I have no idea how these OnLive guys are going to handle frequent hardware updates since high-end games continue to push hardware.
Not really. Most games are essentially using DirectX 9 with additions because that's what the current generation of consoles supports. When the next round of consoles comes out they'll support DirectX 12 and you'll see nothing but DX12 games for the next five years. The boundary-pushing games are a thing of the past simply because that only works if you only release for the PC. Everyone else is tethered to what you can do with the chips found in consoles.
The problem is that this will make gaming much more expensive. We're looking at a low-latency (10 ms or below), high-bandwidth (20 MB/s guaranteed minimum) line per user in order to be able to play fast-paced games like FPSes at 1080p. Given how things are today that would mean an SLA with your ISP. The ISPs aren't going to just bump the speed of their entire network because the content industry says so.
Of course we could also see people playing at lower resolutions and/or the death of the FPS genre. Both are unlikely.
The problem is that with over-prescription you'd need to somehow make people finish their course, even though many don't want to. How yould you go about the problem? One way would be to limit antibiotic administration to trained professionals but that would mean hospitalizing everyone with a fever. I can't think of anything else that can be expected to actually acheive something right now.
Actually, no. It requires the "Plain" build (random read/write access), which is the most basic UDF build that all UDF implementations should support. Packet writing is a means of making it transparent to the user whether he's using a write-once or a rewritable medium, using either the "VAT" or the "Spared" build.
Annoyingly, proper support of plain UDF is not universal and operating systems may or may not properly work with plain UDF filesystems. For example, even though OS X is supposed to properly support plain UDF I still managed to confuse the filesystem layer by using a UDF-formatted USB stick (done on both 10.5 and 10.6).
The large number of UDF revisions and the lack of support some OSes have for them is a large problem. If all major OSes fully supported UDF and treated it as a first-class citizen, we'd have a shoe-in for FAT32's successor.
Why fire insurance? Without a court system the insurance would never pay unless you forced them at gunpoint - and they most likely can afford more guns than you do. Then again, you would be paying them fore "fire insurance" because they have more guns than you do.
UDF? Granted, writing to it is finicky under OS X and I'm not sure how well Windows works but it has been a full FAT replacement for ages. Except people don't seem to bother even perceiving it.
Note that Cameras often create files with names like "DCIM0043.JPG", which fits into the old 8.3 format. This might be in order to avoid having to license the FAT LFN patent.
does the fact that there are no devices in the market mean that you would rather buy something as much unlike a smartphone as possible? if there is no true smartphone, what do you logically buy? something that is closer to the definition (symbian/maemo/droid) or something that is further behind (iphone)?
The fact that my definition of a device class causes said class to contain no devices when everyone else is of the opinion that there are a lot of them would first cause me to re-evaluate whether my definition is correct. Then I'd get the device that best suits my needs without insisting that less interesting (to me) devices don't belong in the device class at all. After all, most people agree that computers running Windows are PCs, even though they don't quite meet my preferences with regards to personal computers.
were you using single program oses on your 600mhz pc with 256mb ram and dedicated graphics? concerning the criticality of msdos apps in some places, i would like to inform you that in many places 128kbps internet is a critical workflow element. does that mean its okay to go back to that?
You're conflating "general-purpose" and "things I like". (And by the way, yes, I did occasionally boot my K6-2 into MS-DOS.)
Legacy applications often aren't nice and it'd be nice to move on (even if that would mean lots of retraining) but that doesn't mean they magically don't exist. They're still around and a general-purpose computer can reasonably be expected to run them, directly or through an emulation layer.
that is what is supposed to change with the advent of smartphones. but look at you.
Well, it's unlikely to change. Certain things just aren't possible - for example it's unfeasible to build a mobile phone with a current desktop-class CPU, a high-end gaming GPU or a full-width 105-key standard keyboard. Mobile devices must be reasonably small (ruling out large and/or heavy components) and they must last for a reasonable time on a finite battery charge (ruling out components with high power consumption).
You're always looking at restrictions simply because certain combinations of features don't make sense (like a mobile phone with an integrated DVD burner). I'm not saying that smartphones can't fulfill many of the same duties a desktop computer does. But you're not going to see people use them in as many diverse roles as a real general-purpose computer because they're not supposed to be one.
usually, cellphone chargers provide that kind of power. and charging is a lot faster than discharging. also note that pandora has a beefier battery than any cellphone.
Note that the Pandora is designed to different requirements than a smartphone, one of which is being a feasible replacement for a notebook - including being able to use arbitrary bus-powered USB devices. Smartphones can't do that - I can't plug one smartphone into another to charge it (and I think even syncing your contacts is unlikely to work). It's a sensible limitation because, as you pointed out, the Pandora's battery is much larger - large enough to push a smartphone past what consumers are likely to tolerate.
Being fully general-purpose would make smartphones less appealing - for instance, they'd need annoyingly large batteries just to be able to supply bus-powered USB devices. Smartphones aren't supposed to do everything, they're supposed to do what people need in their pocket. Which usually boils down to calls, PIM and internet access.
Well, we can't when you keep changing (or not mentioning) the parameters. As I already predicted, you seem to be complaining about the distinct lack of true Scotsmen here.
Note that you contradicted your earlier post by admitting that Symbian-based mobile phones are not general-purpose computers and therefore not smartphones by your own definition, merely devices resembling smartphones more than iPhones do.
No amount of insisting that everyone else is incapable of grasping the concept of a smartphone is helping the fact that your own definition is either inconsistent or implies that there are no such devices on the market.
By the way, I'd like to point out that we used non-multitasking general-purpose computers for quite some time (and there are a number of places where MS DOS applications are still considered a critical workflow element). You can argue that it's common for modern general-purpose computers to have multitasking support but I'd say that's just an OS feature orthogonal to the genereral-purposeness of the device as on general-purpose computers the OS is usually replacable. DOS-era computers are often still able to run a slimmed-down (but still multitasking) Linux; the hardware doesn't depend on a specific operating system to work.
Mobile phones are embedded devices with a limited set of capabilities. This applies to mere phones as well as smartphones. The limited set of capabilities of smartphones can extend to much of what you'd expect from a desktop computer but the platform still wasn't designed to let you run arbitrary machine code on the bare metal. You can't run anything you want, you only can run whatever the non-replacable operating system lets you - thus it's not completely general-purpose in my book.
By the way, as for mobile devices not being able to give enough power: If your mobile device's battery is not capable of supplying 500 mA at 5 V you have a very interesting device indeed. The Pandora shows that it's quite possible to do that. It may cause significant battery drain but there are no technical reasons why it shouldn't work.
(Actually, there might be one smartphone that could be general-purpose - I think Openmoko phones are completely reflashable and have sufficiently open specs to enable people to write compatible replacement OSes. Might be wrong, though.)
In that case I have yet to see one, although the Pandora (note: no actual telephone capabilities built in) might qualify. I expect a general-purpose computer to have features like being a host for the leading external bus (currently USB, RS-232 before that), and being able to run arbitrary compatible OSes (ideally also the ability to drive external monitors although there may be some overlap between general-purpose and embedded devices where external output is omitted). Anything short of that is not general-purpose (note that our definitions here are likely to diverge just as our definitions of smartphones do).
I was able to find Symbian phones that do USB OTG but none that have regular USB host capabilities - and I don't see why a general-purpose computer would use the OTG spec. I also couldn't find any information on how to install alternative bootloaders and/or OSes. This suggests to me that the platform was not designed to be a full replacement for desktop or portable computers. It does, however, act very well as a programmable PDA with a built in telephone.
If we do accept Symbian as being general-purpose, what makes the iPhone less so? Both are user-programmable with the full power of a Turing machine, both offer user input. Yes, the iPhone doesn't have USB host capabilities of any kind but that alone isn't a good definition of a general-purpose computer.
The was I see it, a smartphone is a mobile phone with included PDA functionality (or alternatively a PDA capable of making phone calls). Of course I'm not a true Scotsman but I'd say that the iPhone certainly fulfills these requirements - even if I readily agree that it's not a very good PDA.
Since the GGP was talking about downloading videos (and using Wireshark to make it possible) this subthread is obviously not about any device running iPhone OS. We're talking about capturing Flash video on a PC, which means that Firefox is very much available.
In all these cases, and they cover a large area, the only real challenge is working past the weak obfuscation, if any, to find the actual source URL for the video, which is already in a workable format, and just grabbing it. For individual video capturing use, "View source" and a couple minutes of thinking is usually sufficient, with Wireshark and a filter looking for HTTP GET requests waiting in the wings if that doesn't work.
Wireshark? More like Firefox with one of several user-friendly addons. Flash needs to talk to the browser to fetch the video. At that point it's fairly easy to capture the request (and most decent addons don't even need to do that often because they already know how to deobfuscate many sites' content links).
I actually find it faster to type something like "the produkt will make you happy" or "he is living comfort eagle" than something like "q!h9%2tL".
Random lines from decent-but-not-favourite songs or references to old demos work really well. If your keyboard layout allows easy access to appropriate characters, random foreign songs work well, too. Few people are going to guess "ta några steg åt vänster" or even know which meme it's from.
Why not just use passphrases? Have a minimum password length of 20 characters (encouragng users to use multiple words or whole sentences), that makes (counting only lower-case letters and spaces) 27^20 possible passwords. A dictionary attack would be faster but would still need to run through all permutations of words you can fit into 20+ characters.
Policies like "16 printable ASCII characters with imposed limits" result in 95^16 possibilities - quite a bit more than as 27^20 but if we figure in benign passphrase restrictions like "at least two capital letters" we get 53^20, which is a nice big search space for an uninformed attack. And people might stop posting their passwords where everyone can read them.
Your approach actually shifts the authentication mode: Users aren't identified by what they know (a password) but rather by what they have (a rather cumbersome security token in form of a sheet of paper or an encrypted file).
And now imagine that we finally settle on a common codec (and I'd rather see something Mozilla can afford to implement so VP8 is a godsend). You don't need a codec pack or Flash because the codec is already built into your browser. And when the next version of your favourite OS comes out it'll probably also ship with the codec. No hassle at all, it just works out of the box.
(Also finally smooth web video for OS X users. Flash is not only insecure but also dog slow on Mac OS.)
Do you think the people who argue that drawn child porn turns people into molesters acknowledge the difference between "pedophile" and "child molester"? While their theory might indeed be based on something much more reasonable like what you wrote, there are a lot people who literally think that lolicon can somehow override someone's sexuality and turn them into a child rapist, whether they might have been predisposed or not. That's what happens when reasonable arguments are filtered through the lens of the common tabloid.
Come to think of it, there are some parallels to the old "gay recruitment" theory...
Apparently you assume that everyone runs Windows on x86. If we have multiple competing codecs we will most likely end up locking out people because a particular codec is not available everywhere. Let's say Microsoft comes out with another version of Windows Media. Will their codec be supported on Linux/ARM or OS X/PPC?
I remember the days when video on the web could be anything between WMV, Sorensen, 3ivx, DivX, XviD, RealVideo, MPEG-1, MPEG-2... and those are just the common ones you had to have. It was not nice, especially if you used a platform that didn't have a proper codec for one or more of them.
Also note that many small devices use hardware acceleration to allow playback of video files their CPUs can't handle themselves. With one codec that works. With many codecs we lock out all mobile devices or force users to limit themselves to a subset of media sites that support the same codecs their devices support.
Codecs are an area where standardization is useful. It enables everyone to have access to multimedia - as opposed to everyone with a powerful PC and a supported OS.
You're right about Bungie (remember, they used to develop for Mac OS before Microsoft bought them; now they dont make any Mac games at all) but Valve isn't exactly dogmatically Windows-only. Steam for Mac (along with Cider-ported versions of Valve's Steam games) is slated to launch this month. They realized that Mac users pay for games, too, and are catering towards that market. Seems fairly balanced.
The site is apparently that of a manufacturer of fairly realistic lifesize dolls of children. Nothing inherently bad although some of the pictures show the dolls unclothed.
by allowing lolicon and similar images to become acceptable I feel as if an important social barrier may have been breeched.
And that may be a good thing. Right now pedophilia is anathema. Anything related to the topic is considered so vile that even thinking about the topic itself is dirty. You don't get sensible approaches with that kind of atmosphere.
What would happen if some politician came up with the idea of pedophile support networks and state-funded counseling? An instant career end, that's what would happen. Because pedophilia and everything remotely associated with it is EVIL EVIL EVIL any nuanced approach to the topic is seen as insufficiently anti-pedophile which is the same as pro-pedophile, thus you're unfit for any office anywhere if you don't demonize pedophiles enough.
Progress can't be made if zealotry is the norm.
Now what if drawn child pornography does get some mindshare? People might start to wonder whether victimless* CP is really just as bad as regular CP**. And then they wonder whether the current laws are actually to the benefit of the people. And when we have people wondering whether a picture of Bart Simpson naked really warrants a prison term and an entry in the sexual offender database we're a big step closer to them actually bothering to question other aspects of current policy.
* Yes, we can't prove that a certain artist doesn't use actual children/CP as source material. Yet. We regularly catch companies doing bad things, though, so spot checks etc. can help. Requiring drawn CP creators to submit to random inspections (otherwise not getting admitted into the market) should keep the incidence of abusive drawn CP hitting the market low. And I do expect that the black market would take a large hit if a perfectly legal white market existed.
** The usual argument is that "all child porn turns people into pedophiles". So far I haven't seen any solid scientific support for that hypothesis. There are other hypotheses like "drawn CP allows pedophiles to let off steam without a child getting involved" which are equally unsubstantiated, seem equally sensible to a layman and make it seem a good idea to actually try and find out what is true.
Right now we can't even tell whether drawn CP is good or bad beyond resorting to truthiness.
Slashdot Mirror
I have no idea how these OnLive guys are going to handle frequent hardware updates since high-end games continue to push hardware.
Not really. Most games are essentially using DirectX 9 with additions because that's what the current generation of consoles supports. When the next round of consoles comes out they'll support DirectX 12 and you'll see nothing but DX12 games for the next five years. The boundary-pushing games are a thing of the past simply because that only works if you only release for the PC. Everyone else is tethered to what you can do with the chips found in consoles.
The problem is that this will make gaming much more expensive. We're looking at a low-latency (10 ms or below), high-bandwidth (20 MB/s guaranteed minimum) line per user in order to be able to play fast-paced games like FPSes at 1080p. Given how things are today that would mean an SLA with your ISP. The ISPs aren't going to just bump the speed of their entire network because the content industry says so.
Of course we could also see people playing at lower resolutions and/or the death of the FPS genre. Both are unlikely.
The problem is that with over-prescription you'd need to somehow make people finish their course, even though many don't want to. How yould you go about the problem? One way would be to limit antibiotic administration to trained professionals but that would mean hospitalizing everyone with a fever. I can't think of anything else that can be expected to actually acheive something right now.
Actually, no. It requires the "Plain" build (random read/write access), which is the most basic UDF build that all UDF implementations should support. Packet writing is a means of making it transparent to the user whether he's using a write-once or a rewritable medium, using either the "VAT" or the "Spared" build.
Annoyingly, proper support of plain UDF is not universal and operating systems may or may not properly work with plain UDF filesystems. For example, even though OS X is supposed to properly support plain UDF I still managed to confuse the filesystem layer by using a UDF-formatted USB stick (done on both 10.5 and 10.6).
The large number of UDF revisions and the lack of support some OSes have for them is a large problem. If all major OSes fully supported UDF and treated it as a first-class citizen, we'd have a shoe-in for FAT32's successor.
Of course one doesn't hurt the other.
Why fire insurance? Without a court system the insurance would never pay unless you forced them at gunpoint - and they most likely can afford more guns than you do. Then again, you would be paying them fore "fire insurance" because they have more guns than you do.
UDF? Granted, writing to it is finicky under OS X and I'm not sure how well Windows works but it has been a full FAT replacement for ages. Except people don't seem to bother even perceiving it.
Note that Cameras often create files with names like "DCIM0043.JPG", which fits into the old 8.3 format. This might be in order to avoid having to license the FAT LFN patent.
Of course Google know the public IP of your router. What do you think they see when you talk to their server?
Or at least, that's how it's been for decades in Europe.
I hope I never get to visit the Europe you heard of. It's not anything like that in the one I live in.
does the fact that there are no devices in the market mean that you would rather buy something as much unlike a smartphone as possible? if there is no true smartphone, what do you logically buy? something that is closer to the definition (symbian/maemo/droid) or something that is further behind (iphone)?
The fact that my definition of a device class causes said class to contain no devices when everyone else is of the opinion that there are a lot of them would first cause me to re-evaluate whether my definition is correct. Then I'd get the device that best suits my needs without insisting that less interesting (to me) devices don't belong in the device class at all. After all, most people agree that computers running Windows are PCs, even though they don't quite meet my preferences with regards to personal computers.
were you using single program oses on your 600mhz pc with 256mb ram and dedicated graphics? concerning the criticality of msdos apps in some places, i would like to inform you that in many places 128kbps internet is a critical workflow element. does that mean its okay to go back to that?
You're conflating "general-purpose" and "things I like". (And by the way, yes, I did occasionally boot my K6-2 into MS-DOS.)
Legacy applications often aren't nice and it'd be nice to move on (even if that would mean lots of retraining) but that doesn't mean they magically don't exist. They're still around and a general-purpose computer can reasonably be expected to run them, directly or through an emulation layer.
that is what is supposed to change with the advent of smartphones. but look at you.
Well, it's unlikely to change. Certain things just aren't possible - for example it's unfeasible to build a mobile phone with a current desktop-class CPU, a high-end gaming GPU or a full-width 105-key standard keyboard. Mobile devices must be reasonably small (ruling out large and/or heavy components) and they must last for a reasonable time on a finite battery charge (ruling out components with high power consumption).
You're always looking at restrictions simply because certain combinations of features don't make sense (like a mobile phone with an integrated DVD burner). I'm not saying that smartphones can't fulfill many of the same duties a desktop computer does. But you're not going to see people use them in as many diverse roles as a real general-purpose computer because they're not supposed to be one.
usually, cellphone chargers provide that kind of power. and charging is a lot faster than discharging. also note that pandora has a beefier battery than any cellphone.
Note that the Pandora is designed to different requirements than a smartphone, one of which is being a feasible replacement for a notebook - including being able to use arbitrary bus-powered USB devices. Smartphones can't do that - I can't plug one smartphone into another to charge it (and I think even syncing your contacts is unlikely to work). It's a sensible limitation because, as you pointed out, the Pandora's battery is much larger - large enough to push a smartphone past what consumers are likely to tolerate.
Being fully general-purpose would make smartphones less appealing - for instance, they'd need annoyingly large batteries just to be able to supply bus-powered USB devices. Smartphones aren't supposed to do everything, they're supposed to do what people need in their pocket. Which usually boils down to calls, PIM and internet access.
Well, we can't when you keep changing (or not mentioning) the parameters. As I already predicted, you seem to be complaining about the distinct lack of true Scotsmen here.
Note that you contradicted your earlier post by admitting that Symbian-based mobile phones are not general-purpose computers and therefore not smartphones by your own definition, merely devices resembling smartphones more than iPhones do.
No amount of insisting that everyone else is incapable of grasping the concept of a smartphone is helping the fact that your own definition is either inconsistent or implies that there are no such devices on the market.
By the way, I'd like to point out that we used non-multitasking general-purpose computers for quite some time (and there are a number of places where MS DOS applications are still considered a critical workflow element). You can argue that it's common for modern general-purpose computers to have multitasking support but I'd say that's just an OS feature orthogonal to the genereral-purposeness of the device as on general-purpose computers the OS is usually replacable. DOS-era computers are often still able to run a slimmed-down (but still multitasking) Linux; the hardware doesn't depend on a specific operating system to work.
Mobile phones are embedded devices with a limited set of capabilities. This applies to mere phones as well as smartphones. The limited set of capabilities of smartphones can extend to much of what you'd expect from a desktop computer but the platform still wasn't designed to let you run arbitrary machine code on the bare metal. You can't run anything you want, you only can run whatever the non-replacable operating system lets you - thus it's not completely general-purpose in my book.
By the way, as for mobile devices not being able to give enough power: If your mobile device's battery is not capable of supplying 500 mA at 5 V you have a very interesting device indeed. The Pandora shows that it's quite possible to do that. It may cause significant battery drain but there are no technical reasons why it shouldn't work.
(Actually, there might be one smartphone that could be general-purpose - I think Openmoko phones are completely reflashable and have sufficiently open specs to enable people to write compatible replacement OSes. Might be wrong, though.)
In that case I have yet to see one, although the Pandora (note: no actual telephone capabilities built in) might qualify. I expect a general-purpose computer to have features like being a host for the leading external bus (currently USB, RS-232 before that), and being able to run arbitrary compatible OSes (ideally also the ability to drive external monitors although there may be some overlap between general-purpose and embedded devices where external output is omitted). Anything short of that is not general-purpose (note that our definitions here are likely to diverge just as our definitions of smartphones do).
I was able to find Symbian phones that do USB OTG but none that have regular USB host capabilities - and I don't see why a general-purpose computer would use the OTG spec. I also couldn't find any information on how to install alternative bootloaders and/or OSes. This suggests to me that the platform was not designed to be a full replacement for desktop or portable computers. It does, however, act very well as a programmable PDA with a built in telephone.
If we do accept Symbian as being general-purpose, what makes the iPhone less so? Both are user-programmable with the full power of a Turing machine, both offer user input. Yes, the iPhone doesn't have USB host capabilities of any kind but that alone isn't a good definition of a general-purpose computer.
The was I see it, a smartphone is a mobile phone with included PDA functionality (or alternatively a PDA capable of making phone calls). Of course I'm not a true Scotsman but I'd say that the iPhone certainly fulfills these requirements - even if I readily agree that it's not a very good PDA.
Since the GGP was talking about downloading videos (and using Wireshark to make it possible) this subthread is obviously not about any device running iPhone OS. We're talking about capturing Flash video on a PC, which means that Firefox is very much available.
In all these cases, and they cover a large area, the only real challenge is working past the weak obfuscation, if any, to find the actual source URL for the video, which is already in a workable format, and just grabbing it. For individual video capturing use, "View source" and a couple minutes of thinking is usually sufficient, with Wireshark and a filter looking for HTTP GET requests waiting in the wings if that doesn't work.
Wireshark? More like Firefox with one of several user-friendly addons. Flash needs to talk to the browser to fetch the video. At that point it's fairly easy to capture the request (and most decent addons don't even need to do that often because they already know how to deobfuscate many sites' content links).
I actually find it faster to type something like "the produkt will make you happy" or "he is living comfort eagle" than something like "q!h9%2tL".
Random lines from decent-but-not-favourite songs or references to old demos work really well. If your keyboard layout allows easy access to appropriate characters, random foreign songs work well, too. Few people are going to guess "ta några steg åt vänster" or even know which meme it's from.
Why not just use passphrases? Have a minimum password length of 20 characters (encouragng users to use multiple words or whole sentences), that makes (counting only lower-case letters and spaces) 27^20 possible passwords. A dictionary attack would be faster but would still need to run through all permutations of words you can fit into 20+ characters.
Policies like "16 printable ASCII characters with imposed limits" result in 95^16 possibilities - quite a bit more than as 27^20 but if we figure in benign passphrase restrictions like "at least two capital letters" we get 53^20, which is a nice big search space for an uninformed attack. And people might stop posting their passwords where everyone can read them.
Your approach actually shifts the authentication mode: Users aren't identified by what they know (a password) but rather by what they have (a rather cumbersome security token in form of a sheet of paper or an encrypted file).
And now imagine that we finally settle on a common codec (and I'd rather see something Mozilla can afford to implement so VP8 is a godsend). You don't need a codec pack or Flash because the codec is already built into your browser. And when the next version of your favourite OS comes out it'll probably also ship with the codec. No hassle at all, it just works out of the box.
(Also finally smooth web video for OS X users. Flash is not only insecure but also dog slow on Mac OS.)
Which not only strengthens my argument but also makes my nipples explode with delight.
Do you think the people who argue that drawn child porn turns people into molesters acknowledge the difference between "pedophile" and "child molester"? While their theory might indeed be based on something much more reasonable like what you wrote, there are a lot people who literally think that lolicon can somehow override someone's sexuality and turn them into a child rapist, whether they might have been predisposed or not. That's what happens when reasonable arguments are filtered through the lens of the common tabloid.
Come to think of it, there are some parallels to the old "gay recruitment" theory...
Apparently you assume that everyone runs Windows on x86. If we have multiple competing codecs we will most likely end up locking out people because a particular codec is not available everywhere. Let's say Microsoft comes out with another version of Windows Media. Will their codec be supported on Linux/ARM or OS X/PPC?
I remember the days when video on the web could be anything between WMV, Sorensen, 3ivx, DivX, XviD, RealVideo, MPEG-1, MPEG-2... and those are just the common ones you had to have. It was not nice, especially if you used a platform that didn't have a proper codec for one or more of them.
Also note that many small devices use hardware acceleration to allow playback of video files their CPUs can't handle themselves. With one codec that works. With many codecs we lock out all mobile devices or force users to limit themselves to a subset of media sites that support the same codecs their devices support.
Codecs are an area where standardization is useful. It enables everyone to have access to multimedia - as opposed to everyone with a powerful PC and a supported OS.
You're right about Bungie (remember, they used to develop for Mac OS before Microsoft bought them; now they dont make any Mac games at all) but Valve isn't exactly dogmatically Windows-only. Steam for Mac (along with Cider-ported versions of Valve's Steam games) is slated to launch this month. They realized that Mac users pay for games, too, and are catering towards that market. Seems fairly balanced.
The site is apparently that of a manufacturer of fairly realistic lifesize dolls of children. Nothing inherently bad although some of the pictures show the dolls unclothed.
And that may be a good thing. Right now pedophilia is anathema. Anything related to the topic is considered so vile that even thinking about the topic itself is dirty. You don't get sensible approaches with that kind of atmosphere.
What would happen if some politician came up with the idea of pedophile support networks and state-funded counseling? An instant career end, that's what would happen. Because pedophilia and everything remotely associated with it is EVIL EVIL EVIL any nuanced approach to the topic is seen as insufficiently anti-pedophile which is the same as pro-pedophile, thus you're unfit for any office anywhere if you don't demonize pedophiles enough.
Progress can't be made if zealotry is the norm.
Now what if drawn child pornography does get some mindshare? People might start to wonder whether victimless* CP is really just as bad as regular CP**. And then they wonder whether the current laws are actually to the benefit of the people. And when we have people wondering whether a picture of Bart Simpson naked really warrants a prison term and an entry in the sexual offender database we're a big step closer to them actually bothering to question other aspects of current policy.
* Yes, we can't prove that a certain artist doesn't use actual children/CP as source material. Yet. We regularly catch companies doing bad things, though, so spot checks etc. can help. Requiring drawn CP creators to submit to random inspections (otherwise not getting admitted into the market) should keep the incidence of abusive drawn CP hitting the market low. And I do expect that the black market would take a large hit if a perfectly legal white market existed.
** The usual argument is that "all child porn turns people into pedophiles". So far I haven't seen any solid scientific support for that hypothesis. There are other hypotheses like "drawn CP allows pedophiles to let off steam without a child getting involved" which are equally unsubstantiated, seem equally sensible to a layman and make it seem a good idea to actually try and find out what is true.
Right now we can't even tell whether drawn CP is good or bad beyond resorting to truthiness.