I see where you're going, and in fact I just did that to eliminate this "feature." The problem I have is that there isn't an option to turn off the nagging. The effect that this has is twofold:
1. It encourages people to reboot. (i.e., as intended)
2. It causes people to delay installing the patches because, well, they have to reboot in the first place, and they get sick of the nagging.
So the result is that most people do what I've done, which is "download updates for me but let me choose when to install them." The problem is that a lot of the time they'll wind up not installed. (*I*'ll install them, but God only knows about Joe Bob.)
This kind of thing is rampant in the Windows world. For example, Norton Antivirus (I think it is) has an option to do automatic virus scans on a schedule. This is a GOOD thing. It should be done. Unfortunately, since it doesn't run with the equivelent of a "nice 20" and it insists on hogging the console as well (until you background the thing), a lot of people, including me, just turn the autoscan feature off.
The one exception to this is probably firewalls. When firewalls do this kind of thing and don't play nice, they do it ONCE for an application most of the time, so it doesn't become annoying. Sure, it might crash the whole freaking 3D app when it unceremoniously grabs the desktop to pop up a little bubble dialog, but it should happen once if at all, and that's it. So it isn't the same thing.
Now, while we're on the topic, I might as well get my post downmodded by saying something that Windows tends to do well that I like. Well, Windows specifically doesn't do it, but the various firewalls out there do. You authorize *applications* to either access the net or not, which is nice. Granted, it isn't all that you need for a decent firewall, but it would be nice if Linux made that kind of enforcement fairly transparent. (Of course, make the admin have to turn it on. Don't do it by default or all sorts of stuff will break.)
As for the rest of you, if you think Windows is so bad, why pirate it?
Because hardware manufacturers refuse to release Linux applications and drivers for their hardware so you're locked into the Windows platform? (Try to get a Palm Lifedrive to work under Gpilot. I still haven't gotten that working. Yeah, I know about hacking the Gpilot.xml file; still doesn't work.)
Because the majority of the "non-serious" software out there runs under Windows and nothing else?
Because Microsoft and its cohorts deliberately introduce incompatibilities and refuse to release file format specifications so that you are locked into the Windows platform?
I own four copies of Windows XP. All four are from HP. Three are on desktops, and one is on a laptop. The newest machine is about a year old, and damned if I know where the Windows serial number or whatever the hell they use is. Despite the fact that all four are legal, if I were to suddenly be locked out of functionality because of Micro$hit's paranoia, I'd be justifiably pissed.
Windows, if you define it as the OS including all the bullshit the OEMs (e.g., HP) ship with it, is already loaded down with "questionable" things. Crap like the Weatherbug, half a dozen different applications that want to connect to the Internet for no identifiable reason (so I assume to do something nefarious), etc.
The problem with Windows is, and always has been, that it's intended to be easy for new users. This is good in that it is, well, easy for new users. The problem arises when it becomes annoying for people who actually know what's going on, and when the layers of "user friendliness" becomes anything but by hiding all sorts of stuff from the user.
On my Linux system, I know what it's doing. If nothing else, I can look at the process listing, look at the manual pages for the various applications, and if I'm really suspicious I can strace/ltrace the thing. It's also easy to kill it. On my Windows system, on the other hand, it's very difficult to find out what the hell half the crap that's loaded is. Is it really needed? Is it fluff? Is it for some random feature? Is it something that's a potential security hole? Hell if I know, due to lack of documentation, lack of development tools (unless I buy them), and the lack of source, among other things.
Windows is good for the casual, "I don't care how it works" kind of user, especially if that user doesn't care one wit about data and system security. It's remarkably bad for someone who needs/wants to know how things work, and who does care about data and system security.
Windows has its niche. Granted, it's a pretty big niche. But it isn't mine, and aside from playing games and hotsyncing my Palm (damned Windows-only desktop-side apps), it's a pain in my ass.
Use what works for you. In the case of Windows, don't be surprised when it's ridden with spyware, viruses, security holes, ill-defined applications that are all over the place and you don't know what they do, et cetera. In the case of Linux, don't be surprised when hardware support blows chunks (not that I think this is right, mind you).
All that said, I seriously doubt Microsoft would implement something like that. Historically, they have, but supposedly unintentionally. (I've had NT installations wipe their drives, Windows scribble all over Linux partitions, etc.) If they did it intentionally, the blackhats would be all over it and cost Microsoft more than the entire thing was worth by releasing the next fifteen versions of Slammer targetting different vulnerabilities.
2. Use Microsoft's "recommended" setting to allow it to install service packs.
3. Go away for the weekend.
4. Come back to find out that Windows at some point started a stupid 5 minute countdown to reboot the machine without user intervention.
Even better, try doing anything like playing a game while that stupid dialog keeps coming up every twenty minutes or so, even though you've already told Windoze that you'll "reboot later."
If I tell it to reboot later and it keeps FUBARing applications by popping up a stupid dialog and seizing control of the machine like that, or just spontaneously reboots because I don't see the dialog, then it's effectively the same as a forced shutdown due to a crash.
What the authors are trying to preach in this series are techniques that are used by the "bad guys" so that the "good guys" can learn from them.
I don't want a military bomb expert who couldn't figure out how to build a bomb (albeit a lame one) with some nails, some glass bottles, Clorox, and a bottle of ammonia. Similarly, I feel a hell of a lot better as a network administrator (unpaid -- grumble) knowing how to compromise a system when it comes to securing the damned thing.
STN: How To Own A Continent was good, I thought. Rarely do I pick up a book over coffee in a bookstore, and not only wind up buying the thing for $25 or whatever it was, but read the thing non-stop (except for work) from start to finish in two days.
The STN series is unique in that it focuses on technical details. Some of the fiction itself might be lacking in form, but the reality is that this is not what the authors are really trying to do. They're trying to educate about various techniques in an entertaining way, and in that department I think they do a pretty good job.
One of the criticism about STN: How To Own A Continent that I had, though, was that there didn't seem to be enough technical details. I'd much rather read a book like this and have it go over my head than be able to understand everything without much thought. From this review, it looks like they might have addressed this in STN: How To Steal an Identity.
Most likely, I'll be ducking into a bookstore to buy this thing.
About a year or so ago, I got a wild hair and decided to design a kind of "futuristic city information system" for lack of a better term. I considered the centralized idea, which had some advantages, but the huge drawback was having a single point of failure.
What I finally decided on was having a centralized computer and data repository, with the local nodes being able to handle their functions autonomously. So long as the network is up, the entire mess is effectively Beowulfed to all the other machines and the central system. If the network connection to a local node drops, then control is transferred transparently to that local node.
The catch, of course, is that if you do that you can't store all the data on the centralized server, and you still have to have copies of the programs on the local node. So really, the local nodes still had to handle their local stuff themselves, using the rest of the network for archives, noncritical data storage, and as a processor power boost.
This "web-based world" is the same idea. Being commercially driven, though, it will be entirely different. There will be no local redundancy worth mentioning, because why bother to code it? If you code it, then what keeps someone from buying your software for $100 rather than paying a $15/month subscription fee to your company?
This is one of those ideas that might be good, if you could figure out how to handle all the security issues, but is fundamentally doomed once the almighty dollar gets involved.
I think the "bloat" he's referring to is the fact that running a KDE application results in a wait while a bunch of seemingly (to a non-programmer) inane stuff is loaded, complete with some stupid message about ksyscoa or whatever it's called.
If you run in GNOME, KDE applications look bloated. If you run in KDE, GNOME applications looked bloated. It's a side-effect of having two different frameworks that different applications use.
It's still a honking security hole if it just goes off and runs an untrusted application without user intervention.
Picture this. Let's say that I have a friend, we'll call her Lauren, who happens to use KDE. Since I know what she's using, and I have some reason to compromise her system, I give her a data CD of documents or something. She puts it in. KDE transparently runs the "autorun" on it, which installs a network-accessible backdoor under her account, or (since I have to worry about routers and firewalls and such) checks to see if it can make an outgoing connection. If it does, it "phones home." If not, it just snarfs her mail settings from Kmail, installs a key logger, and mails compressed logs back to me at routine intervals.
Autorun is a bad idea, no matter what OS it's under. It was a bad idea under Windows, and it's a bad idea under something like Linux too.
I've had a hell of a time getting my family familiar with the concept of unmounting media. "My CD is stuck!" Or someone mounts a floppy, ejects it without unmounting, and FUBARing results. It would be nice if KDE does something in this regard for the n00bs out there. (It might already. I've been using Gnome.)
For what it's worth (not much), Good Morning America on ABC is running with this idiotic story. They show video footage of the material obviously not hitting the shuttle, and are running this scare piece on "debris hitting the shuttle."
Of course it is. A human "race" is simply a group of people who, for whatever reason (ideological, religious, geographical, etc.) mated within their own group. This allows mutations which occur within that group to stay there, rather than get folded over the entire genome. Continue this for long enough, and some of those mutations will result in situations where members of different races can't reproduce with each other.
As an ass-backward example, consider sickle cell anemia. The sickle cell trait (Ss), whereby an individual has one copy of the gene but one copy of the non-sickle gene, conveys a bit of immunity to malaria. So there was environmental selection for people who are Ss. People who were ss (two normal genes) died at higher rates from malaria. The catch is that people with two copies of the gene (SS) die from sickle cell anemia, often before they reproduce.
Let's say that for a particular gene locus, you can have one of two versions of a gene. People with XX live fine. People with YY live fine. People with XY drop dead. If 90% of, say, Asians have YY, and 90% of, say, blacks have XX, then you have a very high probability of getting XY in a resultant offspring. In effect, you're speciated. These two races of people can't reproduce with each other at a reasonable rate.
What was necessary for survival in, say, Africa, was entirely different than what was necessary in England. As a result, there was divergent evolution with people in each area subtly specializing in different things. Now, this didn't stop, say, the Saxons and the Africans from being able to reproduce, but it easily could have. It just happens that humans apparently have more redundancy than a butterfly does, which is why simpler organisms are studied for this kind of research.
That doesn't in any way excuse racism, of course, which was the point (I think) of the parent. However, it does raise questions about things like intelligence, athletic performance, and everything else. (Nobody balks at someone putting forth the idea that blacks tend to be better athletes than whites, for example. Change 'athlete' to 'quantum theorists' and everyone blows up. It's the same thing, just a different trait. And yes, this is just an example.)
If I were a doctor, a full medial records + billing application would be worth many thousands (or equivalent of support services for free software). If I am running a bakery, then inventory software is worth far less.
Interestingly enough, such applications go for thousands. The problem is that people just starting out, medical students, etc., can't afford that. There's a massive disparity between the free and low cost software out there, and the $10,000 packages. That isn't surprising.
What is surprising is just how absolutely terrible the OSS in this field is. For reference, in order for me to consider it, the software package has to support a Palm and sync to a desktop application.
And no, I'm not just bitching about it and having no intention to do something about it. As soon as I get my Lifedrive to actually work with gpilot (any hints, guys?) I'm going to start on this.
Not only that, but in a real, honest-to-God emergency, you can send Morse in a wide-band pattern using a car battery.(And it better be a real damned emergency or you're going to get eviscerated by the FCC and the Ham community alike.)
A more profound question is the following. What is the point of amateur radio when the Internet has connected most of the globe?
There are still plenty of places that the Internet doesn't reach, but that isn't the real reason for it. Ham radio is mostly a hobbiest thing (e.g., people using the local disaster frequencies to talk about their dogs) until something happens like the power grid going to hell in an area, a nuke going off, or God only knows what else. At that point, the Internet is going to be full of holes, if it's running in the area at all, the phone system may or may not work or will at the least be jammed all to hell, cell phones will either not work or be jammed, and so on.
At that point, it becomes very, very important and useful for public safety and disaster management to have a communications network that is largely self contained. (Read: Handheld 5W Ham sets, people hooking their base stations up to car batteries or generators, etc.)
This is precisely the reason that as soon as I get rid of some exceedingly annoying medical certifications I have to deal with in the next few weeks, I intend to get a ham license.
With that in mind, O' Slashdot, perhaps those who know this stuff cold can enlighten the forum: Is there much point, all in all, in going for the higher level licenses, or should I just stick with the low-band entry level ones?
That was my point, really. I can guarantee that on power failure a modern desktop PC couldn't recover if you gave it a roadmap. What they could do -- and what I think you're getting at -- is stick a small battery in the thing and when a power failure interrupt is raised, it calls a temporary system halt, shifts the RAM over to some other kind of power supply (or writes it to disk if it can), and saves all the state data. When power is restored, the system comes back up, the OS restores the application and hardware states, and then resumes normal execution.
I'm not sure PoE wold help that much, since the power on the Ethernet line has to come from somewhere. I can see an application on an enterprise level with a huge, centralized UPS that continues powering the Ethernet line, allowing systems to save. But then we're back where we started before, with either having to have a battery in the machine itself anyway, or the router becoming a toaster from having to supply a ton of power on every switch port.
That probably isn't as easy as you might think at first, because while you'll retain the contents of RAM, you'll lose all hardware and processor state information.
It depends on what you want to do. I have a 3GHz AMD with 1GB RAM and about 400GB of HDD space. Among other machines, I also have a 500MHz Athlon with 256MB, a bunch of disk space, and a P133 with 32MB and some reasonable disk space for what I use it for.
The 3GHz will run Windows great, performance wise. It will also run Linux great. No surprise there.
The 500MHz machine runs Windows like crap. It runs Linux really well, all things considered. Compile times are reasonable on it. It handles everything I throw at it, including full routing functions, a VPN, SMTP server, web server, some PHP stuff, etc. etc. X11 might be scary.
The P133 won't run Windows at all. It runs Linux just fine. I use it as a backend fileserver primarily, along with some database functions. It does this just fine. Compile times SUCK GOATS. I mean, they're BAD. I kicked off a compile of gcc 3.4.4 as an upgrade and was surprised (but not shocked) when it was still going 12 hours later. I haven't even tried since I retired this machine from desktop use about 8 years ago, but X11 would probably look like I was running it through a 2400baud dialup.
Some of it is feature bloat. This is a good thing, so long as those features can be turned off when necessary. Some of it is just "bad" programming (bad in the sense that it assumes the user to have more processing power and memory capacity than he does). In the case of my P133, the next project is just to distcc everything over to the 500MHz, 3GHz Athlon (when it isn't in Windows to play games or something), and a P4 3GHz laptop, and cut the P133 out of the loop for its own compilations, because letting it have any part of the compiling will just slow the entire process down.
What is a "practical system?" To most users, that's being able to run a GUI desktop, play some games, do some GUI word processing, etc. My P133 and my 500MHz machine are both pretty unsuited for that, though the latter might be passable. So I don't use them for that.
The significant difference is that if for some bizarre reason I ever wanted to, I could boot up the 20MHz 386 machine I have in a closet with 32MB of RAM, and use it for something using Linux. (I can't think of what, though, short of really low-power file serving or maybe running Postgres to do nothing other than log syslog messages from the network.)
Yes (once), and it's the exact reason I absolutely refuse to use Windows for it, and why I refuse to use C for it. Hell, in that kind of situation, I'd rather use ObjectiveC (not much better for obvious reasons), Python, or Perl, and even then I'm not too happy with it.
It comes down to testing the hell out of the software. Even then, there is always the possibility that you've missed something, and that something will blow up in your face. At least you can do your part to help ensure that it isn't something as mundane as a buffer overrun or screwed up memory free().
You're right, of course. One way to handle this is to set up the virtual economy such that only a small number of virtual people have to deal directly with the basic necessities of survival. The ones that have free time are free to do whatever else they want. To make this work, have some kind of skill system, modelling how someone gets good at woodworking when they do that all the time, for example.
This leads to seeing whether the automata will diversify their labor or take equal shares. Will there be some automata who somehow become freeloaders? Nothing but breeders? Builders? Farmers?
This is an interesting experiment if it can be done robustly enough to actually show something.
Not really. WiFi is always going to be inherently less secure than the equivelent implementation on a physical, wired line because of the nature of radio communications. Anyone within range can intercept it.
As for WiFi's security, it's flawed, and slows down attackers rather than stopping them. WEP can be broken relatively easily, and hiding your SSID doesn't save you either contrary to what some people might think.
The real way to handle WiFi security is to open a VPN with strong encryption to your router, and route everything through that VPN. If you're concerned about unauthorized people syncing to the network, MAC address filter *and* require some kind of cryptographic key exchange with the router prior to opening the communication. The same can apply for wired Ethernet; run a VPN between physically unsecured bits of cable and you bypass that problem.
Slashdot Mirror
1. It encourages people to reboot. (i.e., as intended)
2. It causes people to delay installing the patches because, well, they have to reboot in the first place, and they get sick of the nagging.
So the result is that most people do what I've done, which is "download updates for me but let me choose when to install them." The problem is that a lot of the time they'll wind up not installed. (*I*'ll install them, but God only knows about Joe Bob.)
This kind of thing is rampant in the Windows world. For example, Norton Antivirus (I think it is) has an option to do automatic virus scans on a schedule. This is a GOOD thing. It should be done. Unfortunately, since it doesn't run with the equivelent of a "nice 20" and it insists on hogging the console as well (until you background the thing), a lot of people, including me, just turn the autoscan feature off.
The one exception to this is probably firewalls. When firewalls do this kind of thing and don't play nice, they do it ONCE for an application most of the time, so it doesn't become annoying. Sure, it might crash the whole freaking 3D app when it unceremoniously grabs the desktop to pop up a little bubble dialog, but it should happen once if at all, and that's it. So it isn't the same thing.
Now, while we're on the topic, I might as well get my post downmodded by saying something that Windows tends to do well that I like. Well, Windows specifically doesn't do it, but the various firewalls out there do. You authorize *applications* to either access the net or not, which is nice. Granted, it isn't all that you need for a decent firewall, but it would be nice if Linux made that kind of enforcement fairly transparent. (Of course, make the admin have to turn it on. Don't do it by default or all sorts of stuff will break.)
Because hardware manufacturers refuse to release Linux applications and drivers for their hardware so you're locked into the Windows platform? (Try to get a Palm Lifedrive to work under Gpilot. I still haven't gotten that working. Yeah, I know about hacking the Gpilot .xml file; still doesn't work.)
Because the majority of the "non-serious" software out there runs under Windows and nothing else?
Because Microsoft and its cohorts deliberately introduce incompatibilities and refuse to release file format specifications so that you are locked into the Windows platform?
I own four copies of Windows XP. All four are from HP. Three are on desktops, and one is on a laptop. The newest machine is about a year old, and damned if I know where the Windows serial number or whatever the hell they use is. Despite the fact that all four are legal, if I were to suddenly be locked out of functionality because of Micro$hit's paranoia, I'd be justifiably pissed.
The problem with Windows is, and always has been, that it's intended to be easy for new users. This is good in that it is, well, easy for new users. The problem arises when it becomes annoying for people who actually know what's going on, and when the layers of "user friendliness" becomes anything but by hiding all sorts of stuff from the user.
On my Linux system, I know what it's doing. If nothing else, I can look at the process listing, look at the manual pages for the various applications, and if I'm really suspicious I can strace/ltrace the thing. It's also easy to kill it. On my Windows system, on the other hand, it's very difficult to find out what the hell half the crap that's loaded is. Is it really needed? Is it fluff? Is it for some random feature? Is it something that's a potential security hole? Hell if I know, due to lack of documentation, lack of development tools (unless I buy them), and the lack of source, among other things.
Windows is good for the casual, "I don't care how it works" kind of user, especially if that user doesn't care one wit about data and system security. It's remarkably bad for someone who needs/wants to know how things work, and who does care about data and system security.
Windows has its niche. Granted, it's a pretty big niche. But it isn't mine, and aside from playing games and hotsyncing my Palm (damned Windows-only desktop-side apps), it's a pain in my ass.
Use what works for you. In the case of Windows, don't be surprised when it's ridden with spyware, viruses, security holes, ill-defined applications that are all over the place and you don't know what they do, et cetera. In the case of Linux, don't be surprised when hardware support blows chunks (not that I think this is right, mind you).
All that said, I seriously doubt Microsoft would implement something like that. Historically, they have, but supposedly unintentionally. (I've had NT installations wipe their drives, Windows scribble all over Linux partitions, etc.) If they did it intentionally, the blackhats would be all over it and cost Microsoft more than the entire thing was worth by releasing the next fifteen versions of Slammer targetting different vulnerabilities.
1. Install Windows XP SP 2.
2. Use Microsoft's "recommended" setting to allow it to install service packs.
3. Go away for the weekend.
4. Come back to find out that Windows at some point started a stupid 5 minute countdown to reboot the machine without user intervention.
Even better, try doing anything like playing a game while that stupid dialog keeps coming up every twenty minutes or so, even though you've already told Windoze that you'll "reboot later."
If I tell it to reboot later and it keeps FUBARing applications by popping up a stupid dialog and seizing control of the machine like that, or just spontaneously reboots because I don't see the dialog, then it's effectively the same as a forced shutdown due to a crash.
I don't want a military bomb expert who couldn't figure out how to build a bomb (albeit a lame one) with some nails, some glass bottles, Clorox, and a bottle of ammonia. Similarly, I feel a hell of a lot better as a network administrator (unpaid -- grumble) knowing how to compromise a system when it comes to securing the damned thing.
The STN series is unique in that it focuses on technical details. Some of the fiction itself might be lacking in form, but the reality is that this is not what the authors are really trying to do. They're trying to educate about various techniques in an entertaining way, and in that department I think they do a pretty good job.
One of the criticism about STN: How To Own A Continent that I had, though, was that there didn't seem to be enough technical details. I'd much rather read a book like this and have it go over my head than be able to understand everything without much thought. From this review, it looks like they might have addressed this in STN: How To Steal an Identity.
Most likely, I'll be ducking into a bookstore to buy this thing.
About a year or so ago, I got a wild hair and decided to design a kind of "futuristic city information system" for lack of a better term. I considered the centralized idea, which had some advantages, but the huge drawback was having a single point of failure.
What I finally decided on was having a centralized computer and data repository, with the local nodes being able to handle their functions autonomously. So long as the network is up, the entire mess is effectively Beowulfed to all the other machines and the central system. If the network connection to a local node drops, then control is transferred transparently to that local node.
The catch, of course, is that if you do that you can't store all the data on the centralized server, and you still have to have copies of the programs on the local node. So really, the local nodes still had to handle their local stuff themselves, using the rest of the network for archives, noncritical data storage, and as a processor power boost.
This "web-based world" is the same idea. Being commercially driven, though, it will be entirely different. There will be no local redundancy worth mentioning, because why bother to code it? If you code it, then what keeps someone from buying your software for $100 rather than paying a $15/month subscription fee to your company?
This is one of those ideas that might be good, if you could figure out how to handle all the security issues, but is fundamentally doomed once the almighty dollar gets involved.
You're only going to see those messages if you run the program in a terminal. Otherwise, the output is piped to stderr, which winds up being /dev/null.
If you run in GNOME, KDE applications look bloated. If you run in KDE, GNOME applications looked bloated. It's a side-effect of having two different frameworks that different applications use.
Picture this. Let's say that I have a friend, we'll call her Lauren, who happens to use KDE. Since I know what she's using, and I have some reason to compromise her system, I give her a data CD of documents or something. She puts it in. KDE transparently runs the "autorun" on it, which installs a network-accessible backdoor under her account, or (since I have to worry about routers and firewalls and such) checks to see if it can make an outgoing connection. If it does, it "phones home." If not, it just snarfs her mail settings from Kmail, installs a key logger, and mails compressed logs back to me at routine intervals.
Autorun is a bad idea, no matter what OS it's under. It was a bad idea under Windows, and it's a bad idea under something like Linux too.
I've had a hell of a time getting my family familiar with the concept of unmounting media. "My CD is stuck!" Or someone mounts a floppy, ejects it without unmounting, and FUBARing results. It would be nice if KDE does something in this regard for the n00bs out there. (It might already. I've been using Gnome.)
I boggle.
As an ass-backward example, consider sickle cell anemia. The sickle cell trait (Ss), whereby an individual has one copy of the gene but one copy of the non-sickle gene, conveys a bit of immunity to malaria. So there was environmental selection for people who are Ss. People who were ss (two normal genes) died at higher rates from malaria. The catch is that people with two copies of the gene (SS) die from sickle cell anemia, often before they reproduce.
Let's say that for a particular gene locus, you can have one of two versions of a gene. People with XX live fine. People with YY live fine. People with XY drop dead. If 90% of, say, Asians have YY, and 90% of, say, blacks have XX, then you have a very high probability of getting XY in a resultant offspring. In effect, you're speciated. These two races of people can't reproduce with each other at a reasonable rate.
What was necessary for survival in, say, Africa, was entirely different than what was necessary in England. As a result, there was divergent evolution with people in each area subtly specializing in different things. Now, this didn't stop, say, the Saxons and the Africans from being able to reproduce, but it easily could have. It just happens that humans apparently have more redundancy than a butterfly does, which is why simpler organisms are studied for this kind of research.
That doesn't in any way excuse racism, of course, which was the point (I think) of the parent. However, it does raise questions about things like intelligence, athletic performance, and everything else. (Nobody balks at someone putting forth the idea that blacks tend to be better athletes than whites, for example. Change 'athlete' to 'quantum theorists' and everyone blows up. It's the same thing, just a different trait. And yes, this is just an example.)
Interestingly enough, such applications go for thousands. The problem is that people just starting out, medical students, etc., can't afford that. There's a massive disparity between the free and low cost software out there, and the $10,000 packages. That isn't surprising.
What is surprising is just how absolutely terrible the OSS in this field is. For reference, in order for me to consider it, the software package has to support a Palm and sync to a desktop application.
And no, I'm not just bitching about it and having no intention to do something about it. As soon as I get my Lifedrive to actually work with gpilot (any hints, guys?) I'm going to start on this.
Not only that, but in a real, honest-to-God emergency, you can send Morse in a wide-band pattern using a car battery.(And it better be a real damned emergency or you're going to get eviscerated by the FCC and the Ham community alike.)
At that point, it becomes very, very important and useful for public safety and disaster management to have a communications network that is largely self contained. (Read: Handheld 5W Ham sets, people hooking their base stations up to car batteries or generators, etc.)
This is precisely the reason that as soon as I get rid of some exceedingly annoying medical certifications I have to deal with in the next few weeks, I intend to get a ham license.
With that in mind, O' Slashdot, perhaps those who know this stuff cold can enlighten the forum: Is there much point, all in all, in going for the higher level licenses, or should I just stick with the low-band entry level ones?
Because then they'd demand DRM, and we all know how pirates on the evil Internet respond to having DRM installed in their retinas.
I'm not sure PoE wold help that much, since the power on the Ethernet line has to come from somewhere. I can see an application on an enterprise level with a huge, centralized UPS that continues powering the Ethernet line, allowing systems to save. But then we're back where we started before, with either having to have a battery in the machine itself anyway, or the router becoming a toaster from having to supply a ton of power on every switch port.
Or did I miss something?
The 3GHz will run Windows great, performance wise. It will also run Linux great. No surprise there.
The 500MHz machine runs Windows like crap. It runs Linux really well, all things considered. Compile times are reasonable on it. It handles everything I throw at it, including full routing functions, a VPN, SMTP server, web server, some PHP stuff, etc. etc. X11 might be scary.
The P133 won't run Windows at all. It runs Linux just fine. I use it as a backend fileserver primarily, along with some database functions. It does this just fine. Compile times SUCK GOATS. I mean, they're BAD. I kicked off a compile of gcc 3.4.4 as an upgrade and was surprised (but not shocked) when it was still going 12 hours later. I haven't even tried since I retired this machine from desktop use about 8 years ago, but X11 would probably look like I was running it through a 2400baud dialup.
Some of it is feature bloat. This is a good thing, so long as those features can be turned off when necessary. Some of it is just "bad" programming (bad in the sense that it assumes the user to have more processing power and memory capacity than he does). In the case of my P133, the next project is just to distcc everything over to the 500MHz, 3GHz Athlon (when it isn't in Windows to play games or something), and a P4 3GHz laptop, and cut the P133 out of the loop for its own compilations, because letting it have any part of the compiling will just slow the entire process down.
What is a "practical system?" To most users, that's being able to run a GUI desktop, play some games, do some GUI word processing, etc. My P133 and my 500MHz machine are both pretty unsuited for that, though the latter might be passable. So I don't use them for that.
The significant difference is that if for some bizarre reason I ever wanted to, I could boot up the 20MHz 386 machine I have in a closet with 32MB of RAM, and use it for something using Linux. (I can't think of what, though, short of really low-power file serving or maybe running Postgres to do nothing other than log syslog messages from the network.)
It comes down to testing the hell out of the software. Even then, there is always the possibility that you've missed something, and that something will blow up in your face. At least you can do your part to help ensure that it isn't something as mundane as a buffer overrun or screwed up memory free().
This leads to seeing whether the automata will diversify their labor or take equal shares. Will there be some automata who somehow become freeloaders? Nothing but breeders? Builders? Farmers?
This is an interesting experiment if it can be done robustly enough to actually show something.
As for WiFi's security, it's flawed, and slows down attackers rather than stopping them. WEP can be broken relatively easily, and hiding your SSID doesn't save you either contrary to what some people might think.
The real way to handle WiFi security is to open a VPN with strong encryption to your router, and route everything through that VPN. If you're concerned about unauthorized people syncing to the network, MAC address filter *and* require some kind of cryptographic key exchange with the router prior to opening the communication. The same can apply for wired Ethernet; run a VPN between physically unsecured bits of cable and you bypass that problem.
Yes, security is a pain in the ass.