Rather than learning from this (as IBM and SGI seem to) they've gone the cynical, rationalist way.
I agree with all of your post except the bit about SGI learning from their mistakes. Last I heard, SGI had finished going through the meat grinder and was simply lying limply on the coutertop. Has this changed?
iTunes DRM is WEAK, man. Burn it to CDRW and rip the sucker again, it's as easy as jumping over a subway turnstile. Why are we wasting time with a pointless thing like this, why not crack WMP or something harder with a better payoff?
Because the method you are advocating is lossy, and there are a lot of Linux-using folks who might like to use iTunes.
Mmm...I dunno. Keep in mind that Apple spends marketing money and at least some money running, administering, and expanding the service. Even if the service is break even, it's better than losing money.
I suspect that you're right that the RIAA (well, member publishing companies) make out like bandits on the deal.
Argh. No. Not transparent proxies, unless you're providing an opt-in mechanism (say, a value-added service) to do so. It's *so* frusterating when ISPs start mucking with my network connection (firewall incoming and outgoing SMTP or transparently proxy it, block outgoing DNS, etc).
ISPs bitch about how they can't provide tiered service. Spam blocking, popup blocking, firewalling, are *all* great things to toss in value-added packages to provide tiered service. They, however, drive many people (who don't *want* said "service") to tears.
Another option would be a opaque proxy and to provide a mini application that sets it up.
Much as this drives me nuts as well, think about it. If there's an article on how a building in NYC collapsed in the Times which says that the cause was "structural failure in the concrete", the average reader just doesn't want more detail. Matter of fact, this is probably already more than they want to know. If it was actually the rebar at fault, it's close enough.
So, let's look at the myths that you're complaining about:
DeCSS is used to make copies of DVDs.
The most popular use of DeCSS and ilk has been to allow ripping and recompressing movies into a smaller, home-broadband-friendly format, which can then be redistributed. Short of explaining DiVX and CD burning, this is not an unreasonably far off explanation of what DeCSS is for Joe Public. It gives him a basic idea of what the thing does. It would be atrocious to try to feed this to a judge in a case or a techie, but the same goes for the rebar/concrete above.
Kazaa is an illegal music sharing site.
Also irritating. However, the majority of people out there know about email and the Web, and that's it. As a matter of fact, thanks to Microsoft using "Internet" to refer to the default web browser in XP, this is getting even more irritating -- "do you want to use email or the Internet?". If you don't want to explain the concept of other clients (which takes forever in *person*, much less in a short and snappy news article), this is reasonably on the ball. The "illegal" is pushing the limit -- I'm not entirely comfortable with that. However, it *is* probably true that over 99% of the traffic on Kazaa is illegal.
The music industry has said that "xyz" is legal/illegal....
I agree entirely with you here. It's poor journalism not to also provide an opposing view if you're presenting claims like this, though at *least* they qualify it with "music industry says" or "music industry claims".
An "Internet Virus" has been loosed...
Also irritating, but mostly to advocates of Linux or other mail clients. Few people know what "malware" or "worm" or "trojan" mean, or care. To them, the term "virus" means what "malware" does to you or me. Most people don't really understand or care about spreading mechanisms, other than the easiest method for them to avoid getting screwed over. Most people also use Windows and Windows alone. "Internet Virus" is not unreasonable to describe an Outlook worm in this context.
mp3s are illegally copied music files that Internet users share.
I don't think I've ever seen things put quite this harshly. I wouldn't be surprised to see "Many are illegal..." or "...frequently illegal...", though.
Basically, you *have* to take the context of your audience into account. First, the press is not unbiased. There are newspapers that are liberal, conservative, favor one politician or law over another, etc. It's a lot better than Soviet-era journalism, though. Second of all, if every time someone said "God" in a US newspaper, I suspect that few people feel the need for a Christian god qualification. People simply know what is meant. Sure, there's probably some theological inaccuracies that show up, just as there are technical inaccuracies here, but such is life.
This may seem awfully silly, and it doesn't mean that there's no good reason to backup DVDs. But, for the love of God, it *has* to be easier to just elevate the DVD player so that it's out of his reach. He can do a lot more damage than just to the DVDs -- when I was a child, I grated a bar of soap through the ventilation holes in my parents' stereo system.
The case wasn't taken to the Norwegian Supreme Court, so no legal precedent was made (presumably the reason why it didn't go to the SC). I may have misunderstood things (in the best/. tradition IANANL) but I don't think so.
IANANL (I Am Not A Norwegian Lawyer), but I'm strongly suspecting that the parent poster is correct.
If any prescedent was set, it is presumably that writing and distributing software to bypass copy protections are legal in Norway. This would put Norway smack dab where the United States was legally before the advent of the DMCA. It's legal to write and distribute the tools, but the actual act of copying DVDs is still copyright infringement.
From a legal standpoint, for all but a handful of software developers and security folks, this is not an issue.
It's a shame that Slashdot stories so frequently contain reasoning and conclusions (which may be false) or opinions (which may be inflammatory, misleading, etc). I wish Slashdot editors would simply dump stories that contain stuff like this -- this is what the comment section is for.
Misconception 1 - American programmers are better.
I don't think that this is what folks are really thinking.
Right now, the Indian IT and software development industry has been growing at an insane rate. Demand has been doubling and doubling and doubling. This isn't a criticism of Indians per se. It's just a fact that when you have a market like that, you're going to wind up employing bad along with the good, because it's easy for anyone to get a job. The same thing happened in the US during the dot-com boom. There was a huge amount of overdemand for web designers, so there were a huge number of incompetent web designers running around. In ten years, everything will probably be fine -- graduates will increase to meet demand, poor employees will develop a poor reputation, and so on. However, at the moment, I'd say that it's a reasonable assumption to say that you're likely to get worse work out of a typical Indian software house than a typical American software house at the moment. That doesn't mean that you can't get a good Indian developer, just that it's easier to get a worse one.
Second of all, the Indian software development industry is younger than the American one, and shooting up to meet demand. It's already hard to avoid godawful consultants in the US. When you go to an environment where reputations may not have been established for as long, things just get more difficult.
Misconception 2 - Indians are not innovative.
Sure.
The problem is the reasons most US companies are moving overseas. This isn't Wolfram running out and looking for a new brilliant lead computer scientist (well, they might well hire an Indian, but that's irrelevant -- this sort of behavior isn't forming the bulk of the recent surge). A lot of jobs going overseas are for ordinary drudge work -- web dev, basic system administration. These companies are looking for folks that don't cost $80K/year to write ordinary code. They're tired of paying obscene prices for work that's pretty straightforward. My guess is that this is where the "uninteresting jobs going overseas" idea comes from.
Finally, one person's anecdotal evidence doesn't count for much, but I've had only awful experiences with both India-based IT and India-based software development. There are clearly currently at least some really awful people out there. My guess is that in a few years, as demand and supply equalize, the situation will improve. However, please understand that not all American dislike of outsourcing simply stems from a sense of job threat.
The final issue, which is simply a problem for any international company, is that it can be terribly difficult to understand a thickly accented voice, especially once it's been crammed through the lossy compression that cell phones apply to voices. I'm sure the reverse applies (and hence a US-based programmer might be less desireable to an operation which is primarily India-based). However, it really can be frusterating to have to ask someone to repeat what they've said three or four times. They may show admirable patience in being willing to repeat what they're saying, but it's still a frusterating communications problem. This is especially so during meetings, where it's embarrassing for everyone involved if someone has to be asked to repeat themselves multiple times. Note that I don't see this as much of a problem with email. Even with poor English (and I've seen awfully good written English from many Indians with strong accents), I can usually make out what's being said without trouble.
On the other hand, I'm also sensitive to the fact that the US is edging uncomfortably closer to the old UK anti-Indianism due to labor issues, and I try not to be unfair.
As an aside, I'm also a bit saddened by the fact that there seem to be awfully few India-based open source projects. Perhaps open source springs simply from affluence, and hasn't been as available to many Indians as it has been western European and US-based developers (particularly Scandinavian -- it seems like there's a hell of a lot of Norwegian and Swedish OSS developers). I do wish that more open source projects came from India, though.
Most important servers have one thing running on them (say, your database server or your web server). If that thing gets compromised, you've already lost the game, since the only worthwhile thing on your system was just compromised.
I noticed this back in the 2.5 days. This won't explain what happens behind the scenes, but each nice point in 2.5/2.6 makes a *much* larger difference than in 2.4. Before, a nice difference of 5 was barely perceptable. You needed a difference of 10 to significantly "nice" down a pricess, and around 20 if you *really* wanted to keep a process from affecting another. This was really annoying, because there was essentially no point in adjusting niceness at a granularity of less than 5, and if you wanted to have regular user-level processes with interactive performance, entirely noninteractive daemons, and idle processes, there simply wasn't a large enough nice range to pull this off (you'd have to go to nice level 40 or so).
I'm sure that it's more sophisticated than just this, but I've found that a nice difference of 5 in 2.6 means that the nicer process can almost get starved -- exactly desired behavior. You can simply nice 20 your idle daemons and still have a big range to let users set up their own lower priority processes.
The story submission sounds pretty bad, but if you read the lkml email, it's actually pretty subtle. There's an obscure memory manager bug that requires some pretty serious finagling to be exploited.
Using prelink means startup times are generally great; konq starts up and displays home-dir in about 1/2 a second; mozilla and firebird take around 1 second to start up and display a simple home page.
Mozilla and Firebird both use XUL, not GTK+.
If you want to compare resize times in a GTK+ browser, open up the same page in Konq and dillo.
Actually, the skills level of those Indian programmers is generally acknowledged as being far higher than the average American programmer. They are better educated and work harder to obtain that education. They then work harder within their jobs
Hmm. I can't agree. See, here's my take, based on somewhat anecdotal evidence and a bit of reasoning.
Indians and Asians, at least in the United States, have an excellent academic and engineering reputation. This, however, has a lot less to do with any kind of racial difference than it does to do with deliberate social filtering. The US tries very hard to encourage only skilled, educated workers to immigrate, and allows talented workers to be brought in on H1Bs. These people are the cream of the crop -- the ones smart enough to acquire the right skills and interested enough in bettering themselves that they went through all the work required to end up in the US. There's a good chance that, even ignoring any kind of genetic influence, that they tend to teach their values to their children, so they have driven children as well.
This is where you get an "Indian workers are good" repulation from.
However, it doesn't have anything to do with outsourcing to India, where *Indian resident workers* are used.
Right now, the Indian IT industry is flush. Stupidly flush. There is incredible demand -- you lose a job, you're back in another in a week. Furthermore, it's young, with not a lot of quality enforcement or much reputation to go on. This is an *awful* environment to get good developers from. It's as bad as the dot-com years in the United States. Instead of choosing from the cream of the crop, you're choosing from the dregs.
Jobs that can be moved to cheaper workers effectively will be. China, India, and Indonesia will get a lot of US jobs. But it's not going to be a single smooth movement. It's definitely not as clean as the business rags represent it to be.
Universities are churning out students of ADA, Pascal and Java, most of whom applied to the university thinking of the good fortunes of being in IT around 1998.
In theory, mute beats the problem of using queries and traffic analysis to see who's sharing what.
Mmmf. I'm dubious.
This sounds like a really neat project to play with (I like to bat around P2P ideas as well.).
However, I'm going to assume (I can't tell from the routing document) that something here is incorrect.
The TTL mechanism is UtilityCounter. You attempt to obscure the real TTL by randomly moving it around. However, it's still pretty easy to simply send a number of messages until a TTL range 20 apart is reached. The host distance is then identified. Thus, a map of the MUTE network may be built, though it will take more packets than the GnutellaNet.
The main concerns I have with the MUTE protocol relate to flooding vulnerability. This is the same problem that GnutellaNet suffers from (and I have been working on in my own time). MUTE, however, is *extremely* vulnerable to flooding, far more so than GnutellaNet, for a number of reasons:
* MUTE shoves data packets through the MUTE network. GnutellaNet sends them directly.
* MUTE has phenomenally large TTLs, averaging 100.
One can probably destroy a massive MUTE network (unless I'm missing something in the routing protocol) with no more than a modem by flooding the network with data transfer packets of 32KiB (the largest the MUTE protocol allows) and bogus to virtual addresses.
I'd be interested in knowing whether there's an IRC channel for MUTE, since I'd be interested in poking at the design a bit. If any MUTE developers read this, would you point me in the right direction?
I've used GNOME more in the past, and use GTK+ based apps, but I'm not a huge fan of using GNOME. If I *just* had the GNOME gaim applet for gkrellm, I'd drop the GNOME panel as well.
good-looking (thanks to tigert, jimmac, and the lack of Keramik) than KDE.
Mmf. I also prefer the GNOME-style slightly organic icons, but I think more people I know prefer the more Windows-like KDE icons.
Practically all of the best-of-breed business apps (Mozilla Firebird, OpenOffice, Gnumeric, Evolution, Planner, Gnucash, etc.) are GNOME apps or GNOME-integrated.
Yes. Although it's depressing that Firebird and OO don't use GTK+.
and they prefer the power and configurability that KDE offers.
I think that GNOME 2 made a lot of functionality invisible. GNOME 1 had emacs-style keybindings and user-rebindable accelerators right in your face, and in GNOME 2 you need to flip a config file option to enable these. Doesn't mean that it went away, though.
KDE is much more of a hacker's desktop, and that's probably why Linus uses it.
I dunno -- I'd say it's an even split. Alan Cox uses GNOME. Probably neither uses either if they're testing kernels to avoid reboot time.
Which is why Python will be the supported scripting environment for Userlinux. Not perl, not Ruby, not TCL/TK.
Facinating.
I'd actually say that this is a good choice (and this is coming from someone who's much more familiar with perl than python).
Perl is, frankly, unmaintainable. It's just so damn useful for text processing that it's easy to shove in places, but I think that in the long run, folks are better off using a different language. Ruby isn't used that much, and TCL has been going a long slow fall from grace for a long time.
Python really is a good choice.
Of course, I'm sure that there will be a few crucial systems that rely on perl that will be a pain to replace.
The GPL promotes Free software development, because you are only allowed to create Free applications with it.
Conclusion does not follow.
I could say that the LGPL promotes Free software development, because it lowers the barrier to use for a lot of people and software, and means that more people get brought into the Linux fold (where divide and conquer allows OSS to tear up closed source).
Fine. Bruce is making a simple distribution with one GUI. He's making it to appeal to folks that don't want to muck about with TrollTech, and he said that he's more than happy to see folks making competing KDE-based distributions. That's just not how he wants to spend his time. He even pointed out that you're free to use his work to base your own on.
Most people agree that KDE is more mature and robust than GNOME anyway,
Not the people I talk to. At one point, this was certainly true.
KDE also has more stability from other points of view, for example it doesn't change the default window manager for each major release
KDE has had three major releases. In two of those, the window manager changed.
GNOME has had two major releases. In two of those , the window manager changed.
Not a *huge* difference is window manager stability. Furthermore, it doesn't look real likely that the WM will change in either of them any time soon.
For that matter, the default WM is a distro-specific choice.
Nobody expects all of this to go away if they switch to Linux. There will still be complexity.
He isn't talking about complexity from an end-user or administrator point of view, as far as I can tell, though that is certainly an issue. He was talking about how many resources the UserLinux distribution has.
kcontrol is integrated and pretty much all-encompassing, while GNOME is constantly shifting from CORBA over XML to a binary registry and back.
Uhhh....what? What you just said makes absolutely zero sense.
GNOME has become so bad that they actually added a regedit style "config editor" and apparently really expect users to use it to configure applications.
Actually, I'm not a gconf fan, and think it isn't a good idea -- probably the main drawback of GNOME. That being said, (a) gconf is very, very technically different from the registry, and (b) it's not really intended for end user use.
Anyway, I don't see why we need to standardize on a GUI, and if we do, we standardize on KDE, of course, as it fulfils more of the requirements businesses have, hands down.
You want a KDE-based distro, get off your butt and make a KDE-based distro. This is *not* rocket science -- lots of people already package things for you and you can reuse the non-GNOME portions of UserLinux. It only takes a couple of people to maintain a distro (especially a slow-moving distro that will appeal to businesses). Complaining about this is just so *stupid*. I don't say "those KDE people waste so much time working on their dumb KDE software". It's what they *want to do*. It's their volunteer time. You're free to put in your own volunteer time wherever you want, but don't bash people who are volunteering effort.
Slashdot Mirror
The ISS doesn't have a shower.
Rather than learning from this (as IBM and SGI seem to) they've gone the cynical, rationalist way.
I agree with all of your post except the bit about SGI learning from their mistakes. Last I heard, SGI had finished going through the meat grinder and was simply lying limply on the coutertop. Has this changed?
iTunes DRM is WEAK, man. Burn it to CDRW and rip the sucker again, it's as easy as jumping over a subway turnstile. Why are we wasting time with a pointless thing like this, why not crack WMP or something harder with a better payoff?
Because the method you are advocating is lossy, and there are a lot of Linux-using folks who might like to use iTunes.
Mmm...I dunno. Keep in mind that Apple spends marketing money and at least some money running, administering, and expanding the service. Even if the service is break even, it's better than losing money.
I suspect that you're right that the RIAA (well, member publishing companies) make out like bandits on the deal.
Argh. No. Not transparent proxies, unless you're providing an opt-in mechanism (say, a value-added service) to do so. It's *so* frusterating when ISPs start mucking with my network connection (firewall incoming and outgoing SMTP or transparently proxy it, block outgoing DNS, etc).
ISPs bitch about how they can't provide tiered service. Spam blocking, popup blocking, firewalling, are *all* great things to toss in value-added packages to provide tiered service. They, however, drive many people (who don't *want* said "service") to tears.
Another option would be a opaque proxy and to provide a mini application that sets it up.
Much as this drives me nuts as well, think about it. If there's an article on how a building in NYC collapsed in the Times which says that the cause was "structural failure in the concrete", the average reader just doesn't want more detail. Matter of fact, this is probably already more than they want to know. If it was actually the rebar at fault, it's close enough.
So, let's look at the myths that you're complaining about:
DeCSS is used to make copies of DVDs.
The most popular use of DeCSS and ilk has been to allow ripping and recompressing movies into a smaller, home-broadband-friendly format, which can then be redistributed. Short of explaining DiVX and CD burning, this is not an unreasonably far off explanation of what DeCSS is for Joe Public. It gives him a basic idea of what the thing does. It would be atrocious to try to feed this to a judge in a case or a techie, but the same goes for the rebar/concrete above.
Kazaa is an illegal music sharing site.
Also irritating. However, the majority of people out there know about email and the Web, and that's it. As a matter of fact, thanks to Microsoft using "Internet" to refer to the default web browser in XP, this is getting even more irritating -- "do you want to use email or the Internet?". If you don't want to explain the concept of other clients (which takes forever in *person*, much less in a short and snappy news article), this is reasonably on the ball. The "illegal" is pushing the limit -- I'm not entirely comfortable with that. However, it *is* probably true that over 99% of the traffic on Kazaa is illegal.
The music industry has said that "xyz" is legal/illegal....
I agree entirely with you here. It's poor journalism not to also provide an opposing view if you're presenting claims like this, though at *least* they qualify it with "music industry says" or "music industry claims".
An "Internet Virus" has been loosed...
Also irritating, but mostly to advocates of Linux or other mail clients. Few people know what "malware" or "worm" or "trojan" mean, or care. To them, the term "virus" means what "malware" does to you or me. Most people don't really understand or care about spreading mechanisms, other than the easiest method for them to avoid getting screwed over. Most people also use Windows and Windows alone. "Internet Virus" is not unreasonable to describe an Outlook worm in this context.
mp3s are illegally copied music files that Internet users share.
I don't think I've ever seen things put quite this harshly. I wouldn't be surprised to see "Many are illegal..." or "...frequently illegal...", though.
Basically, you *have* to take the context of your audience into account. First, the press is not unbiased. There are newspapers that are liberal, conservative, favor one politician or law over another, etc. It's a lot better than Soviet-era journalism, though. Second of all, if every time someone said "God" in a US newspaper, I suspect that few people feel the need for a Christian god qualification. People simply know what is meant. Sure, there's probably some theological inaccuracies that show up, just as there are technical inaccuracies here, but such is life.
This may seem awfully silly, and it doesn't mean that there's no good reason to backup DVDs. But, for the love of God, it *has* to be easier to just elevate the DVD player so that it's out of his reach. He can do a lot more damage than just to the DVDs -- when I was a child, I grated a bar of soap through the ventilation holes in my parents' stereo system.
The case wasn't taken to the Norwegian Supreme Court, so no legal precedent was made (presumably the reason why it didn't go to the SC). I may have misunderstood things (in the best /. tradition IANANL) but I don't think so.
IANANL (I Am Not A Norwegian Lawyer), but I'm strongly suspecting that the parent poster is correct.
If any prescedent was set, it is presumably that writing and distributing software to bypass copy protections are legal in Norway. This would put Norway smack dab where the United States was legally before the advent of the DMCA. It's legal to write and distribute the tools, but the actual act of copying DVDs is still copyright infringement.
From a legal standpoint, for all but a handful of software developers and security folks, this is not an issue.
It's a shame that Slashdot stories so frequently contain reasoning and conclusions (which may be false) or opinions (which may be inflammatory, misleading, etc). I wish Slashdot editors would simply dump stories that contain stuff like this -- this is what the comment section is for.
Are you bubbling? Valid nice ranges are -20 to +19.
Sure -- read the whole sentence in my post:
there simply wasn't a large enough nice range to pull this off (you'd have to go to nice level 40 or so).
I'm complaining that you didn't *have* a large enough range, given the tiny difference between each nice level.
Misconception 1 - American programmers are better.
I don't think that this is what folks are really thinking.
Right now, the Indian IT and software development industry has been growing at an insane rate. Demand has been doubling and doubling and doubling. This isn't a criticism of Indians per se. It's just a fact that when you have a market like that, you're going to wind up employing bad along with the good, because it's easy for anyone to get a job. The same thing happened in the US during the dot-com boom. There was a huge amount of overdemand for web designers, so there were a huge number of incompetent web designers running around. In ten years, everything will probably be fine -- graduates will increase to meet demand, poor employees will develop a poor reputation, and so on. However, at the moment, I'd say that it's a reasonable assumption to say that you're likely to get worse work out of a typical Indian software house than a typical American software house at the moment. That doesn't mean that you can't get a good Indian developer, just that it's easier to get a worse one.
Second of all, the Indian software development industry is younger than the American one, and shooting up to meet demand. It's already hard to avoid godawful consultants in the US. When you go to an environment where reputations may not have been established for as long, things just get more difficult.
Misconception 2 - Indians are not innovative.
Sure.
The problem is the reasons most US companies are moving overseas. This isn't Wolfram running out and looking for a new brilliant lead computer scientist (well, they might well hire an Indian, but that's irrelevant -- this sort of behavior isn't forming the bulk of the recent surge). A lot of jobs going overseas are for ordinary drudge work -- web dev, basic system administration. These companies are looking for folks that don't cost $80K/year to write ordinary code. They're tired of paying obscene prices for work that's pretty straightforward. My guess is that this is where the "uninteresting jobs going overseas" idea comes from.
Finally, one person's anecdotal evidence doesn't count for much, but I've had only awful experiences with both India-based IT and India-based software development. There are clearly currently at least some really awful people out there. My guess is that in a few years, as demand and supply equalize, the situation will improve. However, please understand that not all American dislike of outsourcing simply stems from a sense of job threat.
The final issue, which is simply a problem for any international company, is that it can be terribly difficult to understand a thickly accented voice, especially once it's been crammed through the lossy compression that cell phones apply to voices. I'm sure the reverse applies (and hence a US-based programmer might be less desireable to an operation which is primarily India-based). However, it really can be frusterating to have to ask someone to repeat what they've said three or four times. They may show admirable patience in being willing to repeat what they're saying, but it's still a frusterating communications problem. This is especially so during meetings, where it's embarrassing for everyone involved if someone has to be asked to repeat themselves multiple times. Note that I don't see this as much of a problem with email. Even with poor English (and I've seen awfully good written English from many Indians with strong accents), I can usually make out what's being said without trouble.
On the other hand, I'm also sensitive to the fact that the US is edging uncomfortably closer to the old UK anti-Indianism due to labor issues, and I try not to be unfair.
As an aside, I'm also a bit saddened by the fact that there seem to be awfully few India-based open source projects. Perhaps open source springs simply from affluence, and hasn't been as available to many Indians as it has been western European and US-based developers (particularly Scandinavian -- it seems like there's a hell of a lot of Norwegian and Swedish OSS developers). I do wish that more open source projects came from India, though.
Mmmmfff....it's not quite that bad.
Most important servers have one thing running on them (say, your database server or your web server). If that thing gets compromised, you've already lost the game, since the only worthwhile thing on your system was just compromised.
I noticed this back in the 2.5 days. This won't explain what happens behind the scenes, but each nice point in 2.5/2.6 makes a *much* larger difference than in 2.4. Before, a nice difference of 5 was barely perceptable. You needed a difference of 10 to significantly "nice" down a pricess, and around 20 if you *really* wanted to keep a process from affecting another. This was really annoying, because there was essentially no point in adjusting niceness at a granularity of less than 5, and if you wanted to have regular user-level processes with interactive performance, entirely noninteractive daemons, and idle processes, there simply wasn't a large enough nice range to pull this off (you'd have to go to nice level 40 or so).
I'm sure that it's more sophisticated than just this, but I've found that a nice difference of 5 in 2.6 means that the nicer process can almost get starved -- exactly desired behavior. You can simply nice 20 your idle daemons and still have a big range to let users set up their own lower priority processes.
The story submission sounds pretty bad, but if you read the lkml email, it's actually pretty subtle. There's an obscure memory manager bug that requires some pretty serious finagling to be exploited.
USB natively supports up to 127 or so devices in V1, not sure how many in V2.
That's a lot more than four controllers.
Why bother with some obsolete and oddball hardware when you can just get a bunch of usb controllers?
Using prelink means startup times are generally great; konq starts up and displays home-dir in about 1/2 a second; mozilla and firebird take around 1 second to start up and display a simple home page.
Mozilla and Firebird both use XUL, not GTK+.
If you want to compare resize times in a GTK+ browser, open up the same page in Konq and dillo.
Actually, the skills level of those Indian programmers is generally acknowledged as being far higher than the average American programmer. They are better educated and work harder to obtain that education. They then work harder within their jobs
Hmm. I can't agree. See, here's my take, based on somewhat anecdotal evidence and a bit of reasoning.
Indians and Asians, at least in the United States, have an excellent academic and engineering reputation. This, however, has a lot less to do with any kind of racial difference than it does to do with deliberate social filtering. The US tries very hard to encourage only skilled, educated workers to immigrate, and allows talented workers to be brought in on H1Bs. These people are the cream of the crop -- the ones smart enough to acquire the right skills and interested enough in bettering themselves that they went through all the work required to end up in the US. There's a good chance that, even ignoring any kind of genetic influence, that they tend to teach their values to their children, so they have driven children as well.
This is where you get an "Indian workers are good" repulation from.
However, it doesn't have anything to do with outsourcing to India, where *Indian resident workers* are used.
Right now, the Indian IT industry is flush. Stupidly flush. There is incredible demand -- you lose a job, you're back in another in a week. Furthermore, it's young, with not a lot of quality enforcement or much reputation to go on. This is an *awful* environment to get good developers from. It's as bad as the dot-com years in the United States. Instead of choosing from the cream of the crop, you're choosing from the dregs.
Jobs that can be moved to cheaper workers effectively will be. China, India, and Indonesia will get a lot of US jobs. But it's not going to be a single smooth movement. It's definitely not as clean as the business rags represent it to be.
Universities are churning out students of ADA, Pascal and Java, most of whom applied to the university thinking of the good fortunes of being in IT around 1998.
What's the beef with Ada?
In theory, mute beats the problem of using queries and traffic analysis to see who's sharing what.
Mmmf. I'm dubious.
This sounds like a really neat project to play with (I like to bat around P2P ideas as well.).
However, I'm going to assume (I can't tell from the routing document) that something here is incorrect.
The TTL mechanism is UtilityCounter. You attempt to obscure the real TTL by randomly moving it around. However, it's still pretty easy to simply send a number of messages until a TTL range 20 apart is reached. The host distance is then identified. Thus, a map of the MUTE network may be built, though it will take more packets than the GnutellaNet.
The main concerns I have with the MUTE protocol relate to flooding vulnerability. This is the same problem that GnutellaNet suffers from (and I have been working on in my own time). MUTE, however, is *extremely* vulnerable to flooding, far more so than GnutellaNet, for a number of reasons:
* MUTE shoves data packets through the MUTE network. GnutellaNet sends them directly.
* MUTE has phenomenally large TTLs, averaging 100.
One can probably destroy a massive MUTE network (unless I'm missing something in the routing protocol) with no more than a modem by flooding the network with data transfer packets of 32KiB (the largest the MUTE protocol allows) and bogus to virtual addresses.
I'd be interested in knowing whether there's an IRC channel for MUTE, since I'd be interested in poking at the design a bit. If any MUTE developers read this, would you point me in the right direction?
And, as the KDE advocates note, KDE is a far more common default - among all distros, not only those aimed at new users.
Okay, this is going based on memory of the major distros:
KDE:
SuSE
Mandrake
GNOME:
Debian
Red Hat
Bizarre. I feel almost the opposite.
I may be biased as a GNOME user,
I've used GNOME more in the past, and use GTK+ based apps, but I'm not a huge fan of using GNOME. If I *just* had the GNOME gaim applet for gkrellm, I'd drop the GNOME panel as well.
good-looking (thanks to tigert, jimmac, and the lack of Keramik) than KDE.
Mmf. I also prefer the GNOME-style slightly organic icons, but I think more people I know prefer the more Windows-like KDE icons.
Practically all of the best-of-breed business apps (Mozilla Firebird, OpenOffice, Gnumeric, Evolution, Planner, Gnucash, etc.) are GNOME apps or GNOME-integrated.
Yes. Although it's depressing that Firebird and OO don't use GTK+.
and they prefer the power and configurability that KDE offers.
I think that GNOME 2 made a lot of functionality invisible. GNOME 1 had emacs-style keybindings and user-rebindable accelerators right in your face, and in GNOME 2 you need to flip a config file option to enable these. Doesn't mean that it went away, though.
KDE is much more of a hacker's desktop, and that's probably why Linus uses it.
I dunno -- I'd say it's an even split. Alan Cox uses GNOME. Probably neither uses either if they're testing kernels to avoid reboot time.
That doesn't mean that all the config options have to disappear (ahem, Metacity can bite my ass)
Don't use Metacity if you're a power user. There are a hell of a lot of GNOME-supporting WMs.
I've been a fan of sawfish since well before GNOME adopted it, but it's not like there are a dearth of other choices.
Which is why Python will be the supported scripting environment for Userlinux. Not perl, not Ruby, not TCL/TK.
Facinating.
I'd actually say that this is a good choice (and this is coming from someone who's much more familiar with perl than python).
Perl is, frankly, unmaintainable. It's just so damn useful for text processing that it's easy to shove in places, but I think that in the long run, folks are better off using a different language. Ruby isn't used that much, and TCL has been going a long slow fall from grace for a long time.
Python really is a good choice.
Of course, I'm sure that there will be a few crucial systems that rely on perl that will be a pain to replace.
The GPL promotes Free software development, because you are only allowed to create Free applications with it.
Conclusion does not follow.
I could say that the LGPL promotes Free software development, because it lowers the barrier to use for a lot of people and software, and means that more people get brought into the Linux fold (where divide and conquer allows OSS to tear up closed source).
That said, KDE is faster. Much, much faster;
That's...odd. Really odd.
I've generally found Qt to be noticeably (and last time I was using it, which was on an older computer, irritatingly) slower than GTK+.
What things were you doing that were particularly slow?
I'm a user and I want KDE.
Fine. Bruce is making a simple distribution with one GUI. He's making it to appeal to folks that don't want to muck about with TrollTech, and he said that he's more than happy to see folks making competing KDE-based distributions. That's just not how he wants to spend his time. He even pointed out that you're free to use his work to base your own on.
Most people agree that KDE is more mature and robust than GNOME anyway,
Not the people I talk to. At one point, this was certainly true.
KDE also has more stability from other points of view, for example it doesn't change the default window manager for each major release
KDE has had three major releases. In two of those, the window manager changed.
GNOME has had two major releases. In two of those , the window manager changed.
Not a *huge* difference is window manager stability. Furthermore, it doesn't look real likely that the WM will change in either of them any time soon.
For that matter, the default WM is a distro-specific choice.
Nobody expects all of this to go away if they switch to Linux. There will still be complexity.
He isn't talking about complexity from an end-user or administrator point of view, as far as I can tell, though that is certainly an issue. He was talking about how many resources the UserLinux distribution has.
kcontrol is integrated and pretty much all-encompassing, while GNOME is constantly shifting from CORBA over XML to a binary registry and back.
Uhhh....what? What you just said makes absolutely zero sense.
GNOME has become so bad that they actually added a regedit style "config editor" and apparently really expect users to use it to configure applications.
Actually, I'm not a gconf fan, and think it isn't a good idea -- probably the main drawback of GNOME. That being said, (a) gconf is very, very technically different from the registry, and (b) it's not really intended for end user use.
Anyway, I don't see why we need to standardize on a GUI, and if we do, we standardize on KDE, of course, as it fulfils more of the requirements businesses have, hands down.
You want a KDE-based distro, get off your butt and make a KDE-based distro. This is *not* rocket science -- lots of people already package things for you and you can reuse the non-GNOME portions of UserLinux. It only takes a couple of people to maintain a distro (especially a slow-moving distro that will appeal to businesses). Complaining about this is just so *stupid*. I don't say "those KDE people waste so much time working on their dumb KDE software". It's what they *want to do*. It's their volunteer time. You're free to put in your own volunteer time wherever you want, but don't bash people who are volunteering effort.