A quick scan of company servers would be enough to see if they're running Linux or not. And even if you do change the servers to say they are Win 2003 or something, what about social engineering? Calling up the company, saying you're MS tech support, and you found a problem with the company's web servers. "But we run Linux." Gotcha!
Anybody who gets an unsolicited call from MS tech support and doesn't have their crap detectors go off full blast deserves what they've got coming.
Any security-conscious operation isn't going to reveal what they're running anyway, especially not to a complete stranger. Requests for HTTP service should be met with something that isn't a general-purpose computer, and the web server should identify itself as SomeWebServer 3.14 on SomeOperatingSystem. There's really no need for the browser, hackers, Billy Gates, SCO or Netcraft to know.
Actually, I think Google and the other search engines will solve this problem for us. There are so many broken links out there that VeriSign's web servers are going to get a steady stream of queries as the bad links are followed into their ads. Even if they put a robots.txt on their server or use a redirect to another page, they're still going to get queried at least once. I notice that.cx redirects to a URL containing the domain name and a unique ID number, so even plain redirects aren't going to cut it. The risk could be reduced by only returning an address for things that are truly typos, but I don't imagine VeriSign's going to want to reduce their eyeball count by doing that.
God help 'em if people started creating pages that genereted a few hundred thousand bad links...
Side note: I have a friend inside VeriSign who reports that people in the the non-NSI part of the company don't think very highly of the NSI part. (If this were Fark, I'd put an [OBVIOUS] tag on that.)
Diebold is a private company. Diebold makes a closed-source product, much the same as Microsoft. How would you require, by law, that this company expose its source code to the world?
How? You nip the proprietary part of the software process in the bud from day one by separating it from hardware. Hire the hardware vendor or a third party to do the development as a service, with the finished product belonging to whatever board of elections let the contract. They're then free to publish it for review. The Diebolds of the world will still have made the same money because they'd build the same amount of revenue into the development contract.
I'm curious about this. Do you think it's due to the depth of the photosites or lenses that have classic flatness of field problems and edge/corner softness? Have you seen any pictures taken on film with the same lenses?
That's a good question. One of the samples I saw was taken with an unfamiliar mid-range zoom, so I can't entirely discount the optics from being part of the problem. The other was taken with a prime that I've shot film with, and the edges of the digital image weren't consistent with what I'd seen in film (apertures being close or the same). One other factor not taken into account is that neither the Kodak or Cannon full-frame sensors are CCDs, and not having studied their operation, it's hard for me to say how they'll behave around the edges. But they're new sensors, and as happened with CCDs, there'll be improvements.
I agree with the other poster about the irritating multiplying effect of smaller sensors.
My experience with the multiplier factor was comparable to that of getting a new car: it felt different from the old one, but after awhile I got used to it. The other poster's point about losing a lot of picture angle on wides is taken, though. I don't shoot wide much, so for me it's not as much of an annoyance. The other end of the spectrum is that the sports and nature shooters are now effectively getting 600 performance out of their 400s, and some of them like that.
Maybe I'm delusional, but I think I'll do OK with corner-to-corner sharpness with a 24mm and 35mm non-zoom lens vs. some freakish 15-24mm zoom used with an APS-sized sensor DSLR.
It sounds like you're trying to compare shooting full-frame film with a prime to shooting smaller-frame digital with a wide zoom. That's apples-to-oranges: good primes will outdo good zooms any day of the week on any camera. The change over to the smaller sensor is really just a small change in mindset about how your glass will behave. In exchange for the loss of picture angle, the smaller sensor grabs the image from the center, which happens to be the best-performing part of most lenses. Until they get the full-frame sensors right, I can live with that.
...Probably because 11g was ratified as a standard too recently for the chipsets to be available for inclusion in the product. 11b also has a much larger installed base.
It doesn't matter; the wireless unit is an add-on, and if there's demand for an 11g version, they'll do it.
I don't know the details of this, but it has something to do with the distance from the lens to the CCD, and the angle the light hits the CCD. With wide angle lenses in particular, you will get colour shifts - mostly around the edges and the corners. Does somebody here know more about this problem?
The center of the film plane is, in most cases, lined up with the focal point, and the light hitting it arrives at a 90-degree angle. As you get further from the center, the angle decreases.
Each pixel on a CCD is made up of a little well called a photosite that collects photons and reports back how many it caught as the brightness level for that pixel. Like a well you'd dig in the ground, photosites have some depth to them, and as such the best way for light to reach the photosite is straight on, or at a 90-degree angle. Part of the filter that goes over a CCD includes an array of microlenses, each of which takes light from a certain range of angles and realigns it so the most photons fall into the photosite. Without the corrective lenses, different wavelengths of light behave differently when they hit the same photosite at the same angle, which can produce all sorts of undesirable artifacts that are exaggerated as the angle decreases.
The corrective lenses reduce this problem to a certain degree, but like any other bit of optics, there are limits to what they can bend back to the 90-degree angle that makes a CCD photosite work best. That means that if you build a sensor with correction for shorter focal lengths, lenses with longer ones will display the same behavior the correction was supposed to fix. The way around this is to use a sensor small enough that the corrective lenses will work equally well when the focal point is 17mm away from the film plane as when it's 400mm away. That, combined with the yield problems in fabricating full-frame sensors, is the reason you haven't seen many of them yet. The technology's just not there yet.
This isn't directed at you, but I'm still stumped over why people get their undies in a bunch over the smaller sensors. I've been given images from Cannon and Kodak's full-frame offerings, and even with the large pixel count, I'd still have been forced to crop them down to what a smaller sensor would have produced to get uniformly-sharp results. I've shot close to 5,000 frames in the last year, 90% of which came from a DSLR with a smaller-than-35mm-sized frame (and I'm not going to say which it is, because frankly it's immaterial). Not once have I found it a handicap: I still look throught the viewfinder, compose the shot and release the shutter just the same as I've always done with my film bodies.
Yeah, and the difference is that there's no R&D costs with gold, nor does the FDA force the miners to put it through an expensive approval process before it can be molded into your favorite bling.
...Which doesn't change the fact that there are stations with cheap-assed GMs that are still using 20-year-old carts because the tape hasn't broken yet.:-)
Come on, what kind of confused reasoning is that? Digital cameras have shorter focal lengths and larger DOF because digital sensors are smaller than 35mm film.
There's nothing confused about it if you go back and re-read my post. I said consumer digital cameras have short focal lengths, which makes sense because the optics are designed to illuminate the entire imager. On most consumer models, that's a whopping 4x5mm.
Digitals based on SLR bodies, on the other hand, have just as much (or as little) DOF as their film counterparts because the lens geometry is still configured for a full 24x36mm frame. The fact that a 50mm lens behaves like a 75mm lens because of the multiplier does nothing to change the fact that the focal length is still 50mm. The smaller imager just captures something smaller than 24x36mm.
My point here is that a blanket statement declaring that (all) digitals have large depth of field is nowhere near accurate and certainly won't be when a consumer digital point-and-shoot with a 24x36mm imager is available for $200.
I agree that excess DOF is not particularly nice for certain shooting situations. However, for most people, it's a good thing because it makes "getting the shot" much easier.
If you aske me, near-infinite DOF isn't particularly nice in most shooting situations. It does help people using fixed-focus cameras get their subjects in focus more often, but potentially great snapshots get ruined when a sharply-focused background distracts the viewer's eye from the equally-sharply-focused subject.
Actually, I find 35mm DOF to be neither here nor there: it doesn't give satisfactory separation of subject and background in many situations, yet it is small enough to punish inaccurate focussing ruthlessly.
I don't buy that, because the math doesn't bear it out.
Take your garden-variety 50mm f/2 lens on a 35mm body focused at 10 feet. Stopped down to f/22, you get 23.34 feet of DOF to play with, because everything from 6.03 to 29.37 feet will be in focus. (CoC is 0.025, by the way.) Open it up to f/16 and it's 12.46 feet, at f/8 it's 5.09 feet, at f/4 it's 2.43 feet and at f/2 it's 1.21 feet. That's enough range for a whole lot of situations with a single prime lens. And at worst, there's a 6% margin for focusing error. If you're not trying to manually-focus a short-throw autofocus lens, that's not a hard margin to hit.
I think MF and digital each represent better tradeoffs: with MF, I get meaningful separation of subject and background, and with digital, I generally don't have to worry about it at all.
I'm not even going to ask what you think happens when you put a digital back on a medium-format body.:-)
Large DOF has nothing to do with digital vs. analog and everything to do with the fact that the focal length in most consumer digitals is so short that the hyperfocal distance becomes very short. A random shot taken with a DC280 selected from my archive says 6.3mm at f/4, putting the HFD at 1.3 feet and everything from about eight inches to infinity in focus.
I found the excess DOF in my DC280 annoying enough that I sprung for a DSLR, and when I want that much DOF, I can have it.
If Max Kipfer does the same kind of job running OpenBand that he did while he ran Cablevision of Loudoun (bought by Adelphia a few years ago), homeowners can expect overpriced service on a poorly-maintained system.
And since nobody else will likely be able to bury anything in that neighborhood, they'll be stuck with them for POTS and Internet access. Verizon may suck at a lot of things, but they're real good at making sure you have a dial tone when you need to call 911 in the middle of the night. (Fortunately, they can't restrict DBS dishes, so at least there's an escape route for TV.)
It's not so much that BIND is bad (which it isn't), it's that there needs to be some diversity in the software used on the root servers. Otherwise, one gaping security hole discovred in BIND and exploited could render the whole lot of them useless.
And there ain't nothin' worse than trying to make a phone call and discovering that the phone book no longer works.
I'm sure that for Verizon, bandwidth is cheap. They probably have a dozen or so OC48/192/... lines running from my CO to their office downtown.
The RBOCs may have the bandwidth to play with internally, but they still have to buy transit from national ISPs, and that's neither infinite nor cheap. And besides, those big pipes cary tens of thousands of phone lines, 56Ks T1s, T3s, etc. It's not all Internet.
That is why I switched to DSL, it is a always reliable connection, and I know no one is hogging my share of the pipe.
Uh, no, it means nobody's hogging the pipe between you and the DSLAM. Don't think for a minute that because you've got a 768K DSL line the phone company has 768K of its pipe to the rest of the world earmarked for your use only. ISPs work on oversubscription, which is why your service is so cheap.
You'd be surprised at how many devices have embedded firmware.... toys, microwaves, cd players, radios, etc.,... ethernet palm pilot cradle I worked on a while back... barcode scanners... all manner of microcontrollers/firmware on computer peripherals (PCI cards, printers, scanners, etc.), cordless phones (do you think you get the caller ID message without firmware?) etc. etc.
And don't think for a minute that a single one of those devices runs anything that even remotely resembles a general-purpose operating system.
Slashdot Mirror
That's gonna make it awfully difficult for all but the most hearty-bladdered pilots to do long flights like LAX to SYD.
"Uh, roger United 345 heavy, understand your condition. Porta-potty will be waiting at the gate."
Anybody who gets an unsolicited call from MS tech support and doesn't have their crap detectors go off full blast deserves what they've got coming.
Any security-conscious operation isn't going to reveal what they're running anyway, especially not to a complete stranger. Requests for HTTP service should be met with something that isn't a general-purpose computer, and the web server should identify itself as SomeWebServer 3.14 on SomeOperatingSystem. There's really no need for the browser, hackers, Billy Gates, SCO or Netcraft to know.
...skid marks. Plain and simple.
(Sorry. Somebody had to do it.)
God help 'em if people started creating pages that genereted a few hundred thousand bad links...
Side note: I have a friend inside VeriSign who reports that people in the the non-NSI part of the company don't think very highly of the NSI part. (If this were Fark, I'd put an [OBVIOUS] tag on that.)
...Cleveland becomes the world's spam capital.
How? You nip the proprietary part of the software process in the bud from day one by separating it from hardware. Hire the hardware vendor or a third party to do the development as a service, with the finished product belonging to whatever board of elections let the contract. They're then free to publish it for review. The Diebolds of the world will still have made the same money because they'd build the same amount of revenue into the development contract.
That's a good question. One of the samples I saw was taken with an unfamiliar mid-range zoom, so I can't entirely discount the optics from being part of the problem. The other was taken with a prime that I've shot film with, and the edges of the digital image weren't consistent with what I'd seen in film (apertures being close or the same). One other factor not taken into account is that neither the Kodak or Cannon full-frame sensors are CCDs, and not having studied their operation, it's hard for me to say how they'll behave around the edges. But they're new sensors, and as happened with CCDs, there'll be improvements.
I agree with the other poster about the irritating multiplying effect of smaller sensors.
My experience with the multiplier factor was comparable to that of getting a new car: it felt different from the old one, but after awhile I got used to it. The other poster's point about losing a lot of picture angle on wides is taken, though. I don't shoot wide much, so for me it's not as much of an annoyance. The other end of the spectrum is that the sports and nature shooters are now effectively getting 600 performance out of their 400s, and some of them like that.
Maybe I'm delusional, but I think I'll do OK with corner-to-corner sharpness with a 24mm and 35mm non-zoom lens vs. some freakish 15-24mm zoom used with an APS-sized sensor DSLR.
It sounds like you're trying to compare shooting full-frame film with a prime to shooting smaller-frame digital with a wide zoom. That's apples-to-oranges: good primes will outdo good zooms any day of the week on any camera. The change over to the smaller sensor is really just a small change in mindset about how your glass will behave. In exchange for the loss of picture angle, the smaller sensor grabs the image from the center, which happens to be the best-performing part of most lenses. Until they get the full-frame sensors right, I can live with that.
...Probably because 11g was ratified as a standard too recently for the chipsets to be available for inclusion in the product. 11b also has a much larger installed base.
It doesn't matter; the wireless unit is an add-on, and if there's demand for an 11g version, they'll do it.
The center of the film plane is, in most cases, lined up with the focal point, and the light hitting it arrives at a 90-degree angle. As you get further from the center, the angle decreases.
Each pixel on a CCD is made up of a little well called a photosite that collects photons and reports back how many it caught as the brightness level for that pixel. Like a well you'd dig in the ground, photosites have some depth to them, and as such the best way for light to reach the photosite is straight on, or at a 90-degree angle. Part of the filter that goes over a CCD includes an array of microlenses, each of which takes light from a certain range of angles and realigns it so the most photons fall into the photosite. Without the corrective lenses, different wavelengths of light behave differently when they hit the same photosite at the same angle, which can produce all sorts of undesirable artifacts that are exaggerated as the angle decreases.
The corrective lenses reduce this problem to a certain degree, but like any other bit of optics, there are limits to what they can bend back to the 90-degree angle that makes a CCD photosite work best. That means that if you build a sensor with correction for shorter focal lengths, lenses with longer ones will display the same behavior the correction was supposed to fix. The way around this is to use a sensor small enough that the corrective lenses will work equally well when the focal point is 17mm away from the film plane as when it's 400mm away. That, combined with the yield problems in fabricating full-frame sensors, is the reason you haven't seen many of them yet. The technology's just not there yet.
This isn't directed at you, but I'm still stumped over why people get their undies in a bunch over the smaller sensors. I've been given images from Cannon and Kodak's full-frame offerings, and even with the large pixel count, I'd still have been forced to crop them down to what a smaller sensor would have produced to get uniformly-sharp results. I've shot close to 5,000 frames in the last year, 90% of which came from a DSLR with a smaller-than-35mm-sized frame (and I'm not going to say which it is, because frankly it's immaterial). Not once have I found it a handicap: I still look throught the viewfinder, compose the shot and release the shutter just the same as I've always done with my film bodies.
Since you're obviously in possession of an early production D2H, how about posting a couple of side-by-side samples to compare it with your D1S?
Inquiring eyeballs want to know...
Yeah, and the difference is that there's no R&D costs with gold, nor does the FDA force the miners to put it through an expensive approval process before it can be molded into your favorite bling.
A whole Steppenwolf^H^H^H^H^H^H^H^H^H^H^HBeowulf clutster of these things!
...Which doesn't change the fact that there are stations with cheap-assed GMs that are still using 20-year-old carts because the tape hasn't broken yet. :-)
Never mind that ESPN and one of the TV networks (ABC) are both Disney outlets...
There's nothing confused about it if you go back and re-read my post. I said consumer digital cameras have short focal lengths, which makes sense because the optics are designed to illuminate the entire imager. On most consumer models, that's a whopping 4x5mm.
Digitals based on SLR bodies, on the other hand, have just as much (or as little) DOF as their film counterparts because the lens geometry is still configured for a full 24x36mm frame. The fact that a 50mm lens behaves like a 75mm lens because of the multiplier does nothing to change the fact that the focal length is still 50mm. The smaller imager just captures something smaller than 24x36mm.
My point here is that a blanket statement declaring that (all) digitals have large depth of field is nowhere near accurate and certainly won't be when a consumer digital point-and-shoot with a 24x36mm imager is available for $200.
I agree that excess DOF is not particularly nice for certain shooting situations. However, for most people, it's a good thing because it makes "getting the shot" much easier.
If you aske me, near-infinite DOF isn't particularly nice in most shooting situations. It does help people using fixed-focus cameras get their subjects in focus more often, but potentially great snapshots get ruined when a sharply-focused background distracts the viewer's eye from the equally-sharply-focused subject.
Actually, I find 35mm DOF to be neither here nor there: it doesn't give satisfactory separation of subject and background in many situations, yet it is small enough to punish inaccurate focussing ruthlessly.
I don't buy that, because the math doesn't bear it out. Take your garden-variety 50mm f/2 lens on a 35mm body focused at 10 feet. Stopped down to f/22, you get 23.34 feet of DOF to play with, because everything from 6.03 to 29.37 feet will be in focus. (CoC is 0.025, by the way.) Open it up to f/16 and it's 12.46 feet, at f/8 it's 5.09 feet, at f/4 it's 2.43 feet and at f/2 it's 1.21 feet. That's enough range for a whole lot of situations with a single prime lens. And at worst, there's a 6% margin for focusing error. If you're not trying to manually-focus a short-throw autofocus lens, that's not a hard margin to hit.
I think MF and digital each represent better tradeoffs: with MF, I get meaningful separation of subject and background, and with digital, I generally don't have to worry about it at all.
I'm not even going to ask what you think happens when you put a digital back on a medium-format body. :-)
Large DOF has nothing to do with digital vs. analog and everything to do with the fact that the focal length in most consumer digitals is so short that the hyperfocal distance becomes very short. A random shot taken with a DC280 selected from my archive says 6.3mm at f/4, putting the HFD at 1.3 feet and everything from about eight inches to infinity in focus.
I found the excess DOF in my DC280 annoying enough that I sprung for a DSLR, and when I want that much DOF, I can have it.
Never heard of 'em. Must be some fly-by-night operation. :-)
If Max Kipfer does the same kind of job running OpenBand that he did while he ran Cablevision of Loudoun (bought by Adelphia a few years ago), homeowners can expect overpriced service on a poorly-maintained system.
And since nobody else will likely be able to bury anything in that neighborhood, they'll be stuck with them for POTS and Internet access. Verizon may suck at a lot of things, but they're real good at making sure you have a dial tone when you need to call 911 in the middle of the night. (Fortunately, they can't restrict DBS dishes, so at least there's an escape route for TV.)
It's not so much that BIND is bad (which it isn't), it's that there needs to be some diversity in the software used on the root servers. Otherwise, one gaping security hole discovred in BIND and exploited could render the whole lot of them useless.
And there ain't nothin' worse than trying to make a phone call and discovering that the phone book no longer works.
I'm sure that for Verizon, bandwidth is cheap. They probably have a dozen or so OC48/192/... lines running from my CO to their office downtown.
The RBOCs may have the bandwidth to play with internally, but they still have to buy transit from national ISPs, and that's neither infinite nor cheap. And besides, those big pipes cary tens of thousands of phone lines, 56Ks T1s, T3s, etc. It's not all Internet.
That is why I switched to DSL, it is a always reliable connection, and I know no one is hogging my share of the pipe.
Uh, no, it means nobody's hogging the pipe between you and the DSLAM. Don't think for a minute that because you've got a 768K DSL line the phone company has 768K of its pipe to the rest of the world earmarked for your use only. ISPs work on oversubscription, which is why your service is so cheap.
Nothing in the GPL (for those packages that are GPLd) requires that download information be included in press releases.
You'd be surprised at how many devices have embedded firmware. ... toys, microwaves, cd players, radios, etc., ... ethernet palm pilot cradle I worked on a while back ... barcode scanners ... all manner of microcontrollers/firmware on computer peripherals (PCI cards, printers, scanners, etc.), cordless phones (do you think you get the caller ID message without firmware?) etc. etc.
And don't think for a minute that a single one of those devices runs anything that even remotely resembles a general-purpose operating system.