Just doing something different is not innovation, thats differentiation. Innovation involves a certain level of enhancement as well. And I'm not seeing where the ribbon actually innovates.
Most people have no problem knowing where the icons they use and want are. Making them bigger and smaller with use and resizing and reorganizing the toolbars will be nothing but an annoyance.
If its any good, I expect it to be replicated. I really dont think you'll see many other apps doing likewise. Toolbars work fine, thanks much.
Much of the world is built around stealing peoples ideas and putting them to good use. Whether its MS or someone else, as long as their actually putting these ideas to good use, they deserve credit. If MS does take some of the ideas from Singularity and makes a more microkernel-alike system, they deserve massive props. OS-X isnt still built off Mach is it?-- I know they did something off Mach, but wait, OSX was a BSD... *shrug* (its been a while). Anyways I'm pretty sure Apple somehow managed to never let anyone actually reap the benefits of the microkernels they've used, and I know they at one point did something with Mach. So, here's a small cheers to MS.
Linux's religious insistence on being monolithic is egotistical and fucking retarded. It stiffles innovation again and again; video card drivers, the reiserfs fiasco, the years it took to get a userland fs in, &c
Err, the radio said as much itself; "The digital data from the ADC is processed into I and Q format using a direct digital converter (DDC)". Yup. Just a FFT.
You have hit upon one of the more sublime elements of the future. A room will have just a wireless keyboard and mouse, and likely a variety of displays. At some point the concept of a monolithic computer system will need a good once over. "The network is the computer"
/me raises hand as reading mostly the electronic copies
my laptop weights 2.2 pounds. my charles petzold' programming windows must weigh at least 8. guess how many copies i can fit on my laptop?
its a pity oreilly doesnt do the same. i have to find copies of my books on bittorrent. ssshhh, please dont tell anyone. tim please dont hurt me. as a recently-former student i cant wait till i can think of pricing the safari library.
0) stop running your webserver on your frakking DSL : unixshell is sweet for this (Linnode alike with 5x better plans). 1) dont ever do anything interesting 2) do something so interesting slashdot will never publish it 3) run lighttpd and back off your useless CGI
True to the MS style, most D3D books simply tell you how to do whatever you want to do. OpenGL is considerably less about knowing the relatively simple API and more about knowing how the systems actually work. It shows in all the manuals. The official OpenGL red book is a fine tutorial [for the hardcore] on how 3d programming in general works. Good luck finding such explanations in D3D books.
Furthermore, D3D will never extend past the MS platform. Whoo, Xbox360. Like hell you'll ever get to program for it. How many smartphones have D3D acceleration? On the other hand, OpenGL already has OpenGL ES for mobile platforms. Hardware support is small, but it will grow.
I like giving my games hardware acceleration when I'm using the OneTrueOS, be it FreeBSD, Solaris, AIX, hell, even Linux. Lets see you do that in D3D.
Stable API's are good. OpenGL 2.0 is here to stay for a long time. Compare this to DirectX. We're at 9.0c, iirc? We've been here a while, DirectX X is taking a while but thats because its so completely new. Just imagine the DirectX 3->4->5->6->7 evolution which occured in, oh, what, three years? Maybe five? Thats not a standard. Thats just enough kruft to accelerate the bejeezus out of whatever hardware was out at the time.
He's right, if you're interest is in joining a game company, more companies use D3D. But even for people seeking to become professionals, I would advise learning OpenGl. For anyone who isnt sure they want to be a dedicated game programmer their whole life, learn the standard, learn OpenGL.
1000 years of nonlinear history, 1000 plateaus, guns gems and steel, web of life, &c &c. All the good systems theory junk. Godel Escher Bach &tGB.
Must like reading, all hefty reading, but its really good stuff. Further suggestions in this vein would be appreciated.
For a little bit of Sci Fi meets Classics, Ilium was really damned good. I've ahd the sequel sitting on my desk for two weeks and I'm too afraid to pick it up because I know it wont be cant be could never be as good. There's actually a bit of a renisance of really good Sci Fi stuff coming out... Alistar Renolds, Richard K Morgan, &c. I havent read these, I still have to finish 1000 plateaus first, but they're sitting on my desk, reviews are very solid, and I'm confident it'll be agood time.
the cheapest route will be a hackmeal of wireless gear.
The good news is the router. Small and prepackaged, wireless and good ole trusty 802.3 ethernet, 2-7 watts power, some have USB 2.0 connectivity. Most are 133-200 mhz, with the great mode being 200 mhz, however there are some faster. 200 mhz should be plenty for what you want; the routers all have excellent DMA capabilities for shuffling data around. No problems decoding a couple mp3's while doing some file access. As an interesting bonus, a number of the units will run off of basically any sort of voltage differential you can find; some of the Linksys's are reported good at 48v (although i cant imagine the voltage regulator onboard agrees with that statement). I cant remember, but there were a couple which seemed to run fine off of 4-6v. Almost every wall wart is a 12v supply, they just have flexible regulators.
The bad news is its basically a $60 CPU brick + wireless. Not bad at all the price, but you still need $20 USB hard drive enclosure, $100 hard drive (ahem, Seagate 7200.9; 5 year warranty & fast but noticably louder than fluid-bearings), and a usb sound card. I picked up a handful of C-Media 8 channel output cards which are, against all odds here, not entirely crap-- espsecially suprising at ~$20 a pop. So, really, the price isnt that bad. $100 basically for cpu & ram & wireless and the necessary peripherials, sans the HD itself. The suck is the stack of hard drive sound card and router, although I happen to like the modularity. Quite unexpectedly, I've found I'm usually using one or the other peripherial, oddly enough. Anyways, I personally feel it compares favorably to the mini-itx's which seem to run at truly inordate dollar sums, in complete spite of the inordately small ammounts of natural resources manufacturing required. And good luck finding boxes for em.
Carefully check out http://wiki.openwrt.org/TableOfHardware for complete router specs. Since you're going to have a hard drive, you just need the 4mb flash, but the RAM should be at least 16, quite preferably 32 megs. Really, its enough. There are a couple 64meg ones. If you really want to get persnickety, consider one of the uber-fast USB flash sticks and set it up as swap space.;) I should actually try that some day... Anyways, that awsome table will also tell you if a particular router supported by the best Linux for Routers distribution out there, OpenWrt. Sveasoft++ for actual linux users.
I'm having some issues with my USB drive right now, otherwise I'd run some NFS benches for you. I've always found it to be acceptable, but I've never really benched it. I'd suspect USB 2.0 as the primary bottleneck, ethernet as the second, router as the third.
I was going to mention the old DEC boxes with dual pcmcia, but mine are both dead,-- not a good sign-- I cant find the blasted name anywhere, and-- deathknell here-- you want a low power solution.
Since there are no standards for anything but the lowest layers in zigbee, then it is only marginally better than using a proprietary standard.
My biggest qualm is that Zigbee itself is a highly proprietary standard. I'm fine with having computers orchestrate all levels of interaction, being middlemen in the sensor networks to get sensors to talk to each other, but as it stands we've got to sit around with our thumbs up our asses while people release readymade Zigbee solutions... because we cant built hte @#$#@$ ourselves.
Every free sensor net project I've seen has run screaming away from Zigbee, down a layer or two towards 802.15.4. A lot of it is as you state, its simpler easier to build off a basic 802.15.4. But in truth, there is some intercompatibility from Zigbee, and it would be sweet, getting routing &c for free, getting intercompatibility, but the standard is protected by a lot of (largely self-defeating) corporate interest.
I seem to remember somewhere that when Zigbee was being announced and talked about and hyped, it was always toted as a "free(r)" standard than bluetooth. Of course, that went out the window and now you need another corporate license to even think about using the technology. At least they released the spec for academic use, but its basically reverted to bluetooth levels of control.
The stacks are much smaller. 30k v. 200k. If they actually gave a license open source could use though, I'd be willing to be you'd see some amazing 10k Zigbee stacks cropping up. But no ones going to do that because the technology isnt open. I doubt any corporation is going to want to explot the 10k zigbee stack niche either-- if they do it certainly wont be pioneered by a coder.
There's an amazing market built around closed source technology. At some point though, technology just needs saturation, to be used. To develop community and support. Look at Microchip; they're a company built entirely around hobbyists and community. No one uses that stuff in industry; its either an ARM or 8051 depending on if you need processing power or low cost. The best thing Zigbee coudl've done would've been to make a standard hackers could use. Get some bleeding edge people to start pushing it up toward the chasm. We really dont know what this technology is useful for. Its basically lightweight radio i/o. We've never had this before; really, nothing standard. But no, greedy asshole corporate committee work fucks up another could-have-been technology. I predict a hundred proprietary flopped home automation systems. NEXT!
Just look at all the shit hobbyists have put web browsers on. Webservers in 1k of ram with 12k flash, &c &C. In general, I'd say that even now, corporations are rarely the trendsetters. There's always the bleeding edge pushing harder faster.
I will now laugh at our pathetic corporate world. Myren
You can build your own squeezebox-alike cheap as nuts. Its a hack, but its cool. Get an OpenWRT box with a USB output and hook up a USB sound card. Cost should come in under $120. You can also get a ATI Remote Wonder for RF remote control. You should only need one or two recievers for the whole house, and you can sometimes buy the reciever & the remotes seperately on ebay. For now I leave sync issues as DIY. More some day.
The best way to counter tyranny is to provide more openness. Continue developing more open information solutions, invent new ways to connect people, generate more tools that future hackers will use to create their own new connections. As the cycle grows and progresses, the corporate suit's ability to take that all away shrinks. That is how a market works. On the other hand, if the mega corps provide everything the consumer needs, if they notice no difference in the sinister new net, they are free to take it all away from us.
There does, though, have to be some h/w support for shared-disk. Cluster Interconnect, anyone????
What, like SCSI? FC?;) It would re-complexify the situation, but you can just hook up a bunch of SCSI controllers to a chain. Typically this is used for a Hot/Cold situation, but some solutions are Hot/Hot.
More obviously, how about simply making files smarter rather than gluing ever more solutions on to guessingly tease out ever more intelligence from our files?
Everything is not a file, everything is an object. A file is simply a collcetion of data. An object has associate things you can do with it.
In debian there is an/etc/alternatives directory which holds systemwide links like "x-window-manager" & "www-browser"... they're systemwide but they're something.
Whats really needed on this account is better mime type support. Something to associate various types against each other. I propose we start an online web 2.0 folksonomy service for tagging our mime types.
Wow man, I dont know what all these silly clods are babbling about. Linux aint any good if it dont run. Sorry for your shite experience mate, I hope the dev's patched up the new release to work better for your perfectly good hardware.
-----
My god, fucking animal...
Linux kernel itself has half a million kernel boot parameters to support various Dell branded pieces of craps. The idea is that we make things work, not piss on people when their hardware isnt working.
I have no idea what this particular HD Halo argument is over, but I believe you just tried to claim the difference between 840x480 and 1280x768 isnt really distinct...
HAHAHA.
It would be interesting to see what framegrabber Bungie used. I'm not really sure how you go about storing 720p content.
Slashdot Mirror
Just doing something different is not innovation, thats differentiation. Innovation involves a certain level of enhancement as well. And I'm not seeing where the ribbon actually innovates.
Most people have no problem knowing where the icons they use and want are. Making them bigger and smaller with use and resizing and reorganizing the toolbars will be nothing but an annoyance.
If its any good, I expect it to be replicated. I really dont think you'll see many other apps doing likewise. Toolbars work fine, thanks much.
Myren
Much of the world is built around stealing peoples ideas and putting them to good use. Whether its MS or someone else, as long as their actually putting these ideas to good use, they deserve credit. If MS does take some of the ideas from Singularity and makes a more microkernel-alike system, they deserve massive props. OS-X isnt still built off Mach is it?-- I know they did something off Mach, but wait, OSX was a BSD... *shrug* (its been a while). Anyways I'm pretty sure Apple somehow managed to never let anyone actually reap the benefits of the microkernels they've used, and I know they at one point did something with Mach. So, here's a small cheers to MS.
Linux's religious insistence on being monolithic is egotistical and fucking retarded. It stiffles innovation again and again; video card drivers, the reiserfs fiasco, the years it took to get a userland fs in, &c
Myren
If The Register is the National Inqurirer, what is The Inquirer?
OTOH, their useful post count has rapidly converged on zero. They were the next one off the RSS reader island anywho...
Err, the radio said as much itself; "The digital data from the ADC is processed into I and Q format using a direct digital converter (DDC)". Yup. Just a FFT.
is the direct digital converter some magic piece of hardware, or is it just a FFT on samples?
Umm, most unix people DO operate in 4D already. 2D graphical workspace, one depth dimension and one dimension of virtual desktop spaces.
OTOH your project shold be damned fucking funny. Although from the screen shots I have no inkling of idea how its 4D.
Prediction;
The revolution will happen when HCI stops measuring efficiency as "time to complete task" and "task completion %".
Discuss.
You have hit upon one of the more sublime elements of the future. A room will have just a wireless keyboard and mouse, and likely a variety of displays. At some point the concept of a monolithic computer system will need a good once over. "The network is the computer"
/me raises hand as reading mostly the electronic copies
my laptop weights 2.2 pounds. my charles petzold' programming windows must weigh at least 8. guess how many copies i can fit on my laptop?
its a pity oreilly doesnt do the same. i have to find copies of my books on bittorrent. ssshhh, please dont tell anyone. tim please dont hurt me. as a recently-former student i cant wait till i can think of pricing the safari library.
myren
0) stop running your webserver on your frakking DSL : unixshell is sweet for this (Linnode alike with 5x better plans).
1) dont ever do anything interesting
2) do something so interesting slashdot will never publish it
3) run lighttpd and back off your useless CGI
True to the MS style, most D3D books simply tell you how to do whatever you want to do. OpenGL is considerably less about knowing the relatively simple API and more about knowing how the systems actually work. It shows in all the manuals. The official OpenGL red book is a fine tutorial [for the hardcore] on how 3d programming in general works. Good luck finding such explanations in D3D books.
Furthermore, D3D will never extend past the MS platform. Whoo, Xbox360. Like hell you'll ever get to program for it. How many smartphones have D3D acceleration? On the other hand, OpenGL already has OpenGL ES for mobile platforms. Hardware support is small, but it will grow.
I like giving my games hardware acceleration when I'm using the OneTrueOS, be it FreeBSD, Solaris, AIX, hell, even Linux. Lets see you do that in D3D.
Stable API's are good. OpenGL 2.0 is here to stay for a long time. Compare this to DirectX. We're at 9.0c, iirc? We've been here a while, DirectX X is taking a while but thats because its so completely new. Just imagine the DirectX 3->4->5->6->7 evolution which occured in, oh, what, three years? Maybe five? Thats not a standard. Thats just enough kruft to accelerate the bejeezus out of whatever hardware was out at the time.
He's right, if you're interest is in joining a game company, more companies use D3D. But even for people seeking to become professionals, I would advise learning OpenGl. For anyone who isnt sure they want to be a dedicated game programmer their whole life, learn the standard, learn OpenGL.
I happen to enjoy Stroustroups' (sp massacre, sorry) "The C++ Language" in its hardcover 1000+ page with two bookmarks format you insensitive clod.
1000 years of nonlinear history, 1000 plateaus, guns gems and steel, web of life, &c &c. All the good systems theory junk. Godel Escher Bach &tGB.
Must like reading, all hefty reading, but its really good stuff.
Further suggestions in this vein would be appreciated.
For a little bit of Sci Fi meets Classics, Ilium was really damned good. I've ahd the sequel sitting on my desk for two weeks and I'm too afraid to pick it up because I know it wont be cant be could never be as good. There's actually a bit of a renisance of really good Sci Fi stuff coming out... Alistar Renolds, Richard K Morgan, &c. I havent read these, I still have to finish 1000 plateaus first, but they're sitting on my desk, reviews are very solid, and I'm confident it'll be agood time.
Myren
the cheapest route will be a hackmeal of wireless gear.
;) I should actually try that some day... Anyways, that awsome table will also tell you if a particular router supported by the best Linux for Routers distribution out there, OpenWrt. Sveasoft++ for actual linux users.
The good news is the router. Small and prepackaged, wireless and good ole trusty 802.3 ethernet, 2-7 watts power, some have USB 2.0 connectivity. Most are 133-200 mhz, with the great mode being 200 mhz, however there are some faster. 200 mhz should be plenty for what you want; the routers all have excellent DMA capabilities for shuffling data around. No problems decoding a couple mp3's while doing some file access. As an interesting bonus, a number of the units will run off of basically any sort of voltage differential you can find; some of the Linksys's are reported good at 48v (although i cant imagine the voltage regulator onboard agrees with that statement). I cant remember, but there were a couple which seemed to run fine off of 4-6v. Almost every wall wart is a 12v supply, they just have flexible regulators.
The bad news is its basically a $60 CPU brick + wireless. Not bad at all the price, but you still need $20 USB hard drive enclosure, $100 hard drive (ahem, Seagate 7200.9; 5 year warranty & fast but noticably louder than fluid-bearings), and a usb sound card. I picked up a handful of C-Media 8 channel output cards which are, against all odds here, not entirely crap-- espsecially suprising at ~$20 a pop. So, really, the price isnt that bad. $100 basically for cpu & ram & wireless and the necessary peripherials, sans the HD itself. The suck is the stack of hard drive sound card and router, although I happen to like the modularity. Quite unexpectedly, I've found I'm usually using one or the other peripherial, oddly enough. Anyways, I personally feel it compares favorably to the mini-itx's which seem to run at truly inordate dollar sums, in complete spite of the inordately small ammounts of natural resources manufacturing required. And good luck finding boxes for em.
Carefully check out http://wiki.openwrt.org/TableOfHardware for complete router specs. Since you're going to have a hard drive, you just need the 4mb flash, but the RAM should be at least 16, quite preferably 32 megs. Really, its enough. There are a couple 64meg ones. If you really want to get persnickety, consider one of the uber-fast USB flash sticks and set it up as swap space.
I'm having some issues with my USB drive right now, otherwise I'd run some NFS benches for you. I've always found it to be acceptable, but I've never really benched it. I'd suspect USB 2.0 as the primary bottleneck, ethernet as the second, router as the third.
I was going to mention the old DEC boxes with dual pcmcia, but mine are both dead,-- not a good sign-- I cant find the blasted name anywhere, and-- deathknell here-- you want a low power solution.
Good luck
Myren
Since there are no standards for anything but the lowest layers in zigbee, then it is only marginally better than using a proprietary standard.
My biggest qualm is that Zigbee itself is a highly proprietary standard. I'm fine with having computers orchestrate all levels of interaction, being middlemen in the sensor networks to get sensors to talk to each other, but as it stands we've got to sit around with our thumbs up our asses while people release readymade Zigbee solutions... because we cant built hte @#$#@$ ourselves.
Every free sensor net project I've seen has run screaming away from Zigbee, down a layer or two towards 802.15.4. A lot of it is as you state, its simpler easier to build off a basic 802.15.4. But in truth, there is some intercompatibility from Zigbee, and it would be sweet, getting routing &c for free, getting intercompatibility, but the standard is protected by a lot of (largely self-defeating) corporate interest.
Myren
I seem to remember somewhere that when Zigbee was being announced and talked about and hyped, it was always toted as a "free(r)" standard than bluetooth. Of course, that went out the window and now you need another corporate license to even think about using the technology. At least they released the spec for academic use, but its basically reverted to bluetooth levels of control.
The stacks are much smaller. 30k v. 200k. If they actually gave a license open source could use though, I'd be willing to be you'd see some amazing 10k Zigbee stacks cropping up. But no ones going to do that because the technology isnt open. I doubt any corporation is going to want to explot the 10k zigbee stack niche either-- if they do it certainly wont be pioneered by a coder.
There's an amazing market built around closed source technology. At some point though, technology just needs saturation, to be used. To develop community and support. Look at Microchip; they're a company built entirely around hobbyists and community. No one uses that stuff in industry; its either an ARM or 8051 depending on if you need processing power or low cost. The best thing Zigbee coudl've done would've been to make a standard hackers could use. Get some bleeding edge people to start pushing it up toward the chasm. We really dont know what this technology is useful for. Its basically lightweight radio i/o. We've never had this before; really, nothing standard. But no, greedy asshole corporate committee work fucks up another could-have-been technology. I predict a hundred proprietary flopped home automation systems. NEXT!
Just look at all the shit hobbyists have put web browsers on. Webservers in 1k of ram with 12k flash, &c &C. In general, I'd say that even now, corporations are rarely the trendsetters. There's always the bleeding edge pushing harder faster.
I will now laugh at our pathetic corporate world.
Myren
You can build your own squeezebox-alike cheap as nuts. Its a hack, but its cool. Get an OpenWRT box with a USB output and hook up a USB sound card. Cost should come in under $120. You can also get a ATI Remote Wonder for RF remote control. You should only need one or two recievers for the whole house, and you can sometimes buy the reciever & the remotes seperately on ebay. For now I leave sync issues as DIY. More some day.
-Myren
The best way to counter tyranny is to provide more openness. Continue developing more open information solutions, invent new ways to connect people, generate more tools that future hackers will use to create their own new connections. As the cycle grows and progresses, the corporate suit's ability to take that all away shrinks. That is how a market works. On the other hand, if the mega corps provide everything the consumer needs, if they notice no difference in the sinister new net, they are free to take it all away from us.
DRBD is one of the most interesting sw projects out there right now. There's two things on my wishlist:
1. Multicast replication to more than one backup
2. Write support from more than one system.
Coupled with GFS, it would be absolutely astounding. Multiple failure happens, a single backup is not always enough.
There does, though, have to be some h/w support for shared-disk. Cluster Interconnect, anyone????
;) It would re-complexify the situation, but you can just hook up a bunch of SCSI controllers to a chain. Typically this is used for a Hot/Cold situation, but some solutions are Hot/Hot.
What, like SCSI? FC?
More obviously, how about simply making files smarter rather than gluing ever more solutions on to guessingly tease out ever more intelligence from our files?
Everything is not a file, everything is an object. A file is simply a collcetion of data. An object has associate things you can do with it.
It was just a joke in the Ubuntu thread when I said we needed a Web 2.0 tagging type manager! Jeeze luise.. ;-]
In debian there is an /etc/alternatives directory which holds systemwide links like "x-window-manager" & "www-browser"... they're systemwide but they're something.
Whats really needed on this account is better mime type support. Something to associate various types against each other. I propose we start an online web 2.0 folksonomy service for tagging our mime types.
Wow man, I dont know what all these silly clods are babbling about. Linux aint any good if it dont run. Sorry for your shite experience mate, I hope the dev's patched up the new release to work better for your perfectly good hardware.
-----
My god, fucking animal...
Linux kernel itself has half a million kernel boot parameters to support various Dell branded pieces of craps. The idea is that we make things work, not piss on people when their hardware isnt working.
Myren
I have no idea what this particular HD Halo argument is over, but I believe you just tried to claim the difference between 840x480 and 1280x768 isnt really distinct...
HAHAHA.
It would be interesting to see what framegrabber Bungie used. I'm not really sure how you go about storing 720p content.