never since Windows XP have I had any hardware that didn't work out of the box, and most of the time, even before installing the supplied drivers, Windows would have a go at using some kind of default one, no matter how crippled.
Ok, let's go anecdotal ; my mother just replaced her old XP Dell by a shiny new Vista computer. But her USB HP-made flatbed scanner had to be replaced, because there's no drivers for it in vista, and HP didn't bothered to update them. The irony ? Her new quad-core everything but the kitchen sink computer is a HP.
Now, for your tuner problem, it's more than likely that the driver is there and OK, but your tuner needs a firmware that can't be bundled for legal reasons in your distro. How to put the correct fw in the right place is left as an exercise to your imagination.
On a last note, while windows compatibility is pretty much taken for granted when shopping for hardware, it's not that much evident that pieces are not labeled with the necessary windows versions required to support them. Mess with it, and you're out of luck. It's exactly the same for linux, you have to be careful before buying, or to put it in latin, caveat emptor. All the HW I bought since 1998 was known to work with linux beforehand, and most of it still runs flawlessly.
I somehow looked at the project around ver. 0.1.x, but it wasn't even remotely usable then. Your post made me go back and d/l the qemu image (ReactOS 3.10) ready for use. Everybody should give it a try, it's only 34 Mb, qemu for windows included. It works like a breeze on my fedora 10 (under qemu-kvm), with absolutely no config tweaking, and it seems quite stable so far. I'm going to follow this project much more closely from now on, the improvements just have been massive. TY for the tip.
... but when I bought my computer, I asked for vista to be removed and the price refunded. Hoped from shop to shop until I found one that agreed (in fact I was ready for a trek, but the 3rd shop in the street was the good one). He got the deal, and I bought the refund worth of RAM to top of the computer capacity. I was pleased, and so was the seller.
My laptop is an Asus eeepc 900A linux 'edition'. Again, I carefully reviewed the options before buying.
Going along with the article, the first person perspective brings the dilemma of the "silent hero", because that hero is supposed to be you. However, you rarely get to interact with anything beyond acknowledging an NPC's request to retrieve something for them. You don't get to forge a bond with any of the characters based on your personality traits.
Never played DN3D, did you ? I guarantee you're ready to kick ass and chew bubble gum after you leave the game, however well bred and polite you may be before.
Finally, I'd note that if you really don't want a Bayer-pattern sensor, you can get a Fuji camera. They use a type of sensor originally developed by Foveon (which Fuji since bought out) that has individual sensor elements "stacked", to there's a red, green, and blue element at each pixel location. While Fuji's cameras are perfectly good and work quite nicely in general, they're not really a whole lot better than most others...
Foevon sensors are not what I'm looking for. While I do understand they are marginally better suited to b&w photography on principle, the reality of the technology is they have at the moment way less physical pixels than 'classic' sensors, between a 3rd and a 4th if I'm not mistaken. Fuji pulls a marketing stunt by selling 'equivalent' figures based on extrapolation to bayer-layered sensors, which is somewhat close to reality in colour photography, but is meaningless in a b&w context ; you can achieve the same result as the foevon with a bayer-sensor by reducing the resolution to a 4th of the total pixel count, and extracting and averaging 2 by 2 the green pixels to compute only luminance based on those "super pixels". It does look good on screen, but it's kind of small and no more suitable to large prints.
If you look at the picasa album of pictures taken with the Frankencamera (linked from TFA through their homepage), you'll notice an incredible amount of chroma noise and banding. The caption ("can you PS camera do better ?") is ridiculous. I would be shamed if my PS produced that crap.
You missed the detail that those programs are running real-time inside the camera, building panorama's and correcting dynamic range and removing noise, while taking the picture?
Face it : this contraption still delivers phone grade pictures (a bit better because canon's lenses aren't bad) but is boxed like a field camera, and host a computer that has a 10th of the computing power you can put under your desk. At least, I would have expected a full frame 24x36 CCD to match the lens. Now, seeing most everybody is happy with the crapastic in-camera jpegs produced by consumer-grade cameras (and, that is, those 'everybody' are already enough interested in pictures to actually own a digital dedicated camera, many more people are perfectly happy with their phone !), I can't fathom whom might wish to let this thing handle their HDR process. Those already in the know can wait to better post-process at home a shooting session. Others simply won't care nor see the difference.
Moreover, my cameras already tries to remove noise, and my phone (Nokia Navigator) can stitch panos. And both are so bad at it that I never take pictures with my phone (except from the Big SUV parked next to my car I wish to remember the licence plate in case I get a blow in my door's paint), and set NR to low on my camera to use better suited algorithms on my computer.
Agreed. I've got a lx2 myself (for the 60mm tele end of the lx3 zoom is IMO too short for close-up portraits), and I'm currently drooling over the Olympus ep-1. But as soon as light dims, I get back to my dslr. The smallish sensor of the pany gets very noisy real quick.
What I would really love is a CCD or CMOS without bayer color matrix nor antialiasing filter. This camera would be limited to black and white photography (or studio pack shot with 3 exposures behind R,B,G filters), but I expect the result would be outstanding. At the moment, we're fighting with slightly fuzzy shots (thanks to the AA filter) to recreate true pixels intensity after they've been distorted by various digital process. There's been one digital b&w camera, made by kodak. It's said the production number was 80, but from reviews, images were already gorgeous with 6 Mpix 10 years ago. Today, bayer matrix and AA filters are glued on the chip in the manufacturing process, and it's impossible to get rid of it afterwards.
Most settlers and pilgrims went away after an initial explorer returned to tell how beautiful the trip was and how nice the people there were. Or in another scenario, people moved because they would be in harm way should they stay (invasions, wars, etc.) No one in his right mind just went away without the slightest hope of coming back, even if many knew the journey was dangerous.
Turing desserves an apology because any security lookup done during wartime certainly had already uncovered this aspect of his personality, but the brits were more than happy to turn a blind eye on it in spite of the already existing laws.
The later prosecution was application of a double standard in this regard.
Oddly enough, there apparently are some people that still use them. The train ticketing, phone book, and a number of other services are still up and in use.
Not odd at all considering the various threats of Internet, from spam to virus, credit cards frauds, DoS etc. Minitel pretty much insure that whomever you phone is legit provided you don't misstype the phone number. It's a very helpful and desirable feature for some sensitive businesses (chemist ordering prescription drugs, etc.)
And thanks to being a passive terminal, Minitel is immune to virus and trojans by nature. Being so simple, there are no bugs either I'm aware off. And being text only makes for a great bonus to blinds who can plug whatever Braille device they want to use it.
My thinking was about fine art or wedding photographers. More generally, it applies when your workflow starts with a RAW picture and you need to squeeze as much informations as possible out of it. If you work on in-camera made jpegs, there's no real need to turn to 16 bits because any postprocessing will be bad anyway. But as soon as you oversize, rotate, correct perspectives, adjust tones, in fact every operation that creates new pixels via an interpolating algorithm based on neighbours intensity, 16 bits is really the way to get a whole range of smooth in-between colours for smooth shades instead of noisy chroma blobs. The more operations you do on your pictures, the more you add rounding errors and the more likely you'll get patches of uniform (and bad !) colours right in the middle of what you expect to be a clean shade (think skin, for instance). Suppose for instance you create a pixel between one which value on one channel (8 bits) is 126 and another which value is 127 ; your algorithm can choose either value, and this new pixel won't be further discriminated from its neighbour in any further calculation. Whereas in 16 bits, your 2 points would have 2016 and 2032 as intensity value, leaving a range of 16 intensity steps for creating a pixel in-between, this pixel being distinct from its neighbours in any further calculation.
And I have found NO linux software that can apply a profile to a print. So I need a virtual XP just to create printer profiles and print !
You didn't looked hard enough. Cinepaint does this perfectly, and even let you smelt any RGB profile into CMYK or XYZ profiles while showing you the relative gamut transformation in OpenGL 3D no less. Or you can display your image in proof mode, etc.
It's not a question of eye, it's a question of maths. As long as digital photography will create colours by interpolation between RGBG values of a matrix, you'll need empty 'holes' to re-create colours and avoid sampling artefacts. Applying filters (sharpening, border sharpening, USM etc.) also needs room to do their magic, otherwise you're accumulating errors that ultimately show as noise. When your postprocessing is done, then you can downgrade to 8 bits / channel without too much loss for printing or display.
Believe it or else, there's a real difference between a 8 bits and a 16 bits workflow; you'd need to try for yourself to be convinced, but for a pro, it's the difference between work or joblessness.
Anything that requires more than a cursory manipulation is best done with Photoshop, along with absolutely anything that requires color manipulation, 16 bit depth, or CMYK, no matter how simple.
I'm not into illustration, but as far as photography goes, you can do a very decent colour management with Cinepaint. Krita is unbearably slow, and Gimp... well, it's no use shooting at ambulances.
It's helpful when dealing with serious fonts that come in several subtle variants (like bold oldstyle nums) to reduce the included fonts count. Scribus is not a word processor. The adobe counterpart is no better in this light, as far as I can tell, because I had a helluva hard time dealing with a print shop that insisted on re-creating in InDesign a rough I submitted them in pdf. I had to dig the F* manual on internet to teach the typographer how to switch some caps into the alternate glyph of the face.
what a stupid idea really. there's one thing called "parenting" that do it as well. it's simple : you take the little brat with you outside, to show him the world. if you can't, i'm sure there are holidays camps or scouts or whatever goes hiking away from wifi that just do as well.
Re:In laptops, 1024x600 is low-end
on
KDE 4.3 Released
·
· Score: 1
In laptops, 1024x600 (9"/10") is low-end, and a few bargain-basement models have 800x480 (7").
1024x600 and less are netbooks and they ship with icewm (or after some tweaks kde 3.5) for a reason ; my asus 900A now runs WindowMaker for that same reason.
... but I was speaking of LCD (it should work better with caps), because while you and I made the wise choice to stay crt when they became affordable in big sizes, all the people I knew were lured toward the newer LCDs coming in miserable sizes unheard of for years (like 15") ; I even remember LCDs being given marketing "size equivalent" to crts, to try (and most of the time succeed in) luring the buyer into spending more on a tiny LCD than on bigger and better crt.
Conclusion : the general public was long accustomed to tiny screens and WMs had to take that in account.
Slashdot Mirror
never since Windows XP have I had any hardware that didn't work out of the box, and most of the time, even before installing the supplied drivers, Windows would have a go at using some kind of default one, no matter how crippled.
Ok, let's go anecdotal ; my mother just replaced her old XP Dell by a shiny new Vista computer. But her USB HP-made flatbed scanner had to be replaced, because there's no drivers for it in vista, and HP didn't bothered to update them. The irony ? Her new quad-core everything but the kitchen sink computer is a HP.
Now, for your tuner problem, it's more than likely that the driver is there and OK, but your tuner needs a firmware that can't be bundled for legal reasons in your distro. How to put the correct fw in the right place is left as an exercise to your imagination.
On a last note, while windows compatibility is pretty much taken for granted when shopping for hardware, it's not that much evident that pieces are not labeled with the necessary windows versions required to support them. Mess with it, and you're out of luck. It's exactly the same for linux, you have to be careful before buying, or to put it in latin, caveat emptor. All the HW I bought since 1998 was known to work with linux beforehand, and most of it still runs flawlessly.
This answer originating from the next cubicle of the same open space, I guess.
I somehow looked at the project around ver. 0.1.x, but it wasn't even remotely usable then. Your post made me go back and d/l the qemu image (ReactOS 3.10) ready for use. Everybody should give it a try, it's only 34 Mb, qemu for windows included. It works like a breeze on my fedora 10 (under qemu-kvm), with absolutely no config tweaking, and it seems quite stable so far. I'm going to follow this project much more closely from now on, the improvements just have been massive. TY for the tip.
Obviously straight from the hors^W MS marketing department mouth.
... but when I bought my computer, I asked for vista to be removed and the price refunded. Hoped from shop to shop until I found one that agreed (in fact I was ready for a trek, but the 3rd shop in the street was the good one). He got the deal, and I bought the refund worth of RAM to top of the computer capacity. I was pleased, and so was the seller.
My laptop is an Asus eeepc 900A linux 'edition'. Again, I carefully reviewed the options before buying.
Speak with your wallet.
Going along with the article, the first person perspective brings the dilemma of the "silent hero", because that hero is supposed to be you. However, you rarely get to interact with anything beyond acknowledging an NPC's request to retrieve something for them. You don't get to forge a bond with any of the characters based on your personality traits.
Never played DN3D, did you ? I guarantee you're ready to kick ass and chew bubble gum after you leave the game, however well bred and polite you may be before.
Finally, I'd note that if you really don't want a Bayer-pattern sensor, you can get a Fuji camera. They use a type of sensor originally developed by Foveon (which Fuji since bought out) that has individual sensor elements "stacked", to there's a red, green, and blue element at each pixel location. While Fuji's cameras are perfectly good and work quite nicely in general, they're not really a whole lot better than most others...
Foevon sensors are not what I'm looking for. While I do understand they are marginally better suited to b&w photography on principle, the reality of the technology is they have at the moment way less physical pixels than 'classic' sensors, between a 3rd and a 4th if I'm not mistaken. Fuji pulls a marketing stunt by selling 'equivalent' figures based on extrapolation to bayer-layered sensors, which is somewhat close to reality in colour photography, but is meaningless in a b&w context ; you can achieve the same result as the foevon with a bayer-sensor by reducing the resolution to a 4th of the total pixel count, and extracting and averaging 2 by 2 the green pixels to compute only luminance based on those "super pixels". It does look good on screen, but it's kind of small and no more suitable to large prints.
Vade retro, Satanas ! I'm currently fighting the impulse to buy it and you just ruined my efforts of the week.
If you look at the picasa album of pictures taken with the Frankencamera (linked from TFA through their homepage), you'll notice an incredible amount of chroma noise and banding. The caption ("can you PS camera do better ?") is ridiculous. I would be shamed if my PS produced that crap.
You missed the detail that those programs are running real-time inside the camera, building panorama's and correcting dynamic range and removing noise, while taking the picture?
Face it : this contraption still delivers phone grade pictures (a bit better because canon's lenses aren't bad) but is boxed like a field camera, and host a computer that has a 10th of the computing power you can put under your desk. At least, I would have expected a full frame 24x36 CCD to match the lens. Now, seeing most everybody is happy with the crapastic in-camera jpegs produced by consumer-grade cameras (and, that is, those 'everybody' are already enough interested in pictures to actually own a digital dedicated camera, many more people are perfectly happy with their phone !), I can't fathom whom might wish to let this thing handle their HDR process. Those already in the know can wait to better post-process at home a shooting session. Others simply won't care nor see the difference.
Moreover, my cameras already tries to remove noise, and my phone (Nokia Navigator) can stitch panos. And both are so bad at it that I never take pictures with my phone (except from the Big SUV parked next to my car I wish to remember the licence plate in case I get a blow in my door's paint), and set NR to low on my camera to use better suited algorithms on my computer.
Agreed. I've got a lx2 myself (for the 60mm tele end of the lx3 zoom is IMO too short for close-up portraits), and I'm currently drooling over the Olympus ep-1. But as soon as light dims, I get back to my dslr. The smallish sensor of the pany gets very noisy real quick.
What I would really love is a CCD or CMOS without bayer color matrix nor antialiasing filter. This camera would be limited to black and white photography (or studio pack shot with 3 exposures behind R,B,G filters), but I expect the result would be outstanding. At the moment, we're fighting with slightly fuzzy shots (thanks to the AA filter) to recreate true pixels intensity after they've been distorted by various digital process. There's been one digital b&w camera, made by kodak. It's said the production number was 80, but from reviews, images were already gorgeous with 6 Mpix 10 years ago. Today, bayer matrix and AA filters are glued on the chip in the manufacturing process, and it's impossible to get rid of it afterwards.
Most settlers and pilgrims went away after an initial explorer returned to tell how beautiful the trip was and how nice the people there were. Or in another scenario, people moved because they would be in harm way should they stay (invasions, wars, etc.) No one in his right mind just went away without the slightest hope of coming back, even if many knew the journey was dangerous.
Turing desserves an apology because any security lookup done during wartime certainly had already uncovered this aspect of his personality, but the brits were more than happy to turn a blind eye on it in spite of the already existing laws.
The later prosecution was application of a double standard in this regard.
Oddly enough, there apparently are some people that still use them. The train ticketing, phone book, and a number of other services are still up and in use.
Not odd at all considering the various threats of Internet, from spam to virus, credit cards frauds, DoS etc. Minitel pretty much insure that whomever you phone is legit provided you don't misstype the phone number. It's a very helpful and desirable feature for some sensitive businesses (chemist ordering prescription drugs, etc.)
And thanks to being a passive terminal, Minitel is immune to virus and trojans by nature. Being so simple, there are no bugs either I'm aware off. And being text only makes for a great bonus to blinds who can plug whatever Braille device they want to use it.
sorry, I didn't checked my maths. In 16 bits, 126 and 127 are 32256 and 32512 respectively, and the number of steps inbetween is 256.
My thinking was about fine art or wedding photographers. More generally, it applies when your workflow starts with a RAW picture and you need to squeeze as much informations as possible out of it. If you work on in-camera made jpegs, there's no real need to turn to 16 bits because any postprocessing will be bad anyway. But as soon as you oversize, rotate, correct perspectives, adjust tones, in fact every operation that creates new pixels via an interpolating algorithm based on neighbours intensity, 16 bits is really the way to get a whole range of smooth in-between colours for smooth shades instead of noisy chroma blobs. The more operations you do on your pictures, the more you add rounding errors and the more likely you'll get patches of uniform (and bad !) colours right in the middle of what you expect to be a clean shade (think skin, for instance). Suppose for instance you create a pixel between one which value on one channel (8 bits) is 126 and another which value is 127 ; your algorithm can choose either value, and this new pixel won't be further discriminated from its neighbour in any further calculation. Whereas in 16 bits, your 2 points would have 2016 and 2032 as intensity value, leaving a range of 16 intensity steps for creating a pixel in-between, this pixel being distinct from its neighbours in any further calculation.
And I have found NO linux software that can apply a profile to a print. So I need a virtual XP just to create printer profiles and print !
You didn't looked hard enough. Cinepaint does this perfectly, and even let you smelt any RGB profile into CMYK or XYZ profiles while showing you the relative gamut transformation in OpenGL 3D no less. Or you can display your image in proof mode, etc.
It's not a question of eye, it's a question of maths. As long as digital photography will create colours by interpolation between RGBG values of a matrix, you'll need empty 'holes' to re-create colours and avoid sampling artefacts. Applying filters (sharpening, border sharpening, USM etc.) also needs room to do their magic, otherwise you're accumulating errors that ultimately show as noise. When your postprocessing is done, then you can downgrade to 8 bits / channel without too much loss for printing or display.
Believe it or else, there's a real difference between a 8 bits and a 16 bits workflow; you'd need to try for yourself to be convinced, but for a pro, it's the difference between work or joblessness.
Anything that requires more than a cursory manipulation is best done with Photoshop, along with absolutely anything that requires color manipulation, 16 bit depth, or CMYK, no matter how simple.
I'm not into illustration, but as far as photography goes, you can do a very decent colour management with Cinepaint. Krita is unbearably slow, and Gimp... well, it's no use shooting at ambulances.
It's helpful when dealing with serious fonts that come in several subtle variants (like bold oldstyle nums) to reduce the included fonts count. Scribus is not a word processor. The adobe counterpart is no better in this light, as far as I can tell, because I had a helluva hard time dealing with a print shop that insisted on re-creating in InDesign a rough I submitted them in pdf. I had to dig the F* manual on internet to teach the typographer how to switch some caps into the alternate glyph of the face.
Speaking of "gem in the rough", dcraw is an extremely good RAW images to tiff converter. Command line, but there are plenty graphic wrappers for it.
what a stupid idea really. there's one thing called "parenting" that do it as well. it's simple : you take the little brat with you outside, to show him the world. if you can't, i'm sure there are holidays camps or scouts or whatever goes hiking away from wifi that just do as well.
In laptops, 1024x600 (9"/10") is low-end, and a few bargain-basement models have 800x480 (7").
1024x600 and less are netbooks and they ship with icewm (or after some tweaks kde 3.5) for a reason ; my asus 900A now runs WindowMaker for that same reason.
... but I was speaking of LCD (it should work better with caps), because while you and I made the wise choice to stay crt when they became affordable in big sizes, all the people I knew were lured toward the newer LCDs coming in miserable sizes unheard of for years (like 15") ; I even remember LCDs being given marketing "size equivalent" to crts, to try (and most of the time succeed in) luring the buyer into spending more on a tiny LCD than on bigger and better crt.
Conclusion : the general public was long accustomed to tiny screens and WMs had to take that in account.