You completely missed my point. Take geometry, for instance. Mathematics is a kind of 'pure' science, since we don't have to worry about philosophical problems like the limits or nature of observability-- we don't actually observe anything. But we still need to accept some basic facts, like the idea of a point. Try to prove that a point has one dimension. You can't. This is because we define this thing called a 'point' to be a thing that has one dimension. In fact, if we try to be more specific with our definition, we find that our understanding of the thing becomes less and less precise. Therefore, our understanding of a point is axiomatic, and we simply must accept it. Euclid attempted to compile a list of all the axioms of geometry, and many of them are still important tools in a modern mathematician's toolset.
Science has a few axioms. One central belief of science you mentioned is that valid theory must be observable and testable. Why? We know, a priori, that science can't prove that anything is true, since all empirical evidence is, by definition a posteriori, and unable to prove anything. There are an infinite number (or at least a finite number so large as to be practically infinite) of testable alternatives to your theory. So why bother? Well, science posits that, with evidence, we can approximate the truth. As far as I am aware, this assertion is untestable, and yet it is a central tenet of science. This is essentially science's dogma.
But as I mentioned before, these kinds of beliefs are OK in my book, since, like with the mathematician's acceptance of the idea of the point, these beliefs make the process of science possible. Logic requires that there be some premise to go on in order to function. Where do you start? You have to accept something, somewhere, without proof.
Your blanket disregard for my statement was simply illogical. I hope you don't plan on getting very far in science, since the scientific method has a rather deep dependence on, er... logic.
No, OP is correct. Science is a system of belief. You believe in certain axioms, which are unproveable by definition. You accept them because it makes the conversation possible. For instance, in order to believe in science, you need to have faith that there is such a thing as 'rationality'-- namely, that the world is intelligible, logically consistent, and so on. It's conceivable that our universe is unintelligible to us. Science may not explicitly reject this idea, but it certainly does so implicitly.
I personally believe that accepting rationality as an axiom is an easy compromise between unhealthy skepticism and full-out acceptance of religious doctrine, because "skepticism" doesn't even make sense without "rationality" (you can't say that an argument is 'invalid' unless you can say that the argument is 'irrational'). The difference between science and something like Christianity is that science has a built-in mechanism for self-correction. It may not work quickly, or perfectly, but you can be damn sure that over time, scientific theory will better fit the facts. This is not possible for a system of belief that relies on doctrine.
I don't see any problems with accepting that science is a system of belief, because as a former student of philosophy, I know the kinds of problems you can run into when you start throwing everything out. Read some Kant or Descartes and you'll see what I mean. Quagmire. My cavalier attitude gives certain anally-retentive people an icky feeling.
Besides, science brought us lasers. Think about that. When did Jesus last bring us lasers?
Good alternatives to password authentication have been around in the UNIX world for a long time. You can use certificates, keys, one-time-passwords, and so on. Since no one is trying to lock you in to a particular solution, you can often mix and match them to suit your particular needs.
I've been using ssh keys quite successfully, and we're currently implementing a VPN setup using authpf. We're using ssh keys + passwords for this. If a machine is stolen; it's not a big deal. The theif has the key but not the password. Passwords are useless unless you have the key. So authentication is essentially tied to a machine and a user in this case. It's not foolproof, but it works well enough for our needs.
You might want to consider running your Windows apps within a virtualization environment, so that you can take advantage of some of the existing authentication solutions out there for other host OSes. Can't do Fast User Switching on a Windows box? Not a problem. Just switch between VMs. Don't know if this would be as fast as you want, but if you're smart enough, it'll be a lot cheaper.
They can at least vote on the voter referendums. I think it's safe to assume that our hypothetical ignoramuses are at least smart enough to figure out if they want to allow alcohol sales in grocery stores or not. If they aren't confident enough to fill in the rest, then they can leave those parts blank.
Yeah, here's the valid argument: common carrier status. ISPs are going to argue that it shouldn't be their responsiblity. This is where your analogy fails-- ISPs are the road; your machine is the car. LIke you said, the road builders have nothing to do with your shitty car.
But yeah, SPAM is a scourge. We need to treat it like one. Microsoft desperately needs to clean up their act. Someone I met recently called Windows a "virus runtime environment". It got some chuckles, but it's also true.
There's a good reason for that-- language fads change so often, that CS programs would be irresponsible if they didn't focus on the universal stuff. The choice of language doesn't really matter-- someone who knows the basics, like algorithms and data structures, can pick up a new language quickly. There is a huge difference in quality between code written by someone who knows the language and not the algorithms, and the code written by someone who knows the algorithms and not the language. The person in the latter case only needs to learn the syntax; the person in the former case needs to reinvent the wheel (or at least know where to quickly learn about wheels).
You have to do what every other network engineer/sysadmin on the planet did: work as tech support until you have the experience. As shitty as tech support can get, trust me, it's valuable experience when you become an engineer. Mostly because of all the years of having to deal with frustrated users, you've slowly accumulated the knowledge that end-users are fscking idiots;^)
But in all seriousness, that experience does put your decisions into perspective. You know EXACTLY how much pain just yanking that network cable will cause, and you know WHY it's more of a challenge to roll out Linux to your desktops than Windows. Employers want to be sure that you don't just have book knowledge. Just suck it up and be the broken-keyboard-swapper for awhile; if you're smart, you'll move to more interesting things quickly.
The whole common carrier thing is more complicated than this. Of course, I agree that Verizon should not have dropped these people even though most of us find the site's content to be questionable, if not offensive. But there are lots of kinds of traffic that an ISP needs to filter, and they have been traditionally reluctant to do so.
For example, DOS attacks. You could argue that DOS attacks prevent the system from working correctly, and so they must be blocked by th ISP to ensure proper functionality. And I think that most network engineers would agree with you. But see, here, we're starting down the slope a little-- when we say a common carrier does not block any traffic, we don't really mean "any" traffic. It just blocks the really bad stuff. In any case, we could probably make a strong argument that blocking this stuff doesn't violate their common carrier status. Same thing goes for DNS cache poisoners, BGP manipulators, and so on.
OK, now what about RPC worms? Certainly, they make make Help Desk techs unhappy, and probably a fair number of sysadmins as well, but is it breaking the system? Not at all-- the Internet is working fine, delivering those RPC buffer overflow attacks to their unsuspecting victims, turning their machines into botnets and whatnot. But a lot of people have been pushing for ISPs to do something about this-- certainly, I have contacted a number of ISP abuse departments myself. So here's another class of traffic that we want the ISPs to deal with-- filter-- before it even gets to us. Does this violate their common carrier status? I don't know-- it's in a grey area.
Here's another one-- SPAM. SPEWS is a great example of a DNSRBL whose sole existence is to place pressure on ISPs to drop abusive customers. Actually doing so certainly technically violates the idea of a common carrier, even if it doesn't violate the spirit, unless you really consider SPAM to be a valuable bit of Internet communication. Some spammers might, I guess. But I think there's probably nearly universal support for ISPs to drop the really awful spammers.
The point of all this is that common carrier status is not so cut-and-dry. Where do we draw the line? There's also the question of a company's right not to do business with someone. If an ISP decides not to accept the business of a questionable website, is this filtering/censorship, or just a good business decision? I don't think this is at all clear.
Man pages. Lots of them. For everything, even system calls. And an experience blissfully free of dependency hell, which, as a Gentoo user, I'm sure you're quite familiar with.
Of course, you can pretty much say goodbye to bleeding edge stuff and complicated GUI apps like Ardour, etc.
A guy who used to work for me described it this way: people want a computing appliance.
You're right, of course, but the problem is that computers are not appliances. Appliances are anathema to the whole idea of general-purpose computers, which are supposed to be flexible. You have to have some training to know how to frame a problem in a way that a computer can solve it for you. Bad interfaces aside, I find that many people who run into issues with computers do so because computers are so extremely literal. Most people simply cannot, or refuse to, wrap their minds around a computational task. So yeah, people are stupid and/or lazy. I think the GP is spot-on.
That said, the UNIX philosophy deals with these kinds of problems in an elegant way. You learn small, flexible tools one piece at a time. 'cat' is as useful to a guru as it is to a newbie. I learn new things about 'egrep' every day, and egrep only does one basic task. And with those two tools-- and some other simple ones-- I run circles around the inflexible software designed for the boneheaded network engineers I work with. People who are too lazy to do it right the first time, and empty their pockets looking for an easy solution. Learn the fucking tools, guys! Using your BRAIN is the easy way.
Toasters and cars have rigidly-defined roles. Toasters toast and cars drive (although a previous vehicle of mine toasted as well... long story). The simplest "role" I can think of for a computer is that of a "thinking machine". You need to THINK to use it-- it's a tool for thinking! With some ingenuity, people can make computers that run toaster and cars. That's powerful!
I don't really give a shit if people don't want to use UNIX. I'm happy using it myself. GNOME and KDE help lessen the learning curve for people who are used to GUIs, but if someone wants to learn how to use a real computer, they're going to have to get dirty. "bit twiddling" is how you get dirty.
Because we already have interprocess communication, pipes, symbolic links, a variety of fast and/or reliable filesystems to choose from, POSIX-compatibility, privilege separation, a real application privilege model, and on and on. So we don't need hacks like virtualized registries, better Windows APIs (because we don't need to use the undocumented internal APIs... because our application model is not some huge monolithic thing that requires RPC to talk to other programs), yada, yada. Last I checked, you could use Ardour for real audio production; JACK and ALSA seem to work pretty well on Linux, and the audio subsystem on the Mac OS doesn't seem too shabby either. Aqua already offloads tasks to the GPU, and I understand there's been work in Xorg to do the same thing. What are we missing?
Vista is a bloated piece of crap. If you want to play games, fine, buy Vista. If you just need do work, here's a nickel kid, go buy yourself a real computer.
Not really. Mac OS X is Mach-O/BSD with a GNU, BSD, and Apple userland. Actually, it's pretty fucking weird, especially if you think you can carry a lot of your BSD knowledge over to Mac OS. I mean... bash is the default shell. That's pretty un-BSD if you ask me.
The OpenBSD source repository is already in CVS format, everyone knows how to use it, and it works. What's needed are bug fixes. Hence OpenCVS. What you'll find out if you stick around with BSD for long is that the developers like new features in the OS, but hate them in the toolchain.
Re:Hardware Crypto Accelerators
on
OpenBSD 4.0 Released
·
· Score: 2, Informative
I have a Soekris vpn1401 and it works well, although I don't believe all the features are supported. IIRC, this is because hifn has not been forthcoming with their documentation. The vpn1201 is known to work as well. I'm not sure if later revisions (like the lan1461) work-- OpenBSD does not have a good relationship with hifn at the moment. BTW, I haven't done any benchmarking with my 1401, but the machine handles crypto much faster with than without it. That's all I can say.
I can't believe that anyone puts up with this nowadays. We need to tell people to get themselves over to http://openbsd.org/ftp.html (or whichever you prefer) and download a real OS, where you can do exactly what you want. I spent the weekend at NYCBSDCon and it was really exciting to be around people who could just do what they wanted with their software when they had a good idea. I can't believe anyone puts up with this shit anymore-- especially since we know the next version coming around is going to be even more of a PITA than we one we already have. I'm dreading walking back into work tomorrow and having to go through the same old litany of shit like "why can't I open Microsoft Works documents with Microsoft Word?" You can't because your OS wasn't designed with anything in mind except taking your money.
Beer and yeast infections are two sides of the same coin, too. PGP/GPG protects you. DRM protects media companies against you. See where I'm going with this? Saying that BSD will soon have DRM is like saying that beer is a yeast infection. True, in some sense, but totally misleading. You're implying that encryption is B-A-D.
Not only that-- why do we have to have the results so soon? Why not wait a week or two to make sure the votes are counted correctly? I mean, we have to keep the bozos for 4 years-- what's two weeks of waiting?
WTF are you talking about? DRM in BSD? Not in my lifetime. If you're talking about public-key cryptography, well yeah. I wouldn't use a free system that didn't have that. They're two sides of the same coin, the difference being that you're not forced to use public-key cryptography.
Being from Boston, I am obligated to like Sam Adams. But the locals really like Harpoon. If you care for IPAs, you need to try Harpoon IPA the next time you're around here.
Thanks for the great interview. Just finished setting up another MySQL DB just this morning. I suggest you pick up a Duchesse de Bourgogne (Flemish beer) as a reward for all your hard work:^)
Sounds to me like you realize GIMP's UI shortcomings but are afraid to admit it. Is this a little sore spot?
I do use xv, BTW. Yeah, it's completely different. I personally have no trouble adapting, but if I were to give the GIMP to our Art department here, I'd be taking it back by the end of the day. The $20k site license is worth not having to hear the cacophony of screaming over GIMP's UI.
Dell's stuff is OK. Their main selling point is management software that comes bundled with the hardware. It's true-- if you want a simple 1U with integrated GbE, PERC RAID controller, and so on, you can't get much cheaper than a Dell machine.
IF, that is, you actually want to use Dell's software. We'd been running OpenBSD for years on Dell hardware, and then Dell threw us for a loop and started substituting in integrated Adaptec controllers to replace the AMI ones for RAID-- without notice. This was around the time when OpenBSD was threatening to remove the Adaptec drivers from the tree. So for awhile, we bought AMI PCI-E cards to replace the Adaptecs, and that worked fine, but then I did the math... and I realized that Dell's value was in the software, which we weren't using at all.
For a lot cheaper, I can put together a 3U machine from "commodity" parts-- all the parts can be bought off the shelf except for the chassis and PSU (which I get from SuperMicro). All of the parts are exactly what I want, well-supported in OpenBSD, and replaceable. The last point is a major consideration for us. What do you do when an integrated controller goes bad in a Dell machine? If it's under warranty, you call and get it replaced. That can take days (although to Dell's credit, you can get it much faster if you pay for a more expensive service plan). With commodity parts, all of your servers have the same basic configuration-- you keep extra parts on the shelf (even the mobo and chassis), and you still save. The big savings comes from not having to pay $1-2k/yr per machine for a support contract.
The eureka moment came to me while I was waiting for parts to arrive for a machine that was covered under a contract. I was thinking-- if only I had spare parts on hand this would have been fixed by now...
Anyway, all of Dell's other stuff, consumer items, is just the same as anyone else's. It's probably good, sure, but it's probably the same (or very similar) to what you'd get from HP or whoever else builds PCs these days. In my mind, it's crazy to buy a machine like a Dell through a VAR like PC Connection when Dell is essentially a VAR themselves these days. They don't manufacture anything unique. Dell obviously cuts costs by selling directly to consumers, so their volume discount almost cancels out the assembly cost. But you could probably build the exact same machine by buying parts from NewEgg.
The problem is that the GIMP's UI is not consistent with any other UI I've ever come across. Sure, it's probably easy-to-use once you already know it. But why not make it the same as other apps? Photoshop, among many other apps, has a bounding window-- this helps keep your workspace organized. Photoshop has palettes that actually float, instead of getting lost behind numerous other windows. You don't have to navigate through menu hell just to apply a filter. And so on. I'm sure these points have been flogged to death on lists before, but IMHO, until the GIMP changes these very simple UI items, most graphics professionals will stay away. Which is too bad-- the GIMP is otherwise a great program.
Slashdot Mirror
You completely missed my point. Take geometry, for instance. Mathematics is a kind of 'pure' science, since we don't have to worry about philosophical problems like the limits or nature of observability-- we don't actually observe anything. But we still need to accept some basic facts, like the idea of a point. Try to prove that a point has one dimension. You can't. This is because we define this thing called a 'point' to be a thing that has one dimension. In fact, if we try to be more specific with our definition, we find that our understanding of the thing becomes less and less precise. Therefore, our understanding of a point is axiomatic, and we simply must accept it. Euclid attempted to compile a list of all the axioms of geometry, and many of them are still important tools in a modern mathematician's toolset.
Science has a few axioms. One central belief of science you mentioned is that valid theory must be observable and testable. Why? We know, a priori, that science can't prove that anything is true, since all empirical evidence is, by definition a posteriori, and unable to prove anything. There are an infinite number (or at least a finite number so large as to be practically infinite) of testable alternatives to your theory. So why bother? Well, science posits that, with evidence, we can approximate the truth. As far as I am aware, this assertion is untestable, and yet it is a central tenet of science. This is essentially science's dogma.
But as I mentioned before, these kinds of beliefs are OK in my book, since, like with the mathematician's acceptance of the idea of the point, these beliefs make the process of science possible. Logic requires that there be some premise to go on in order to function. Where do you start? You have to accept something, somewhere, without proof.
Your blanket disregard for my statement was simply illogical. I hope you don't plan on getting very far in science, since the scientific method has a rather deep dependence on, er... logic.
No, OP is correct. Science is a system of belief. You believe in certain axioms, which are unproveable by definition. You accept them because it makes the conversation possible. For instance, in order to believe in science, you need to have faith that there is such a thing as 'rationality'-- namely, that the world is intelligible, logically consistent, and so on. It's conceivable that our universe is unintelligible to us. Science may not explicitly reject this idea, but it certainly does so implicitly.
I personally believe that accepting rationality as an axiom is an easy compromise between unhealthy skepticism and full-out acceptance of religious doctrine, because "skepticism" doesn't even make sense without "rationality" (you can't say that an argument is 'invalid' unless you can say that the argument is 'irrational'). The difference between science and something like Christianity is that science has a built-in mechanism for self-correction. It may not work quickly, or perfectly, but you can be damn sure that over time, scientific theory will better fit the facts. This is not possible for a system of belief that relies on doctrine.
I don't see any problems with accepting that science is a system of belief, because as a former student of philosophy, I know the kinds of problems you can run into when you start throwing everything out. Read some Kant or Descartes and you'll see what I mean. Quagmire. My cavalier attitude gives certain anally-retentive people an icky feeling.
Besides, science brought us lasers. Think about that. When did Jesus last bring us lasers?
She was on On Point last night. Everyone else on the list is either braindead or actually dead (or nonexistent). Maybe that was a criteria.
SMB/CIFS, MAPI, Microsoft DNS, RPC over HTTP, Office APIs, and so on... The list is here.
Good alternatives to password authentication have been around in the UNIX world for a long time. You can use certificates, keys, one-time-passwords, and so on. Since no one is trying to lock you in to a particular solution, you can often mix and match them to suit your particular needs.
I've been using ssh keys quite successfully, and we're currently implementing a VPN setup using authpf. We're using ssh keys + passwords for this. If a machine is stolen; it's not a big deal. The theif has the key but not the password. Passwords are useless unless you have the key. So authentication is essentially tied to a machine and a user in this case. It's not foolproof, but it works well enough for our needs.
You might want to consider running your Windows apps within a virtualization environment, so that you can take advantage of some of the existing authentication solutions out there for other host OSes. Can't do Fast User Switching on a Windows box? Not a problem. Just switch between VMs. Don't know if this would be as fast as you want, but if you're smart enough, it'll be a lot cheaper.
I suppose that would explain why the Chinese ban them.
They can at least vote on the voter referendums. I think it's safe to assume that our hypothetical ignoramuses are at least smart enough to figure out if they want to allow alcohol sales in grocery stores or not. If they aren't confident enough to fill in the rest, then they can leave those parts blank.
Yeah, here's the valid argument: common carrier status. ISPs are going to argue that it shouldn't be their responsiblity. This is where your analogy fails-- ISPs are the road; your machine is the car. LIke you said, the road builders have nothing to do with your shitty car.
But yeah, SPAM is a scourge. We need to treat it like one. Microsoft desperately needs to clean up their act. Someone I met recently called Windows a "virus runtime environment". It got some chuckles, but it's also true.
There's a good reason for that-- language fads change so often, that CS programs would be irresponsible if they didn't focus on the universal stuff. The choice of language doesn't really matter-- someone who knows the basics, like algorithms and data structures, can pick up a new language quickly. There is a huge difference in quality between code written by someone who knows the language and not the algorithms, and the code written by someone who knows the algorithms and not the language. The person in the latter case only needs to learn the syntax; the person in the former case needs to reinvent the wheel (or at least know where to quickly learn about wheels).
You have to do what every other network engineer/sysadmin on the planet did: work as tech support until you have the experience. As shitty as tech support can get, trust me, it's valuable experience when you become an engineer. Mostly because of all the years of having to deal with frustrated users, you've slowly accumulated the knowledge that end-users are fscking idiots ;^)
But in all seriousness, that experience does put your decisions into perspective. You know EXACTLY how much pain just yanking that network cable will cause, and you know WHY it's more of a challenge to roll out Linux to your desktops than Windows. Employers want to be sure that you don't just have book knowledge. Just suck it up and be the broken-keyboard-swapper for awhile; if you're smart, you'll move to more interesting things quickly.
The whole common carrier thing is more complicated than this. Of course, I agree that Verizon should not have dropped these people even though most of us find the site's content to be questionable, if not offensive. But there are lots of kinds of traffic that an ISP needs to filter, and they have been traditionally reluctant to do so.
For example, DOS attacks. You could argue that DOS attacks prevent the system from working correctly, and so they must be blocked by th ISP to ensure proper functionality. And I think that most network engineers would agree with you. But see, here, we're starting down the slope a little-- when we say a common carrier does not block any traffic, we don't really mean "any" traffic. It just blocks the really bad stuff. In any case, we could probably make a strong argument that blocking this stuff doesn't violate their common carrier status. Same thing goes for DNS cache poisoners, BGP manipulators, and so on.
OK, now what about RPC worms? Certainly, they make make Help Desk techs unhappy, and probably a fair number of sysadmins as well, but is it breaking the system? Not at all-- the Internet is working fine, delivering those RPC buffer overflow attacks to their unsuspecting victims, turning their machines into botnets and whatnot. But a lot of people have been pushing for ISPs to do something about this-- certainly, I have contacted a number of ISP abuse departments myself. So here's another class of traffic that we want the ISPs to deal with-- filter-- before it even gets to us. Does this violate their common carrier status? I don't know-- it's in a grey area.
Here's another one-- SPAM. SPEWS is a great example of a DNSRBL whose sole existence is to place pressure on ISPs to drop abusive customers. Actually doing so certainly technically violates the idea of a common carrier, even if it doesn't violate the spirit, unless you really consider SPAM to be a valuable bit of Internet communication. Some spammers might, I guess. But I think there's probably nearly universal support for ISPs to drop the really awful spammers.
The point of all this is that common carrier status is not so cut-and-dry. Where do we draw the line? There's also the question of a company's right not to do business with someone. If an ISP decides not to accept the business of a questionable website, is this filtering/censorship, or just a good business decision? I don't think this is at all clear.
Man pages. Lots of them. For everything, even system calls. And an experience blissfully free of dependency hell, which, as a Gentoo user, I'm sure you're quite familiar with.
Of course, you can pretty much say goodbye to bleeding edge stuff and complicated GUI apps like Ardour, etc.
A guy who used to work for me described it this way: people want a computing appliance.
You're right, of course, but the problem is that computers are not appliances. Appliances are anathema to the whole idea of general-purpose computers, which are supposed to be flexible. You have to have some training to know how to frame a problem in a way that a computer can solve it for you. Bad interfaces aside, I find that many people who run into issues with computers do so because computers are so extremely literal. Most people simply cannot, or refuse to, wrap their minds around a computational task. So yeah, people are stupid and/or lazy. I think the GP is spot-on.
That said, the UNIX philosophy deals with these kinds of problems in an elegant way. You learn small, flexible tools one piece at a time. 'cat' is as useful to a guru as it is to a newbie. I learn new things about 'egrep' every day, and egrep only does one basic task. And with those two tools-- and some other simple ones-- I run circles around the inflexible software designed for the boneheaded network engineers I work with. People who are too lazy to do it right the first time, and empty their pockets looking for an easy solution. Learn the fucking tools, guys! Using your BRAIN is the easy way.
Toasters and cars have rigidly-defined roles. Toasters toast and cars drive (although a previous vehicle of mine toasted as well... long story). The simplest "role" I can think of for a computer is that of a "thinking machine". You need to THINK to use it-- it's a tool for thinking! With some ingenuity, people can make computers that run toaster and cars. That's powerful!
I don't really give a shit if people don't want to use UNIX. I'm happy using it myself. GNOME and KDE help lessen the learning curve for people who are used to GUIs, but if someone wants to learn how to use a real computer, they're going to have to get dirty. "bit twiddling" is how you get dirty.
Because we already have interprocess communication, pipes, symbolic links, a variety of fast and/or reliable filesystems to choose from, POSIX-compatibility, privilege separation, a real application privilege model, and on and on. So we don't need hacks like virtualized registries, better Windows APIs (because we don't need to use the undocumented internal APIs... because our application model is not some huge monolithic thing that requires RPC to talk to other programs), yada, yada. Last I checked, you could use Ardour for real audio production; JACK and ALSA seem to work pretty well on Linux, and the audio subsystem on the Mac OS doesn't seem too shabby either. Aqua already offloads tasks to the GPU, and I understand there's been work in Xorg to do the same thing. What are we missing?
Vista is a bloated piece of crap. If you want to play games, fine, buy Vista. If you just need do work, here's a nickel kid, go buy yourself a real computer.
Not really. Mac OS X is Mach-O/BSD with a GNU, BSD, and Apple userland. Actually, it's pretty fucking weird, especially if you think you can carry a lot of your BSD knowledge over to Mac OS. I mean... bash is the default shell. That's pretty un-BSD if you ask me.
The OpenBSD source repository is already in CVS format, everyone knows how to use it, and it works. What's needed are bug fixes. Hence OpenCVS. What you'll find out if you stick around with BSD for long is that the developers like new features in the OS, but hate them in the toolchain.
I have a Soekris vpn1401 and it works well, although I don't believe all the features are supported. IIRC, this is because hifn has not been forthcoming with their documentation. The vpn1201 is known to work as well. I'm not sure if later revisions (like the lan1461) work-- OpenBSD does not have a good relationship with hifn at the moment. BTW, I haven't done any benchmarking with my 1401, but the machine handles crypto much faster with than without it. That's all I can say.
I can't believe that anyone puts up with this nowadays. We need to tell people to get themselves over to http://openbsd.org/ftp.html (or whichever you prefer) and download a real OS, where you can do exactly what you want. I spent the weekend at NYCBSDCon and it was really exciting to be around people who could just do what they wanted with their software when they had a good idea. I can't believe anyone puts up with this shit anymore-- especially since we know the next version coming around is going to be even more of a PITA than we one we already have. I'm dreading walking back into work tomorrow and having to go through the same old litany of shit like "why can't I open Microsoft Works documents with Microsoft Word?" You can't because your OS wasn't designed with anything in mind except taking your money.
Beer and yeast infections are two sides of the same coin, too. PGP/GPG protects you. DRM protects media companies against you. See where I'm going with this? Saying that BSD will soon have DRM is like saying that beer is a yeast infection. True, in some sense, but totally misleading. You're implying that encryption is B-A-D.
Ah, what's the point explaining? You're an idiot.
Not only that-- why do we have to have the results so soon? Why not wait a week or two to make sure the votes are counted correctly? I mean, we have to keep the bozos for 4 years-- what's two weeks of waiting?
WTF are you talking about? DRM in BSD? Not in my lifetime. If you're talking about public-key cryptography, well yeah. I wouldn't use a free system that didn't have that. They're two sides of the same coin, the difference being that you're not forced to use public-key cryptography.
Being from Boston, I am obligated to like Sam Adams. But the locals really like Harpoon. If you care for IPAs, you need to try Harpoon IPA the next time you're around here.
:^)
Thanks for the great interview. Just finished setting up another MySQL DB just this morning. I suggest you pick up a Duchesse de Bourgogne (Flemish beer) as a reward for all your hard work
Sounds to me like you realize GIMP's UI shortcomings but are afraid to admit it. Is this a little sore spot?
I do use xv, BTW. Yeah, it's completely different. I personally have no trouble adapting, but if I were to give the GIMP to our Art department here, I'd be taking it back by the end of the day. The $20k site license is worth not having to hear the cacophony of screaming over GIMP's UI.
Dell's stuff is OK. Their main selling point is management software that comes bundled with the hardware. It's true-- if you want a simple 1U with integrated GbE, PERC RAID controller, and so on, you can't get much cheaper than a Dell machine.
IF, that is, you actually want to use Dell's software. We'd been running OpenBSD for years on Dell hardware, and then Dell threw us for a loop and started substituting in integrated Adaptec controllers to replace the AMI ones for RAID-- without notice. This was around the time when OpenBSD was threatening to remove the Adaptec drivers from the tree. So for awhile, we bought AMI PCI-E cards to replace the Adaptecs, and that worked fine, but then I did the math... and I realized that Dell's value was in the software, which we weren't using at all.
For a lot cheaper, I can put together a 3U machine from "commodity" parts-- all the parts can be bought off the shelf except for the chassis and PSU (which I get from SuperMicro). All of the parts are exactly what I want, well-supported in OpenBSD, and replaceable. The last point is a major consideration for us. What do you do when an integrated controller goes bad in a Dell machine? If it's under warranty, you call and get it replaced. That can take days (although to Dell's credit, you can get it much faster if you pay for a more expensive service plan). With commodity parts, all of your servers have the same basic configuration-- you keep extra parts on the shelf (even the mobo and chassis), and you still save. The big savings comes from not having to pay $1-2k/yr per machine for a support contract.
The eureka moment came to me while I was waiting for parts to arrive for a machine that was covered under a contract. I was thinking-- if only I had spare parts on hand this would have been fixed by now...
Anyway, all of Dell's other stuff, consumer items, is just the same as anyone else's. It's probably good, sure, but it's probably the same (or very similar) to what you'd get from HP or whoever else builds PCs these days. In my mind, it's crazy to buy a machine like a Dell through a VAR like PC Connection when Dell is essentially a VAR themselves these days. They don't manufacture anything unique. Dell obviously cuts costs by selling directly to consumers, so their volume discount almost cancels out the assembly cost. But you could probably build the exact same machine by buying parts from NewEgg.
The problem is that the GIMP's UI is not consistent with any other UI I've ever come across. Sure, it's probably easy-to-use once you already know it. But why not make it the same as other apps? Photoshop, among many other apps, has a bounding window-- this helps keep your workspace organized. Photoshop has palettes that actually float, instead of getting lost behind numerous other windows. You don't have to navigate through menu hell just to apply a filter. And so on. I'm sure these points have been flogged to death on lists before, but IMHO, until the GIMP changes these very simple UI items, most graphics professionals will stay away. Which is too bad-- the GIMP is otherwise a great program.