA binary-only module. That solves the kernel versioning issues - although there may be ways to use out-dated modules in newer builds (like a 2.2.17 module in 2.2.18, but definately not 2.2.17 in 2.4.3). But if there is, I don't know what it is.
Of course, in the perfect Open Source world, you can recompile the modules every new kernel, but if people want vendor Linux driver support, some sacrafices must be made...
And I still need to check if Linux does SMP on Alphas anyway - I seem to recall it does, but I'm not 100% sure.
Unfortunately, you never were able to launch your Launch Arcos into space to colonize the moon...