As with anything, it's horses for courses to some extent. I switched from Nokia to Sony many years ago (back when phone interfaces were all simple and often elegant). I eventually found the Sony interface was getting too awkward, and went back to Nokia hoping for something as intuitive as I used to have.
I totally agree that S60 is a bad tradeoff between complexity and flexibility as compared with Nokia's older phones, but I prefer it to Sony's for the tasks I do most.
Here's hoping the iPhone really does combine offer a better combination of the two. I'm also in the UK, and echo your comment about Orange. I'd love it if they'd pair with Virgin and actually have a chance of delivering okay customer service (IMO). Jobs and Branson both appear to be value-driven entrepeneurs - Jobs always keen on elegance, Branson on great customer service, niether of them necessarily doing things other people can't, but always doing it differently. They don't always hit the target (G4 Cube, Virgin TV/broadband horror stories I've come across, although that may be due to underlying infrastructure issues), but at least they're driving improvement in unusual directions.
I fully agree that it's sad, but while it isn't fully rational, it might be at least partially so.
No human is fully rational, so it sometimes pays to be rational about your irrationality. If you just can't help wanting something that badly, it might be more rational to indulge yourself than to get all depressed not buying it. Of course, it means you're lacking in willpower, but hey, maybe you've learned to accept that, too!
Once you start talking about being completely rational, you should really get down to brass tacks and justify your continued existence before making any other decision. Since I've not seen a justification for our existence that more than a minority agree with (unless you allow 'god wishes it' without specifying which god - surely cheating), I don't think it's a great way to operate.
I'd do the copy-and-paste (and a few other context-related features) with the same 'right-click' mechanism as recent Apple laptops - a two-fingered click brings up a little contextual menu.
Anyone from Apple reading this? Isn't the API for this already built into Cocoa? How hard can it be?
I'd imagine a web-based iPhone app could probably work as a pretty good IM client.
I'm hugely disappointed there's no proper API - I think OSX's API is one of its most compelling aspects - but the web-based system should be capable of filling some gaps.
This is true to a certain extent, and more so with TVs than computers. However, when a computer is Bob, you may well use the computer to do the interception/cracking or whatever. At this point, we're back to Bob being receiver and cracker, or to Eve being able to control Bob, depending how you look at it.
Either way, it's not quite as clear-cut in all situations.
They tried roughly this already. It didn't work. Why? Because entrants to the new 'Works with Mac OS' market didn't build commodity boxes with super-slim margins, they built high-margin, high-performance machines that didn't necessarily kill Apple's sales, but wiped out the bulk of its profits. This is how Apple ended up at $12 a share, just before Jobs came back.
This isn't about 'vision' or 'consumer electronics vs. computers', it's about economics. If your profit comes from hardware, you have to sell profitable hardware. Apple can't compete economically in the OS market because there's a monopoly. They can compete in the hardware market. The only thing that has any chance of competing with MS on OS alone is 'guerilla' OSes like Linux. Conventional business, like conventional warfare against an army 100x the size of yours, just won't work.
Ironically, the best way to destroy MS might be to separate Mac OS and Mac hardware into two companies - it would immediately become clear the extent to which Microsoft is a monopoly, which only has a competitor because that competitor isn't really in the same market!
I think it's a shame you've been modded down, AC or not, because you're precisely correct. Piracy as used when the import duty on tea to Britain was something like 100% (or more, I can't recall) did not protest paying for the product at all, it protested the interlocutor.
We now have technology sufficient to the task of actually distributing the cost of mkaing the movie fairly directly to the consumer. No-one's figured out a way to make it happen, but the market appears to be 'aware' on some level that it should. A key element of this is that a film/song should actually get cheaper as more people buy it. Perhaps the reasons the model hasn't been tried is it could potentially involve refunding earlier customers the difference.
This may sound far-fetched, but think about it: cost of the system running this will quickly be minimal compared to capital outlay on films. The only (admittedly large) problem is getting that capital together in advance. Perhaps a subscription fee, which is potentially refunded depending on usage but only after a year? So if 100 million people worldwide watch a $250mn film (add $50mn profit and infrastructure costs, say), they pay $3 for the privelege, and that is all they pay ever, for every format or device they ever watch it on, because format-shifting has negligible cost.
I know this is a pipe-dream (although a large film co. could shift to this model gradually and remain profitable, I think), but to anyone who asks when would be a good time to stop pirating: when the system is run like this, that's when.
There are some very good points here, and it puts me in mind of how bandwidth is paid for in general. Someone running a website is charged for bandwidth usage by their provider, yet they are providing the content that means people want to use the service at all. They also charge the end user, but generally not per-quantity.
Would anyone else agree that a better model might be end-users paying a rate for data downloaded which really represented the cost, and content providers not paying? It seems fairer to me - you pay for the stuff you want, and that's how the network is supported. I know a lot of the big publishers are 'evil corporations', and one's instinct might be that they should pay for distributing stuff, but ultimately they'll pass the costs on one way or another anyway, and internet distribution (bar advertising) is unique in that it's genuinely on-demand use of the infrastructure. Actually, I think I've just found the flaw in my theory - bandwidth-intensive flash adverts that drive up costs, although that could be solved at the browser end, and probably would if people were really paying for data.
They could apply different costs to different methods of getting data, so when your phone moves out of WiFi, you start using 3G or EDGE and the price per byte goes up. The market as a whole would be much free, I think, than under the current model.
Not sure if this relates, but a study of the children and grandchildren of women subjected to extreme malnourishment during WWII showed some interesting results.
Firstly, whether the mother or the foetus lost weight depended on how far along the foetus was - I can't remember exactly, but I believe if the foetus was almost capable of surviving on its own, the mother's body basically sacrificed itself.
Perhaps more weirdly, while the babies of malnourished mothers were not more likely to be smaller as adults, their children were. So there was some kind of knock-on effect on the grandchildren of the original women. I think they ascribed this to biochemical changes in the middle generation rather than any genetic effect, but they couldn't pinpoint it. Perhaps some very subtle DNA alteration kicked in, and the children of malnourished mothers produced smaller offspring (more likely to survive in a low-food environment) 'just in case'.
This is a really good point, and very well put. Allow me to adjust what I said slightly. While the 'service' I would like to buy would include the ability to obtain appropriate copies for all my playing devices, in perpetuity, I would not feel hard-done-by if I wasn't allowed to pass copies to a friend for free. CDs can remain CDs, you're quite right; I think it's the 'service' model which has yet to be properly worked out.
Of course, as TFA says, it may not be possible for them to stop me, but I'm just saying my problem with pay-for is not the lack of notional permission to reproduce it, but the actual utility of the format. As someone said earlier, the current situation is one of legislation, litigation, and cartels maintaining market values massively distant from actual marginal value.
Perhaps the ideal system would not DRM files per se, but simply tag them as mine (as Apple's higher quality music does). The key change from the current system would be that distributing the music should not result in some ridiculous damages case, but simply the revocation of that item for that user - you break the contract, you lose the service, as with your warranty example.
Hmm, thinking about it, I'm not so sure it would work, but it'll be interesting to see how Apple's new system pans out.
Your last point is the key for me. If I'm going to pay, it must be more functional than the free version. I want to pay a single fee for each song/show/film, which gives me the ability to obtain it in a variety of formats including new ones as they arrive. Technology moves so fast that even if you're allowed to buy the 'best' current version it will be defunct in shorter and shorter spaces of time - either because of a higher-quality (eg DVD to 720 to 1080) or more flexible (eg mp3, divx) version.
Damn straight, and this brings in the most important restriction on a child's use of the internet, the telly, and computer games - time.
I would have no problem with kids watching/browsing/playing, were it not for all the advertising gradually warping their brains, and the largely one-way interaction atrophying their social skills. Couple of hours a day, max.
Uhh... except of course I hated my parents not letting me play Street Fighter II every waking moment outside school hours. Mind you, four-player Wii I would not object to.
Conversely, some things are intuitive at first, but fall down later because they are not logical, so you can't apply things you've learned one place elsewhere.
The best interfaces are logical, and expose the logic as intuitively and quickly as possible. I guess the problem is working out what someone will be able to interpret quickly. Case in point: icon-only control systems may take up less space, but unless the pictures actually represent something the user already knows about, they are less easy to intuit than written names.
Sorry, I should have said: I quite appreciate Linux is much less DIY than it used to be, if you go for Ubuntu, and all the PCs I set up and ran in our company used it (three, alongside a Mac). I meant a genuine compliment: there is more scope for DIY on linux than Mac, if you want it.
Further to that, I guess when I say polish I mean something more than system-wide gloss. I would gladly exchange aqua for something faster, if it maintained all the functionality (which includes being unobtrusive - I personally think the pale grey is less distracting than many schemes). What I'm prepared to pay for is the level of ease-of-use of certain specific programmes that really are best-in-class. Video editing is the key one for me (plus I've been using Photoshop since 2.0, and I've grown too close to it to give it up), but I'm sure other people have other things. I also think mac programmes in general are more consistent with one another than on other platforms, and this cuts down learning time.
Quite agree with your price point - it would be a huge plus if you could build a mac from scratch, and all my HD systems are hand-made to keep the cost down - but it's worth thinking about the true cost: if I can get even 5% more done each day because I've got a superb programme running on a very predictable OS/hardware combo that I barely have to maintain, I'll easily make back the cost of the computer, and in my experience I get quite a bit more than a 5% bonus. One-off unusual tasks with videos and images are quicker on the mac, and I can still write bash scripts for anything that needs automating.
Plus, that's more subjective than he allows for. Despite being a born-and-bred Mac user, when I found my version of Fontographer wouldn't play with OS X (this was back in the day), I didn't have any problems at all firing up X11 and installing a free UNIX tool I'd found on the 'net. I was fairly new to the command-line, and totally new to X, but had no problems. I've also been very pleased with my results compiling MPlayer myself, again as-per instructions with no expertise.
Let's hope OpenOffice for OS X stabilises asap, removing a major gripe of his which I do agree with. Having said that, there are a lot of really excellent progs around for reasonable amounts of money - his use of Office and Photoshop to illustrate that point is a bit silly.
Personally, I think OS X offers a good choice between pay-for-polish and DIY-for-free. Linux is better for the latter, but until it offers the polish I am prepared to pay for, I won't switch (besides, the most useful linux stuff is on the mac IMO, OO excepted). Windows I can't comment on: never had to anything interesting with it outside excel.
This is a really good point. I use.mac for synching certain things, but it's no way robust enough, and it has the relationship kinda wrong. Files on the remote server are too distinct from local copies. There needs to be a stronger relationship between the two to prevent errors, breakable of course but only explicitly (I'm sure others would disagree, but I'd at least like the option).
I think that, unfortunately, what you're missing in your post is collaboration. It's all very well syncing whole files, and rarely having conflicts because it's always you using the data, but a collaborative system must be more sophisticated. Check-in check-out helps, but incremental changes with a transaction system and rollback is better (particularly if a user is likely to go off-network without checking something back in).
A limitation of current collaborative systems, quite apart from using them offline, is that access control is often retarded, or not integrated with anything else, or totally exclusive to users of the system. I want something where I can create accounts for outsiders who aren't registered with whatever service I'm using so they can use certain data I allow them to. If they register, fine. If not, fine.
Eventually someone will do all the above well, thus integrating distributed back-up, filesharing and off-line use.
What about the 'temporal dithering' mentioned? I assume this means rapidly moving between two different 6-bit values in certain colour channels. If it's fast enough, there's no way the human eye could distinguish between this and a 'true' colour (inverted commas because any RGB representation isn't really the colour it seems anyway, our eyes just can't distinguish actual wavelengths). I think this might relate to the 'sparkle' someone mentioned earlier - it's not just noise-dithering, it's dynamic-noise-dithering.
Can someone who knows lots about LCD screens (dis)confirm this?
I've not found VNC to be fast enough on my home network for things that refresh quickly like movies, but then I've just set it up to allow the mac mini I have as a server to be accessed from various machines - I have no doubt a carefully configured Linux setup would be more efficient. The WAN is certainly the key for devices that stay with you, like a laptop or phone. Either way, as soon as there's enough bandwidth to stream what's on the monitor at a decent clip, there's no longer any need to have device-local storage or processing power beyond what's needed to run the (very thin) OS.
Just imagine, you put your phone down on a table, it projects a screen onto the nearest wall and a keyboard onto the desk, and off you go! (Yes, I know this would use crazy amounts of power, can't a guy dream?)
I reckon there's a lot of sense to that.
One can even envisage a situation where you pay a hire fee for a given level of processing power, another for your storage, and another for bandwidth. Any of these could be pay-as-you-go or pay monthly, so if you suddenly need to process something huge, you pay your money and it's done in seconds, or you let it chug away and do it on the cheap on your contract. Similarly, you can have streaming access to your data fast enough that you're effectively carrying everything in your phone, or you can have it cache locally and be more selective.
An alternative to this is that the central 'router' contain almost everything - CPU(s), HDs, RAM - and then you have a variety of clients. So your PC becomes just a terminal, as does your 'phone'.
There would be three types of device: the 'server' (which most people would probably just think of as 'the PC'), terminals (anything capable of full IO with the server) and very lightweight devices like the fridge, which don't give you access to everything, but can now communicate.
I like this system because as soon as bandwidth is large enough to handle piping the graphics to your terminal, the need to shove all that processing power and storage into handheld devices and laptops disappears. Much simpler, and probably much cheaper in the end.
Absolutely. Speech recognition will only ever be useful when you aren't at a keyboard - "Tea, Earl grey, hot" as you walk in the door springs to mind - or when the computer can acutally work out what you mean from what you said: "Loop through this array of polygons, checking for any doubles, and draw the remainder to screen if they are within the viewport, oh, and parallelise that while you're at it", and it writes out all the math and logic for you.
Personally, I agree with the grandparent that flexibility is the way forward. I would prefer a tool-based paradigm that separates executables (actions to perform) from documents (data on which to operate), so it's more like picking up a pen to do drawing, or scissors to do cutting.
It would allow CLI-style use of multiple independent tools on the same data sets, while providing a far more intuitive interface than anything else around. It would require widespread data format standards, but such things are already around to some extent (XML). Imagine that boundary between document types breaking down, it would be like being back at primary school again, and gluing bits of pasta (vertex data, to be rendered in 3D where possible!) and glitter (an image format of your choosing) round the edge of your essay (written in a text editor, styled in a typesetter).
The great thing about it is, interfaces could be much more experimental on a per-tool basis without losing ease-of-use, because each one could be intuitive in its particular role. So paint programs use your fingers or a stylus, 3D stuff uses a haptic glove if available, text editors can be spoken/typed/written, etc. All you would need is for the overarching tool-selection interface to be simple and consistent, and the file-manager I suppose, but then again why not let people choose their file manager?
Slashdot Mirror
As with anything, it's horses for courses to some extent. I switched from Nokia to Sony many years ago (back when phone interfaces were all simple and often elegant). I eventually found the Sony interface was getting too awkward, and went back to Nokia hoping for something as intuitive as I used to have.
I totally agree that S60 is a bad tradeoff between complexity and flexibility as compared with Nokia's older phones, but I prefer it to Sony's for the tasks I do most.
Here's hoping the iPhone really does combine offer a better combination of the two. I'm also in the UK, and echo your comment about Orange. I'd love it if they'd pair with Virgin and actually have a chance of delivering okay customer service (IMO). Jobs and Branson both appear to be value-driven entrepeneurs - Jobs always keen on elegance, Branson on great customer service, niether of them necessarily doing things other people can't, but always doing it differently. They don't always hit the target (G4 Cube, Virgin TV/broadband horror stories I've come across, although that may be due to underlying infrastructure issues), but at least they're driving improvement in unusual directions.
I fully agree that it's sad, but while it isn't fully rational, it might be at least partially so.
No human is fully rational, so it sometimes pays to be rational about your irrationality. If you just can't help wanting something that badly, it might be more rational to indulge yourself than to get all depressed not buying it. Of course, it means you're lacking in willpower, but hey, maybe you've learned to accept that, too!
Once you start talking about being completely rational, you should really get down to brass tacks and justify your continued existence before making any other decision. Since I've not seen a justification for our existence that more than a minority agree with (unless you allow 'god wishes it' without specifying which god - surely cheating), I don't think it's a great way to operate.
I'd do the copy-and-paste (and a few other context-related features) with the same 'right-click' mechanism as recent Apple laptops - a two-fingered click brings up a little contextual menu.
Anyone from Apple reading this? Isn't the API for this already built into Cocoa? How hard can it be?
I'd imagine a web-based iPhone app could probably work as a pretty good IM client.
I'm hugely disappointed there's no proper API - I think OSX's API is one of its most compelling aspects - but the web-based system should be capable of filling some gaps.
This is true to a certain extent, and more so with TVs than computers. However, when a computer is Bob, you may well use the computer to do the interception/cracking or whatever. At this point, we're back to Bob being receiver and cracker, or to Eve being able to control Bob, depending how you look at it.
Either way, it's not quite as clear-cut in all situations.
Surely that should be 'In Soviet Russia, a Beowulf cluster of Giant Penguins imagines you'.
They tried roughly this already. It didn't work. Why? Because entrants to the new 'Works with Mac OS' market didn't build commodity boxes with super-slim margins, they built high-margin, high-performance machines that didn't necessarily kill Apple's sales, but wiped out the bulk of its profits. This is how Apple ended up at $12 a share, just before Jobs came back.
This isn't about 'vision' or 'consumer electronics vs. computers', it's about economics. If your profit comes from hardware, you have to sell profitable hardware. Apple can't compete economically in the OS market because there's a monopoly. They can compete in the hardware market. The only thing that has any chance of competing with MS on OS alone is 'guerilla' OSes like Linux. Conventional business, like conventional warfare against an army 100x the size of yours, just won't work.
Ironically, the best way to destroy MS might be to separate Mac OS and Mac hardware into two companies - it would immediately become clear the extent to which Microsoft is a monopoly, which only has a competitor because that competitor isn't really in the same market!
I think it's a shame you've been modded down, AC or not, because you're precisely correct. Piracy as used when the import duty on tea to Britain was something like 100% (or more, I can't recall) did not protest paying for the product at all, it protested the interlocutor.
We now have technology sufficient to the task of actually distributing the cost of mkaing the movie fairly directly to the consumer. No-one's figured out a way to make it happen, but the market appears to be 'aware' on some level that it should. A key element of this is that a film/song should actually get cheaper as more people buy it. Perhaps the reasons the model hasn't been tried is it could potentially involve refunding earlier customers the difference.
This may sound far-fetched, but think about it: cost of the system running this will quickly be minimal compared to capital outlay on films. The only (admittedly large) problem is getting that capital together in advance. Perhaps a subscription fee, which is potentially refunded depending on usage but only after a year? So if 100 million people worldwide watch a $250mn film (add $50mn profit and infrastructure costs, say), they pay $3 for the privelege, and that is all they pay ever, for every format or device they ever watch it on, because format-shifting has negligible cost.
I know this is a pipe-dream (although a large film co. could shift to this model gradually and remain profitable, I think), but to anyone who asks when would be a good time to stop pirating: when the system is run like this, that's when.
Is the stone a measurement of mass? I thought the Imperial mass measure was a 'slug'.
There are some very good points here, and it puts me in mind of how bandwidth is paid for in general. Someone running a website is charged for bandwidth usage by their provider, yet they are providing the content that means people want to use the service at all. They also charge the end user, but generally not per-quantity.
Would anyone else agree that a better model might be end-users paying a rate for data downloaded which really represented the cost, and content providers not paying? It seems fairer to me - you pay for the stuff you want, and that's how the network is supported. I know a lot of the big publishers are 'evil corporations', and one's instinct might be that they should pay for distributing stuff, but ultimately they'll pass the costs on one way or another anyway, and internet distribution (bar advertising) is unique in that it's genuinely on-demand use of the infrastructure. Actually, I think I've just found the flaw in my theory - bandwidth-intensive flash adverts that drive up costs, although that could be solved at the browser end, and probably would if people were really paying for data.
They could apply different costs to different methods of getting data, so when your phone moves out of WiFi, you start using 3G or EDGE and the price per byte goes up. The market as a whole would be much free, I think, than under the current model.
Not sure if this relates, but a study of the children and grandchildren of women subjected to extreme malnourishment during WWII showed some interesting results.
Firstly, whether the mother or the foetus lost weight depended on how far along the foetus was - I can't remember exactly, but I believe if the foetus was almost capable of surviving on its own, the mother's body basically sacrificed itself.
Perhaps more weirdly, while the babies of malnourished mothers were not more likely to be smaller as adults, their children were. So there was some kind of knock-on effect on the grandchildren of the original women. I think they ascribed this to biochemical changes in the middle generation rather than any genetic effect, but they couldn't pinpoint it. Perhaps some very subtle DNA alteration kicked in, and the children of malnourished mothers produced smaller offspring (more likely to survive in a low-food environment) 'just in case'.
I love that 'bootstrap routine and program code of an life form building nano-machine' constitutes 'layman's terms' on slashdot.
This is a really good point, and very well put. Allow me to adjust what I said slightly. While the 'service' I would like to buy would include the ability to obtain appropriate copies for all my playing devices, in perpetuity, I would not feel hard-done-by if I wasn't allowed to pass copies to a friend for free. CDs can remain CDs, you're quite right; I think it's the 'service' model which has yet to be properly worked out.
Of course, as TFA says, it may not be possible for them to stop me, but I'm just saying my problem with pay-for is not the lack of notional permission to reproduce it, but the actual utility of the format. As someone said earlier, the current situation is one of legislation, litigation, and cartels maintaining market values massively distant from actual marginal value.
Perhaps the ideal system would not DRM files per se, but simply tag them as mine (as Apple's higher quality music does). The key change from the current system would be that distributing the music should not result in some ridiculous damages case, but simply the revocation of that item for that user - you break the contract, you lose the service, as with your warranty example.
Hmm, thinking about it, I'm not so sure it would work, but it'll be interesting to see how Apple's new system pans out.
Your last point is the key for me. If I'm going to pay, it must be more functional than the free version. I want to pay a single fee for each song/show/film, which gives me the ability to obtain it in a variety of formats including new ones as they arrive. Technology moves so fast that even if you're allowed to buy the 'best' current version it will be defunct in shorter and shorter spaces of time - either because of a higher-quality (eg DVD to 720 to 1080) or more flexible (eg mp3, divx) version.
Damn straight, and this brings in the most important restriction on a child's use of the internet, the telly, and computer games - time.
I would have no problem with kids watching/browsing/playing, were it not for all the advertising gradually warping their brains, and the largely one-way interaction atrophying their social skills. Couple of hours a day, max.
Uhh... except of course I hated my parents not letting me play Street Fighter II every waking moment outside school hours. Mind you, four-player Wii I would not object to.
Er, you mean they'll never make any? ;)
Conversely, some things are intuitive at first, but fall down later because they are not logical, so you can't apply things you've learned one place elsewhere.
The best interfaces are logical, and expose the logic as intuitively and quickly as possible. I guess the problem is working out what someone will be able to interpret quickly. Case in point: icon-only control systems may take up less space, but unless the pictures actually represent something the user already knows about, they are less easy to intuit than written names.
Sorry, I should have said: I quite appreciate Linux is much less DIY than it used to be, if you go for Ubuntu, and all the PCs I set up and ran in our company used it (three, alongside a Mac). I meant a genuine compliment: there is more scope for DIY on linux than Mac, if you want it.
Further to that, I guess when I say polish I mean something more than system-wide gloss. I would gladly exchange aqua for something faster, if it maintained all the functionality (which includes being unobtrusive - I personally think the pale grey is less distracting than many schemes). What I'm prepared to pay for is the level of ease-of-use of certain specific programmes that really are best-in-class. Video editing is the key one for me (plus I've been using Photoshop since 2.0, and I've grown too close to it to give it up), but I'm sure other people have other things. I also think mac programmes in general are more consistent with one another than on other platforms, and this cuts down learning time.
Quite agree with your price point - it would be a huge plus if you could build a mac from scratch, and all my HD systems are hand-made to keep the cost down - but it's worth thinking about the true cost: if I can get even 5% more done each day because I've got a superb programme running on a very predictable OS/hardware combo that I barely have to maintain, I'll easily make back the cost of the computer, and in my experience I get quite a bit more than a 5% bonus. One-off unusual tasks with videos and images are quicker on the mac, and I can still write bash scripts for anything that needs automating.
Plus, that's more subjective than he allows for. Despite being a born-and-bred Mac user, when I found my version of Fontographer wouldn't play with OS X (this was back in the day), I didn't have any problems at all firing up X11 and installing a free UNIX tool I'd found on the 'net. I was fairly new to the command-line, and totally new to X, but had no problems. I've also been very pleased with my results compiling MPlayer myself, again as-per instructions with no expertise.
Let's hope OpenOffice for OS X stabilises asap, removing a major gripe of his which I do agree with. Having said that, there are a lot of really excellent progs around for reasonable amounts of money - his use of Office and Photoshop to illustrate that point is a bit silly.
Personally, I think OS X offers a good choice between pay-for-polish and DIY-for-free. Linux is better for the latter, but until it offers the polish I am prepared to pay for, I won't switch (besides, the most useful linux stuff is on the mac IMO, OO excepted). Windows I can't comment on: never had to anything interesting with it outside excel.
This is a really good point. I use .mac for synching certain things, but it's no way robust enough, and it has the relationship kinda wrong. Files on the remote server are too distinct from local copies. There needs to be a stronger relationship between the two to prevent errors, breakable of course but only explicitly (I'm sure others would disagree, but I'd at least like the option).
I think that, unfortunately, what you're missing in your post is collaboration. It's all very well syncing whole files, and rarely having conflicts because it's always you using the data, but a collaborative system must be more sophisticated. Check-in check-out helps, but incremental changes with a transaction system and rollback is better (particularly if a user is likely to go off-network without checking something back in).
A limitation of current collaborative systems, quite apart from using them offline, is that access control is often retarded, or not integrated with anything else, or totally exclusive to users of the system. I want something where I can create accounts for outsiders who aren't registered with whatever service I'm using so they can use certain data I allow them to. If they register, fine. If not, fine.
Eventually someone will do all the above well, thus integrating distributed back-up, filesharing and off-line use.
What about the 'temporal dithering' mentioned? I assume this means rapidly moving between two different 6-bit values in certain colour channels. If it's fast enough, there's no way the human eye could distinguish between this and a 'true' colour (inverted commas because any RGB representation isn't really the colour it seems anyway, our eyes just can't distinguish actual wavelengths). I think this might relate to the 'sparkle' someone mentioned earlier - it's not just noise-dithering, it's dynamic-noise-dithering.
Can someone who knows lots about LCD screens (dis)confirm this?
I've not found VNC to be fast enough on my home network for things that refresh quickly like movies, but then I've just set it up to allow the mac mini I have as a server to be accessed from various machines - I have no doubt a carefully configured Linux setup would be more efficient. The WAN is certainly the key for devices that stay with you, like a laptop or phone. Either way, as soon as there's enough bandwidth to stream what's on the monitor at a decent clip, there's no longer any need to have device-local storage or processing power beyond what's needed to run the (very thin) OS.
Just imagine, you put your phone down on a table, it projects a screen onto the nearest wall and a keyboard onto the desk, and off you go! (Yes, I know this would use crazy amounts of power, can't a guy dream?)
I reckon there's a lot of sense to that. One can even envisage a situation where you pay a hire fee for a given level of processing power, another for your storage, and another for bandwidth. Any of these could be pay-as-you-go or pay monthly, so if you suddenly need to process something huge, you pay your money and it's done in seconds, or you let it chug away and do it on the cheap on your contract. Similarly, you can have streaming access to your data fast enough that you're effectively carrying everything in your phone, or you can have it cache locally and be more selective.
An alternative to this is that the central 'router' contain almost everything - CPU(s), HDs, RAM - and then you have a variety of clients. So your PC becomes just a terminal, as does your 'phone'.
There would be three types of device: the 'server' (which most people would probably just think of as 'the PC'), terminals (anything capable of full IO with the server) and very lightweight devices like the fridge, which don't give you access to everything, but can now communicate.
I like this system because as soon as bandwidth is large enough to handle piping the graphics to your terminal, the need to shove all that processing power and storage into handheld devices and laptops disappears. Much simpler, and probably much cheaper in the end.
Absolutely. Speech recognition will only ever be useful when you aren't at a keyboard - "Tea, Earl grey, hot" as you walk in the door springs to mind - or when the computer can acutally work out what you mean from what you said: "Loop through this array of polygons, checking for any doubles, and draw the remainder to screen if they are within the viewport, oh, and parallelise that while you're at it", and it writes out all the math and logic for you.
Personally, I agree with the grandparent that flexibility is the way forward. I would prefer a tool-based paradigm that separates executables (actions to perform) from documents (data on which to operate), so it's more like picking up a pen to do drawing, or scissors to do cutting.
It would allow CLI-style use of multiple independent tools on the same data sets, while providing a far more intuitive interface than anything else around. It would require widespread data format standards, but such things are already around to some extent (XML). Imagine that boundary between document types breaking down, it would be like being back at primary school again, and gluing bits of pasta (vertex data, to be rendered in 3D where possible!) and glitter (an image format of your choosing) round the edge of your essay (written in a text editor, styled in a typesetter).
The great thing about it is, interfaces could be much more experimental on a per-tool basis without losing ease-of-use, because each one could be intuitive in its particular role. So paint programs use your fingers or a stylus, 3D stuff uses a haptic glove if available, text editors can be spoken/typed/written, etc. All you would need is for the overarching tool-selection interface to be simple and consistent, and the file-manager I suppose, but then again why not let people choose their file manager?