The Google study didn't include drives running at the insane temperatures they may reach in consumer hardware. Especially the 7200RPM drives. I had a drive running at 55-60C in a tiny box made for a mini-ITX board. 45C is not hard to reach in a normal but badly cooled tower case. Google's study topped out at somewhere about 30C and the graph was going up at that point.
The problem is that SSL happens before any HTTP does, and SSL is a general mechanism that can be used for any kind of TCP connection.
How does the webserver know what to give you when foo.com and bar.com map to the same IP address, and the browser requests something like index.html that exists on both? This works only because when the browser makes the request it also tells the webserver which domain it was trying to access. The browser sends something like this:
GET/index.html HTTP/1.1 Host: foo.com
Now, this breaks for SSL, because SSL happens before the connection is established, so there's no way to decide which certificate to use based on the domain.
To fix to this is adding the support directly to SSL. rfc4336 contains a mechanism to do this with TLS.
Some of those geeks work at places like universities and large companies and make purchase decisions. Others give advice to less knowledgeable people.
An important thing to note here is that a dedicated soundcard is no longer a necessary component of a computer due to onboard sound. A large part of Creative's market are going people who decide on their own to buy a soundcard for some reason, and which card they choose will depend quite heavily on some geek's opinion.
4 CPUs is quite different than the days of Socket A where I could almost grab a Duron or Athlon from any system a drop it in another. There was a Socket A for every speed, series and cache size that AMD had in production it seemed.
First, sorry, I was unclear. I meant I found 4-core CPUs that fit in the 1027 socket. I'm not sure how many of them exist.
But anyway, while the socket A compatibility was nice, it's not all that great in practice. That you can use an Athlon XP 3200+ in a board with a 133MHz FSB, when the CPU supports a 200 MHz FSB doesn't make it a great idea. Same goes in reverse. It's a waste to combine components in such a way that one of them will never be able to work optimally.
Anyways, that example blew up in my face... but I meant to show that even if you don't mind paying top dollar for the top tier product, you don't even get a great product. So, I'll stick to my guns on the original intent of my post; Google and Apple still deliver a solid product at the end of the day, and that's why people like them.
Thing is, there's no unconditional "top tier" hardware. It depends on what you want to do with it.
For example, if your bottleneck is RAM speed, then adding more cores won't help you much. If you run a database then you absolutely want reliability even if speed suffers a bit. If your workload accesses data randomly, then a solid state drive could be a huge improvement. On the other hand if you access data linearly, normal hard disks could probably give you the required performance for much cheaper. If you have one single threaded application, then multiple cores are pointless. But if you concurrently process lots of small requests, a 128 core box from Sun might be the best thing ever.
Spending lots of cash on high end hardware doesn't guarantee you'll get the best performance possible. Depending on what you actually need, it's quite possible that something for half the price would have worked better.
Apple makes great hardware? Since when? I lost count of the number of posts from people talking about laptops that had to be shipped due to "logic board failure", some of them multiple times. iPods have a battery that's not replaceable by the user, even though it's the one component that's absolutely guaranteed to fail. I've seen many posts to the effect of "Don't buy first generation Apple stuff".
Now, maybe Apple makes things that are on average better quality than cheap PCs, but they don't seem to make bulletproof hardware by any measure.
Unfortunately, these days you just can't seem to find things worth the money or time they cost. If I stood outside with a hand full of cash and waved it at everyone passing by and offering that fistful of cash for the first person who could deliver to me a motherboard that's fully functional, delivers performance, is upgradeable, delivers on every bullet point on the box, fits in a standard 1U case with heatsink, and has drivers for Linux, I still couldn't it.
Well, that's kind of a niche market. But what do you mean by "has drivers for Linux"? Linux needs to recognize the disk controller for DMA, and the network card, but that's about it. You can ignore the integrated video for a server.
I've got a Dual Socket 1207 if AMD ever releases processors for it - and it promised to work with quad core when they came out... except for that hardware bug that prevented it. Oh, yeah, and that memory interface that was going to be forwards compatible fell through, too. Did I mention DMA doesn't work with Asus' Linux drivers? I shelled out $300 (mind you, I'm a student, that's big cash) for a dual socket motherboard because I wanted the best money could buy and I still got screwed. By Asus, nonetheless. The only reason I put up with them is because the alternative is BioStar who unashamedly pumps out crap that limps along for a year before dying. At least Apple and Google are delivering!
I don't get where's the part where you got screwed. You didn't specify the exact board, but it seems it should have a nForce chipset. Linux supports that fine with DMA. I also don't understand what you mean by "if AMD ever releases processors for it". I looked and there are 4 core 1027 AMD CPUs in existence. Your board may need a BIOS update to recognize them, though. Regarding memory, those boards use registered ECC memory, and so far I've not seen DDR2 800 with ECC anywhere anyway. You use ECC RAM because you want reliability, not the fastest speed possible.
I didn't like the cookie valid until 2038 back when I found about it, didn't get what's that great about gmail and still don't (I like kmail much better), am not interested in the slightest in using their office applications, and Google's merging with doubleclick didn't improve things either. Doubleclick used to have a quite bad reputation on slashdot some years ago, btw.
One of these days I'm going to figure out how to make google go through tor.
And since other posters mentioned it, I don't like Apple either.
I don't think you understand. I don't care about the specific issue. I certainly don't care about the OS 9 memory model. I'm simply using it as an example to illustrate my point. If you want, I can rewrite the whole thing for mice with one button, or try to find somebody under the delusion of that the Linux audio system doesn't suck, and the configuration of dmix is really intuitive, and use that.
Pick any deficiency in any software or hardware. Watch a fanboy claiming first that it doesn't exist, then that 99% of the users aren't affected, then that some convoluted workaround is absolutely not a big deal. It's all the same.
That you never used a Mac. Yeah, we all know that now.
Actually, yes I have. Not for long though, and for little more than web browsing and IRC.
That however doesn't have anything to do with the subject though, because I'm not commenting on the hardware, but on human behavior. It doesn't really matter whether it's Linux, Mac, Amiga or Windows. It also doesn't matter which specific feature is the issue.
One of the defining characteristics of a fanboy is being irrational, and your 99% claim simply doesn't stand up to logic. A fixed allocation is fine for a tool like a calculator. However there's no way it's going to work well for a tool that works on datasets of very variable size. Such as word processors, spreadsheets, photoshop and other office tools.
Memory is expensive, so most computers don't have that much of it. That means the default memory limit has to be low enough to work on the smallest machine it can be used on. It also has to leave room for whatever else will be running. So if you can expect Word to be started in a 32MB box, the limit must be a fraction of that. Send a large enough document, and the user will have to bump the value up.
I looked around for info on why this is necessary. My understanding is that OS 9 runs in an environment without a MMU, where each application needs a contiguous chunk of address space. That means it can't ask for more later, so it's got to get what it needs upfront. And there's no memory protection either. This is in a time where there existed OS/2, Windows 95, and Linux which had both things.
Now the issue isn't that OS 9 lacked this or that. All the alternatives also lacked one thing or another. The issue here is attitude. A realistic person would tell you that making the user to manually define how much memory an application is allowed to use is rather not ideal, but it's worth bearing with that quirk because the system itself is pretty nice. A fanboy though is going to first claim the issue doesn't exist at all, then that it only affects a tiny minority, then that it exists, but it's a completely trivial matter. (Note that still, most people have no clue what RAM is, what it's for, what's the difference between RAM and disk space, and how much they have of each). And after all that when the new version comes out without those issues they'll start evangelizing the new version with those improvements as the selling point, forgetting they claimed before it didn't bother them before in the slightest.
So it seems, as long as the choice is *NOT* Microsoft.
I dunno about other people, but I don't like Microsoft's behavior, not MS per se. Say, MS SQL Server is actually a pretty decent piece of software. I don't think anybody would complain much if they ported that to Linux.
In fact, I think a Linux version of MS Office, with ODF support would be very welcome. But they'll never to that, because it wouldn't help keep and extend their monopoly. Take the whole OOXML mess happening right now. That's precisely the sort of thing that annoys people. MS can't just compete, they've got to rule with an iron fist, and if things aren't going their way, they'll try to squish whoever's necessary to get what they want.
If MS behaved in a more IBM-like manner I doubt people would protest nearly that much.
Unless we're discussing the scheduler!!
Since when? Granted, the scheduler is a complex matter. It's a complicated piece of software, and a very important part of the kernel. It affects everybody using it. It's hard to benchmark because you can bet there's somebody with an obscure configuration who finds changes produce a much worse result than the old version. And even worse, perception of performance is subjective. Then there's that there are big egos involved when working on something this important.
But for all of that, nobody's claiming the Linux scheduler is perfect. On the contrary, there's been much debate, benchmarking and testing, and the scheduler changed several times. Linux started as a single CPU operating system. Now it can use more CPUs than Windows. There clearly have been improvements.
In fact, I think people get a mistaken perception of this. It's not that Linux is the only place where there are huge arguments about how the scheduler should work. It's that at Apple and MS they probably also have huge arguments but out of public view, and people aren't used to seeing this sort of thing in the open.
The should include it but won't & for good reason, the Vorbis codec is rapidly becoming dated and was almost all hype. Whatever happened to bitrate peeling anyway? I love FLAC:)
Well, I had to use an example and that came to mind. I think it would have made a better argument a few years ago, as with the growth of available space compression became less important. I still have an old but perfectly functional 1 GB flash player. With that capacity, ogg's quality advantage vs mp3 is nice.
It still has advantages though. MP3 and AAC have licensing issues. FLAC is too big for many purposes. I've seen games (Unreal Tournament I think?) with an ogg soundtrack. If you need audio to be compact, and don't want to pay licensing fees, I can't think anything besides Vorbis to do it.
bitrate peeling seems to have died, btw, which is a pity. Sounded like a nice idea.
Unless, of course, you dare suggest that people be free to choose proprietary software. Then all that freedom-lovin' goes right out the window, and you are a terrible person for wanting to use Photoshop instead of the GIMP.
Eh? There's plenty proprietary software on Linux. Vmware and commercial games for instance. I've never, ever seen somebody claiming it to be wrong to use vmware or NWN on Linux. There's a preference towards free software of course when possible.
Meh. Too few people use Ogg for it to really matter. If it were a popular, widely-used format, it might get a response. Can't you add a script to your music software to convert your oggs when you load the iPod?
Well, most of my collection is in it. And why would I buy an iPod, then convert my collection with loss of quality instead of buying a player that works without extra messing with it in the first place?
What I'm saying is that the fanboys' perception is that whatever is current is the absolute perfection. Take OS 9 for instance, where you had to set manually the amount of memory an application could use. If you brought up that setting memory is an oddly unfriendly characteristic for an OS that aims to be user friendly, you'd get a reply along the lines of "Lets see, I select the app, "get info"-->Memory and then set the amount. What's hard about that?"
Of course now that OS X is here, it's OS X what became the definition of perfection. I get the impression that many people refuse to acknowledge the existence of any faults until they're fixed, then the subject is quietly forgotten. For instance:
Everybody swore that a mouse with more than one button isn't needed, until Apple suddenly released a computer with one. The memory limit was an "advantage", because Windows would die a swap death, and "Whgat New user is going to jump in and go manilulate large excel files?", anyway? etc.
This attitude turns many people off, because: It creates a feeling that there's some sort of apple collective that many people aren't interested to join, because for them computers are a tool and not an object of religious worship. It creates the feeling of that it's hard to get a honest opinion about anything because many people are dedicated to sweeping faults under the rug. And it creates an impression of inflexibility: Either you accept the package in full, or you'd better not get it at all, because there's nothing in the middle.
It's a difference in philosophy. Linux is about freedom and choice. If you say "Linux lacks X", most of the time if you get a negative reply it'll be something like "well, go fix it, the source's there". You generally won't be flamed to a crisp for daring to suggest that say, the state of audio in Linux isn't ideal. Constructive criticism could get a positive reply. Take the guy who did Linux boot benchmarking -- it quickly resulted in optimizations of the process.
Now try to do the same with Apple. Apple is about the "experience". Either you get it, or you can go look somewhere else. If you try to suggest the iPod, iPhone or something else isn't ideal you'll often get a reply from somebody who thinks nothing Apple makes might be a bit imperfect, and that if you don't like it, something is wrong with you. Mac OS was perfect before OS X came out. I've never seen a fan reply to the complaint of the iPod's lack of ability to play Ogg Vorbis as "You know, they should really include that". If it was a Linux device somebody would have added that within a month of the iPod's release.
Come again? Democracy works only if the population is informed and in control.
How can you have a working democracy if the population isn't free? Let's see what Wikipedia says about that:
Political freedom is the absence of interference with the sovereignty of an individual by the use of coercion or aggression. The members of a free society would have full dominion over their public and private lives. The opposite of a free society would be a totalitarian state, which highly restricts political freedom in order to regulate almost every aspect of behavior.
Note: I'm going with Wikipedia here intentionally because I expect it to be biased towards the American concepts in such matters.
So for example. Take away the freedoms of association, assembly, press, religion and speech. Quite totalitarian. What kind of democracy are you going to get in a state where citizens can't form associations or political parties, press is restricted to follow the government agenda, the religion is whatever the state says it is and nothing else, and if you say something unpopular you suddenly vanish?
Actually, no. The mistake you make is thinking somebody is interested in you personally. Most often they aren't.
They're not after your vacation's photos. They're after your internet connection, so that you're the one blamed for sending a huge amount of spam.
The way around this is by using the best security you can. If you use WPA, and the neighbour is open, guess who the spammer is going to choose for their operations. Same goes for house and car security: It doesn't need to be perfect, just good enough to be more inconvenient to break through than the other ones nearby.
You can't secure a connection against abuse. Exploits can be as simple as crafting an URL, and completely normal internet activity can involve illegal content. Posting death threats for instance.
If you refer to Applied Cryptography, it's actually a very interesting and mostly readable book. IMO it's worth reading for anybody working on anything crypto related. At least part of it.
About half of it is about the theory. This includes things like cryptographic protocols, such as: How to get two parties to exchange information in such a way that one of them can't back off it once they started? A lot of that is quite esoteric, but it's quite interesting, and mostly understandable to a programmer without needing math knowledge.
The other half is a description of many crypto algorithms, such as, how DES works. This is followed by source listings. That part is probably too hard to understand for people without the relevant knowledge. This is probably is interesting for a cryptographer and includes discussion on how the algorithms came to be, but probably won't be useful to 99% of programmers as most people aren't going to implement DES from scratch.
Then there's Secrets and Lies, which IMO is a bad book for a technical person. It makes a fine gift for somebody who really doesn't get how is all this security stuff supposed to work, but for me it's a "Well, DUH" one after another. It explains what is a DoS, what is the problem with biometrics, etc. Anybody interested in the subjects will be quite familiar with all of that already.
OK, read what I said. That the disorder has to interfere with the person in a clinically significant way. Then read your link.
E. The avoidance, anxious anticipation, or distress in the feared social or performance situation(s) interferes significantly with the person's normal routine, occupational (academic) functioning, or social activities or relationships, or there is marked distress about having the phobia.
The problem is the vagueness of things like "interferes significantly". It also depends on the point of view of the one doing that evaluation, especially for children.
For example, many people still consider any kind of online communication to be "not real". Even if online you're capable of effectively leading a group of 50 people (which IMO is impressive since it's hard to force people online to do anything), that pales in comparison with the social value of getting drunk in a bar with people you barely know.
Nobody agrees on what "furry" is. I'd say it's like "computer geek". It points to a recognizable trait (this person likes computers), but inside that category you have very different people, from script kiddies to experts able to design a CPU, build a whole computer and make it work. It's wide enough to include people that are nothing alike to the point of not being able to understand each other, and people with violent disagreements. Take for example Linux, Windows and Mac fanatics. Each of those can be easily in the "computer geek" category, but it's very likely that they'll never agree on what an user friendly OS should be like.
Same goes for furries. The wide definition is something like "somebody who likes anthropomorphic animals". That includes many different people. Some are simply interested in the concept (or like to draw anthros). Some love a particular species with RL experience and identify with it. Some identify with a particular species despite having never been anywhere near it. Some identify with species that don't exist at all (mythological, or something completely new). Some take it seriously and for them it's a major component of their lives. A very few take it REALLY seriously and would go to extremes like body modification. Some like furry porn and some don't.
The more extreme behaviors aren't that common. AFAIK, very few own fursuits. They're expensive and far from everybody is interested.
In SL far from everybody is into it for the porn. I'd say your fiancee should visit Luskwood if she comes to SL, it's a PG furry place. No porn of any sort there.
Does that mean that an autistic savant, or just somebody who happened to have a calculator handy isn't human?
I have bc (a commandline calculator) running quite often. If you asked me that, I'd copy/paste, press enter, copy/paste the answer, then say "12381557655576425121, why do you ask?". By that standard you'd think I'm not human either.
Most modern cameras have a LCD screen. That means it's going to be impractical to force the user to always look into the viewfinder. Also most decent cameras can take photos on a timer. That means the eye data must be stored.
But this in turn makes it very easy to store anything as eye data. I could use my cat's eye for identification. Or find a high resolution photo on the web, print it, and try to get the camera to accept somebody else's eye.
I like this way much more than coming up with something new because it means I'd be able to keep my XML generating shell scripts, and just filter the output through a text to binary converter.
The issue has nothing to do with needing NAT. The issue is that we like the convenience of our NAT boxes, and there aren't any consumer priced firewalls that I'm aware of. I'd be delighted to use a sub $100 wireless router that had an honest to god firewall capability configurable via a friendly web client.
Well, I use a Linux box for this. Get a cheap Pentium, attach a decent sized heatsink to it, use a CompactFlash to IDE adapter and a CF card, and you should be able to run that from a fanless power supply. With such low power usage you could probably even harmlessly disable the fan on a normal supply. The hardware can be found for free.
You can also get a Linksys WRT54G, those are below $100, and there's alternate firmware for it with plenty of functionality. Some googling suggests the WRT54G can be made work with IPv6 fine.
Slashdot Mirror
You must be kidding. Furry areas are a very small part of what there is in SL.
Not sure what you mean by "the whole island", if you mean the mainland, then certainly not.
The Google study didn't include drives running at the insane temperatures they may reach in consumer hardware. Especially the 7200RPM drives. I had a drive running at 55-60C in a tiny box made for a mini-ITX board. 45C is not hard to reach in a normal but badly cooled tower case. Google's study topped out at somewhere about 30C and the graph was going up at that point.
How does the webserver know what to give you when foo.com and bar.com map to the same IP address, and the browser requests something like index.html that exists on both? This works only because when the browser makes the request it also tells the webserver which domain it was trying to access. The browser sends something like this: Now, this breaks for SSL, because SSL happens before the connection is established, so there's no way to decide which certificate to use based on the domain.
To fix to this is adding the support directly to SSL. rfc4336 contains a mechanism to do this with TLS.
Some of those geeks work at places like universities and large companies and make purchase decisions. Others give advice to less knowledgeable people.
An important thing to note here is that a dedicated soundcard is no longer a necessary component of a computer due to onboard sound. A large part of Creative's market are going people who decide on their own to buy a soundcard for some reason, and which card they choose will depend quite heavily on some geek's opinion.
First, sorry, I was unclear. I meant I found 4-core CPUs that fit in the 1027 socket. I'm not sure how many of them exist.
But anyway, while the socket A compatibility was nice, it's not all that great in practice. That you can use an Athlon XP 3200+ in a board with a 133MHz FSB, when the CPU supports a 200 MHz FSB doesn't make it a great idea. Same goes in reverse. It's a waste to combine components in such a way that one of them will never be able to work optimally.
Thing is, there's no unconditional "top tier" hardware. It depends on what you want to do with it.
For example, if your bottleneck is RAM speed, then adding more cores won't help you much. If you run a database then you absolutely want reliability even if speed suffers a bit. If your workload accesses data randomly, then a solid state drive could be a huge improvement. On the other hand if you access data linearly, normal hard disks could probably give you the required performance for much cheaper. If you have one single threaded application, then multiple cores are pointless. But if you concurrently process lots of small requests, a 128 core box from Sun might be the best thing ever.
Spending lots of cash on high end hardware doesn't guarantee you'll get the best performance possible. Depending on what you actually need, it's quite possible that something for half the price would have worked better.
Now, maybe Apple makes things that are on average better quality than cheap PCs, but they don't seem to make bulletproof hardware by any measure.
Well, that's kind of a niche market. But what do you mean by "has drivers for Linux"? Linux needs to recognize the disk controller for DMA, and the network card, but that's about it. You can ignore the integrated video for a server.
I don't get where's the part where you got screwed. You didn't specify the exact board, but it seems it should have a nForce chipset. Linux supports that fine with DMA. I also don't understand what you mean by "if AMD ever releases processors for it". I looked and there are 4 core 1027 AMD CPUs in existence. Your board may need a BIOS update to recognize them, though. Regarding memory, those boards use registered ECC memory, and so far I've not seen DDR2 800 with ECC anywhere anyway. You use ECC RAM because you want reliability, not the fastest speed possible.
Count me as an exception.
I didn't like the cookie valid until 2038 back when I found about it, didn't get what's that great about gmail and still don't (I like kmail much better), am not interested in the slightest in using their office applications, and Google's merging with doubleclick didn't improve things either. Doubleclick used to have a quite bad reputation on slashdot some years ago, btw.
One of these days I'm going to figure out how to make google go through tor.
And since other posters mentioned it, I don't like Apple either.
I don't think you understand. I don't care about the specific issue. I certainly don't care about the OS 9 memory model. I'm simply using it as an example to illustrate my point. If you want, I can rewrite the whole thing for mice with one button, or try to find somebody under the delusion of that the Linux audio system doesn't suck, and the configuration of dmix is really intuitive, and use that.
Pick any deficiency in any software or hardware. Watch a fanboy claiming first that it doesn't exist, then that 99% of the users aren't affected, then that some convoluted workaround is absolutely not a big deal. It's all the same.
Actually, yes I have. Not for long though, and for little more than web browsing and IRC.
That however doesn't have anything to do with the subject though, because I'm not commenting on the hardware, but on human behavior. It doesn't really matter whether it's Linux, Mac, Amiga or Windows. It also doesn't matter which specific feature is the issue.
One of the defining characteristics of a fanboy is being irrational, and your 99% claim simply doesn't stand up to logic. A fixed allocation is fine for a tool like a calculator. However there's no way it's going to work well for a tool that works on datasets of very variable size. Such as word processors, spreadsheets, photoshop and other office tools.
Memory is expensive, so most computers don't have that much of it. That means the default memory limit has to be low enough to work on the smallest machine it can be used on. It also has to leave room for whatever else will be running. So if you can expect Word to be started in a 32MB box, the limit must be a fraction of that. Send a large enough document, and the user will have to bump the value up.
I looked around for info on why this is necessary. My understanding is that OS 9 runs in an environment without a MMU, where each application needs a contiguous chunk of address space. That means it can't ask for more later, so it's got to get what it needs upfront. And there's no memory protection either. This is in a time where there existed OS/2, Windows 95, and Linux which had both things.
Now the issue isn't that OS 9 lacked this or that. All the alternatives also lacked one thing or another. The issue here is attitude. A realistic person would tell you that making the user to manually define how much memory an application is allowed to use is rather not ideal, but it's worth bearing with that quirk because the system itself is pretty nice. A fanboy though is going to first claim the issue doesn't exist at all, then that it only affects a tiny minority, then that it exists, but it's a completely trivial matter. (Note that still, most people have no clue what RAM is, what it's for, what's the difference between RAM and disk space, and how much they have of each). And after all that when the new version comes out without those issues they'll start evangelizing the new version with those improvements as the selling point, forgetting they claimed before it didn't bother them before in the slightest.
First, there's no "me" in there, that's a random post I googled up to have something to point at.
Second, thanks a lot for confirming precisely what I was saying.
Out of curiosity, what do you think is wrong with Schneier's work in security consulting?
I dunno about other people, but I don't like Microsoft's behavior, not MS per se. Say, MS SQL Server is actually a pretty decent piece of software. I don't think anybody would complain much if they ported that to Linux.
In fact, I think a Linux version of MS Office, with ODF support would be very welcome. But they'll never to that, because it wouldn't help keep and extend their monopoly. Take the whole OOXML mess happening right now. That's precisely the sort of thing that annoys people. MS can't just compete, they've got to rule with an iron fist, and if things aren't going their way, they'll try to squish whoever's necessary to get what they want.
If MS behaved in a more IBM-like manner I doubt people would protest nearly that much.
Since when? Granted, the scheduler is a complex matter. It's a complicated piece of software, and a very important part of the kernel. It affects everybody using it. It's hard to benchmark because you can bet there's somebody with an obscure configuration who finds changes produce a much worse result than the old version. And even worse, perception of performance is subjective. Then there's that there are big egos involved when working on something this important.
But for all of that, nobody's claiming the Linux scheduler is perfect. On the contrary, there's been much debate, benchmarking and testing, and the scheduler changed several times. Linux started as a single CPU operating system. Now it can use more CPUs than Windows. There clearly have been improvements.
In fact, I think people get a mistaken perception of this. It's not that Linux is the only place where there are huge arguments about how the scheduler should work. It's that at Apple and MS they probably also have huge arguments but out of public view, and people aren't used to seeing this sort of thing in the open.
Well, I had to use an example and that came to mind. I think it would have made a better argument a few years ago, as with the growth of available space compression became less important. I still have an old but perfectly functional 1 GB flash player. With that capacity, ogg's quality advantage vs mp3 is nice.
It still has advantages though. MP3 and AAC have licensing issues. FLAC is too big for many purposes. I've seen games (Unreal Tournament I think?) with an ogg soundtrack. If you need audio to be compact, and don't want to pay licensing fees, I can't think anything besides Vorbis to do it.
bitrate peeling seems to have died, btw, which is a pity. Sounded like a nice idea.
Eh? There's plenty proprietary software on Linux. Vmware and commercial games for instance. I've never, ever seen somebody claiming it to be wrong to use vmware or NWN on Linux. There's a preference towards free software of course when possible.
Well, most of my collection is in it. And why would I buy an iPod, then convert my collection with loss of quality instead of buying a player that works without extra messing with it in the first place?
Of course, Apple can improve things.
What I'm saying is that the fanboys' perception is that whatever is current is the absolute perfection. Take OS 9 for instance, where you had to set manually the amount of memory an application could use. If you brought up that setting memory is an oddly unfriendly characteristic for an OS that aims to be user friendly, you'd get a reply along the lines of "Lets see, I select the app, "get info"-->Memory and then set the amount. What's hard about that?"
Of course now that OS X is here, it's OS X what became the definition of perfection. I get the impression that many people refuse to acknowledge the existence of any faults until they're fixed, then the subject is quietly forgotten. For instance:
Everybody swore that a mouse with more than one button isn't needed, until Apple suddenly released a computer with one.
The memory limit was an "advantage", because Windows would die a swap death, and "Whgat New user is going to jump in and go manilulate large excel files?", anyway?
etc.
This attitude turns many people off, because: It creates a feeling that there's some sort of apple collective that many people aren't interested to join, because for them computers are a tool and not an object of religious worship. It creates the feeling of that it's hard to get a honest opinion about anything because many people are dedicated to sweeping faults under the rug. And it creates an impression of inflexibility: Either you accept the package in full, or you'd better not get it at all, because there's nothing in the middle.
IMO, the Linux zealots are less scary.
It's a difference in philosophy. Linux is about freedom and choice. If you say "Linux lacks X", most of the time if you get a negative reply it'll be something like "well, go fix it, the source's there". You generally won't be flamed to a crisp for daring to suggest that say, the state of audio in Linux isn't ideal. Constructive criticism could get a positive reply. Take the guy who did Linux boot benchmarking -- it quickly resulted in optimizations of the process.
Now try to do the same with Apple. Apple is about the "experience". Either you get it, or you can go look somewhere else. If you try to suggest the iPod, iPhone or something else isn't ideal you'll often get a reply from somebody who thinks nothing Apple makes might be a bit imperfect, and that if you don't like it, something is wrong with you. Mac OS was perfect before OS X came out. I've never seen a fan reply to the complaint of the iPod's lack of ability to play Ogg Vorbis as "You know, they should really include that". If it was a Linux device somebody would have added that within a month of the iPod's release.
How can you have a working democracy if the population isn't free? Let's see what Wikipedia says about that:
Note: I'm going with Wikipedia here intentionally because I expect it to be biased towards the American concepts in such matters.
So for example. Take away the freedoms of association, assembly, press, religion and speech. Quite totalitarian. What kind of democracy are you going to get in a state where citizens can't form associations or political parties, press is restricted to follow the government agenda, the religion is whatever the state says it is and nothing else, and if you say something unpopular you suddenly vanish?
Actually, no. The mistake you make is thinking somebody is interested in you personally. Most often they aren't.
They're not after your vacation's photos. They're after your internet connection, so that you're the one blamed for sending a huge amount of spam.
The way around this is by using the best security you can. If you use WPA, and the neighbour is open, guess who the spammer is going to choose for their operations. Same goes for house and car security: It doesn't need to be perfect, just good enough to be more inconvenient to break through than the other ones nearby.
You can't secure a connection against abuse. Exploits can be as simple as crafting an URL, and completely normal internet activity can involve illegal content. Posting death threats for instance.
If you refer to Applied Cryptography, it's actually a very interesting and mostly readable book. IMO it's worth reading for anybody working on anything crypto related. At least part of it.
About half of it is about the theory. This includes things like cryptographic protocols, such as: How to get two parties to exchange information in such a way that one of them can't back off it once they started? A lot of that is quite esoteric, but it's quite interesting, and mostly understandable to a programmer without needing math knowledge.
The other half is a description of many crypto algorithms, such as, how DES works. This is followed by source listings. That part is probably too hard to understand for people without the relevant knowledge. This is probably is interesting for a cryptographer and includes discussion on how the algorithms came to be, but probably won't be useful to 99% of programmers as most people aren't going to implement DES from scratch.
Then there's Secrets and Lies, which IMO is a bad book for a technical person. It makes a fine gift for somebody who really doesn't get how is all this security stuff supposed to work, but for me it's a "Well, DUH" one after another. It explains what is a DoS, what is the problem with biometrics, etc. Anybody interested in the subjects will be quite familiar with all of that already.
The problem is the vagueness of things like "interferes significantly". It also depends on the point of view of the one doing that evaluation, especially for children.
For example, many people still consider any kind of online communication to be "not real". Even if online you're capable of effectively leading a group of 50 people (which IMO is impressive since it's hard to force people online to do anything), that pales in comparison with the social value of getting drunk in a bar with people you barely know.
Nobody agrees on what "furry" is. I'd say it's like "computer geek". It points to a recognizable trait (this person likes computers), but inside that category you have very different people, from script kiddies to experts able to design a CPU, build a whole computer and make it work. It's wide enough to include people that are nothing alike to the point of not being able to understand each other, and people with violent disagreements. Take for example Linux, Windows and Mac fanatics. Each of those can be easily in the "computer geek" category, but it's very likely that they'll never agree on what an user friendly OS should be like.
Same goes for furries. The wide definition is something like "somebody who likes anthropomorphic animals". That includes many different people. Some are simply interested in the concept (or like to draw anthros). Some love a particular species with RL experience and identify with it. Some identify with a particular species despite having never been anywhere near it. Some identify with species that don't exist at all (mythological, or something completely new). Some take it seriously and for them it's a major component of their lives. A very few take it REALLY seriously and would go to extremes like body modification. Some like furry porn and some don't.
The more extreme behaviors aren't that common. AFAIK, very few own fursuits. They're expensive and far from everybody is interested.
In SL far from everybody is into it for the porn. I'd say your fiancee should visit Luskwood if she comes to SL, it's a PG furry place. No porn of any sort there.
Does that mean that an autistic savant, or just somebody who happened to have a calculator handy isn't human?
I have bc (a commandline calculator) running quite often. If you asked me that, I'd copy/paste, press enter, copy/paste the answer, then say "12381557655576425121, why do you ask?". By that standard you'd think I'm not human either.
Most modern cameras have a LCD screen. That means it's going to be impractical to force the user to always look into the viewfinder. Also most decent cameras can take photos on a timer. That means the eye data must be stored.
But this in turn makes it very easy to store anything as eye data. I could use my cat's eye for identification. Or find a high resolution photo on the web, print it, and try to get the camera to accept somebody else's eye.
Here you go: http://en.wikipedia.org/wiki/Efficient_XML_Interchange
I like this way much more than coming up with something new because it means I'd be able to keep my XML generating shell scripts, and just filter the output through a text to binary converter.
If you manage to crack that, try it at 127.249.17.156
Well, I use a Linux box for this. Get a cheap Pentium, attach a decent sized heatsink to it, use a CompactFlash to IDE adapter and a CF card, and you should be able to run that from a fanless power supply. With such low power usage you could probably even harmlessly disable the fan on a normal supply. The hardware can be found for free.
You can also get a Linksys WRT54G, those are below $100, and there's alternate firmware for it with plenty of functionality. Some googling suggests the WRT54G can be made work with IPv6 fine.