Having been in the exact same situation I've tried all sorts of different solutions and I'd say the best current solution is NTFS, which is out of the box natively supported on both OSX and Windows (natch) and also available R/O in the default linux kernel as well as having strong R/W support now via ntfs-3g. Of course fat32 still works just fine for this application, but it's getting a little long in the tooth as far as advanced features and modern storage needs go (c'mon what is up with those weak filesize limits)!?!? And I've had some limited success with using ext2/3 on windows and linux but found that the windows kernel driver for ext2 was not very stable in my config and the userspace tools to read/write ext3 in windows was far too kludgy for my tastes; I haven't had a chance to try ext2/3 on OSX.
Look, you are missing my whole point. IF you shoot 2gb at a time, THEN BT becomes the bottle neck and THEN you are probably not the type of person that this device would be useful to. However, many people would get plenty of use out of this technology AND many people do have unlimited 3g data AND live in areas with coverage AND don't care if their phone is transmitting for minutes-hours after they are done shooting. Just look at the popularity of camera phones... Think about the number of people who aren't shooting huge photos on DSLR cameras and want to take snapshots... I was simply pointing out that, in your previous post, you were taking the worst possible scenario and using that single handedly to try and create a false dichotomy wherein you propose that the only use you have for a camera would preclude the usefulness of a BT PAN including a camera and therefore would make it useless. However, as someone who does regularly take photographs and sends them over a 3g connection, I am saying that I (and I'm sure many many others) would certainly be a consumer of this technology.
Also the first part of your second post is going on about shooting in nature and not having any 3g coverage... Well certainly there is more 3g coverage (in my country at least) than there is area covered by open wifi access points. And if you are shooting through 2GB cards in 3 hours I'd bet dollars to donuts that you'd be best served by having multiple cards and doing a "traditional" transfer to local for postprocessing and forgoing the entire wireless transfer scheme altogether....
Your initial argument reads like one of the pro hybrid car arguments where they base the maximum range of a car on the presumption that it starts with the batteries fully charged. It is a true statement, but it is being made about a situation that is generally atypical. Using that as the crux of your argument is the straw man I was calling you out on...
This next part might come out like gobblydegook since i'm just doing it as I'm going along...
So lets split the difference and say that the average digital camera user falls somewhere between your example and what I say my typical usage is. Instead of 3.5MB per photo of your DSLR or my 1.25MB point and shoots we'll call an average picture 2.4MB. And your 3.5MB pictures fill a 2GB card in 3 hours, lets call that 195 photos/hr while my highest rate, while on vacation and taking waaay more than more normal rate, I clocked something like 30/hr at most. That gives us an average of 112 photos at 2.5 MB each per hour. Now you gave an estimated speed of 3.5MB per minute for BT+3g which is 210 of your DSLR photos per hour or something like 9.7 hours to upload 3 hours worth of photography, or 6.7 extra hours of uploading after shooting. I OTOH would take 54 hours of shooting at my stated rate/resolution to shoot 2gb and would have been uploading at faster than realtime, my card would never hit capacity as long as I had a signal and/or enough memory on my phone to cache the data (coincidentally my 3g handset has 2GB of addon memory and 64MB internal). So our theoretical "average" photog would take 4.6 hours to shoot 2gb worth of photos while uploading at about 70% of realtime incurring a 30% time to upload penalty...
Well, not sure if that math holds up... maybe i'll make a spreadysheet or something... but for now I'm gonna go enjoy dinner;)
You are making a straw man argument. If you are using a PAN which includes a 3g handset then you are bringing the network with you. As such, photos could be uploaded on the fly. So unless you shoot in contnuous mode and hold down the shutter button to intentionally fill the 2gb of storage you wouldn't be waiting "9 hours" to transfer your photos, the photos could instead be transferred seamlessly in the background.
If you are able to get a sustained 64kBps all the way from your camera to your webhost you will not be posting in 100% realtime, although you'd be damn near it and I'd also say it is unlikely that most people would be shooting over 2GB of pictures in any given 9 hours stretch. The bottle neck in nearly all cases will be the photographer, not the transfer speed.
I'd much prefer an SD card with a bluetooth adaptor built in that could leverage the 3G wireless internet connection which is the true core of the PAN (Personal Area Network) that is always touted as being the logical goal of the bluetooth architecture. I mean really, BT chipsets are far more optimized for power than wifi and comes with far fewer limitation as to the connections it can make. Let the devices choose the path of least resistance to the internet, be it tunnel over a phone, pda, laptop, or whatever the marketplace has in store next.
honestly I think that the working group that came up with BT designed it for exactly this sort of purpose. It'd be stupid not to also use this type of integration between PAN components to further enhance the meta data richness of the content created by the camera. GPS, PDA, camera, 3gphone, and headset sounds like a pretty good recipe for being your own gargoyle. I for one wouldn't mind being able to publish video, photo, sound, and location data at a moment's notice directly to the internet. If we are bound to live in surveillance state I'd sure like to get a good grip on the technology before Big Brother does.
Sure, just show me how I'm going to get a 4 hour repair/replace onsite technician response contract for this $3000 beige box NAS device. I'll buy 12 of them. Until then I'm gonna stick to my Netapps;)
WTF? Are you 12 years old now or something? 10GB was orders of magnitude larger than anything I could possibly hope for when I was "9 or 10" years old.
1989 Nissan 240SX, which is one of the most aerodynamic vehicles on the planet (0.26cD) cD isn't even close to the whole story on aerodynamics, it is a totally meaningless measure unless you also know the frontal area of the car as well as the rolling resistance. Even the slipperiest of vehicles is going to have significant aerodynamic resistance if it present a large area; likewise any motorcycle is going to have a cD much higher than almost any passenger car however it will also have about 1/8 the frontal surface...
Now you know!
And knowing is some fractional portion of some sort of battle, or so I've been told.
I'm just pointing out that your comment was a pure speculation on what people would or wouldn't do with no factual basis to point to. I personally have seen systems which do appear to be working. You on the other hand just throw the suggestion out the window as something that people are obviously not going to participate in. Support and proof are not interchangeable words... I feel no burden to prove my experience to you however I can share that experience as support for my claims. You have done neither...
Not sure where you are located, but here (Oakland, CA) there are CFL recycling bins at Ikea, Home Depot, the Waste Transfer Station, and many other hardware stores and other locations where you would purchase CFls... Usually the bins have quite a few bulbs in them which would seem to indicate that, contrary to your unsupported claim, people will and do recycle and properly dispose of items which may not be covered in their normal curbside pickup.
Sun Microsystems Inc. SunOS 5.10 Generic January 2005 # psrinfo 0 on-line since 04/14/2007 21:27:04 1 on-line since 04/14/2007 21:27:06 2 on-line since 04/14/2007 21:27:06 3 on-line since 04/14/2007 21:27:06 4 on-line since 04/14/2007 21:27:06 5 on-line since 04/14/2007 21:27:06 6 on-line since 04/14/2007 21:27:06 7 on-line since 04/14/2007 21:27:06 8 on-line since 04/14/2007 21:27:06 9 on-line since 04/14/2007 21:27:06 10 on-line since 04/14/2007 21:27:06 11 on-line since 04/14/2007 21:27:06 12 on-line since 04/14/2007 21:27:06 13 on-line since 04/14/2007 21:27:06 14 on-line since 04/14/2007 21:27:06 15 on-line since 04/14/2007 21:27:06 16 on-line since 04/14/2007 21:27:06 17 on-line since 04/14/2007 21:27:06 18 on-line since 04/14/2007 21:27:06 19 on-line since 04/14/2007 21:27:06 20 on-line since 04/14/2007 21:27:06 21 on-line since 04/14/2007 21:27:06 22 on-line since 04/14/2007 21:27:06 23 on-line since 04/14/2007 21:27:06 # uname -a SunOS [redacted] 5.10 Generic_118833-24 sun4v sparc SUNW,Sun-Fire-T1000
There are people who will build "it, whatever it is" however I think you are drastically understating the difficulty it is to get a board with the complexity that a modern motherboard requires actually made. Hobby type board manufacturers rely on the fact the most hobby electronics use 1or 2 sided boards, not 6-10 layer boards with exact specs on trace length and orientation that is needed to fit and support the huge number of very high speed components that is needed for even the most basic P4 or Athlon.
OTOH if you "opensource motherboard" is an 8088 clone built with discrete components you just might be able to get that done...
Niagara is wicked fast. It works great for highly parallelized tasks, however it only has a single FPU which makes it pretty much worthless for a lot of the things that you'd want to use a high-end server for. 24 threads and only one FPU does not make for fast ops at all tasks....
It does have 300% more blue LEDs than the last gen sun hardware though;)
I recall reading an article in a magazine a few years back which was talking about the research being done into using Perflurocarbon based blood replacement. My memory of the article is hazy, but iirc there was some mention of some human trials being done in the poorer parts of Africa where the demand for blood cannot be met due to the lack of refrigeration and the abundance of disease in the population.
You know what would be awesome? If people read the articles that they linked too...
Open collector BJTs (which are usually NPN) exhibits faster fall time and greater current handling capabilities than FET, but have other problems. One of them is that they consume a lot of power.
Possible problems
As mentioned above, open-collector devices can handle more current, but they also have higher current minimums for correct operation. Even in the "off" state, open-collectors have some few nanoamps of leakage current (the exact amount varies with temperature.)
Besides, isn't patent licensing part of the reason nVidia and Ati won't release fully OSS drivers? I believe Intel has patents on certain memory bus related technologies which are used by both nVidia and Ati.
AFAICT (IANAL) this is exactly correct. There needs to be some sort of tipping point reached wherein the projected cost of IP based liability issues is outweighed by potential revenue to be gained through sales based on open drivers.
I personally am not too opposed to binary globs with an open ABI; unfortunately I, as well as most linux users (and potential ATI/NVidia customers) have very little say as to what gets accepted into the mainstream kernel.
I would tend to disagree, as would nearly any police or military armed combat instructor. Practical studies have shown time and again that an assaliant within 20 feet of his target and armed with an edged weapon vs a defender with a handgun will be victorious in 9 out of 10 encounters. The factors which make this possible include the fact that the agressor has the dual advantages of stealth and suprise. There are numerous police training videos where highly trained an combat ready armed police have deadly blows delivered with a training blade before they can even have their service weapon out of the holster. The ones who do manage to get the gun from their holster are unable to aim and fire with even the slimest notion of accuracy or timeliness.
The situations where a gun trumps a knife are those in open areas with distances between target and shooter over 20 feet and/or the agressor is the one with the firearm and the edged weapon is used for defense.
The fact of the matter is that regardless of the weapons employed, the agressor has many many cards stacked in their favor and the defender has very little in the way of reaction time in which to change the circumstances of the situation. This training video shows the effectiveness of a knife vs. a holstered weapon at 12 feet. Clearly even a trained police officer who has fore knowledge of the impending attack cannot sufficiently defend themselves.
The Force Science Research Center, found the following while studying this exact encounter situation:
*Once he perceives a signal to do so, the AVERAGE officer requires 1.5 seconds to draw from a snapped Level II holster and fire one unsighted round at center mass. Add 1/4 of a second for firing a second round, and another 1/10 of a second for obtaining a flash sight picture for the average officer.
*The fastest officer tested required 1.31 seconds to draw from a Level II holster and get off his first unsighted round.The slowest officer tested required 2.25 seconds.
*For the average officer to draw and fire an unsighted round from a snapped Level III holster, which is becoming increasingly popular in LE because of its extra security features, takes 1.7 seconds.
*Meanwhile, the AVERAGE suspect with an edged weapon raised in the traditional "ice-pick" position can go from a dead stop to level, unobstructed surface offering good traction in 1.5-1.7 seconds.
*The "fastest, most skillful, most powerful" subject FSRC tested "easily" covered that distance in 1.27 seconds. Intense rage, high agitation and/or the influence of stimulants may even shorten that time, Lewinski observes.
*Even the slowest subject "lumbered" through this distance in just 2.5 seconds. Bottom line: Within a 21-foot perimeter, most officers dealing with most edged-weapon suspects are at a decided - perhaps fatal - disadvantage if the suspect launches a sudden charge intent on harming them. "Certainly it is not safe to have your gun in your holster at this distance," Lewinski says, and firing in hopes of stopping an activated attack within this range may well be justified.
And keep in mind, the officers in this study were both aware of the impending attack and were mentally prepared to draw down and fire on the agressor.
AND VHS has a strong history of legal precedent allowing consumer recording and place/time shifting.... The same doesn't exist for digital media and the legal world seems to be swaying in the direction of keeping that from ever happening.
It would be insanely trivial for someone with an EE background to make a jumper emulator which is operated via parallel port that could switch the drive (or any other jumper for that matter), the issue is whether or not the drive's firmware honers the jumper setting at anytime or if it is a "check once" type of setting. I suspect that the firmware only cares about the jumper setting when the disk is first spun up, but I could be wrong.
While this is probably sound advice in theory, I'm sure that anyone following it will quickly find that the answer they receive from their local FD will be of virtually no value. The reason being that most localities spend very little money on staffing for their hazardous waste collection facilities which leaves them available to the citizenry only during a few days a week and usually for only a small subset of normal business hours. If my experience is anything like the majority of people on Slashdot, you generally will not want to take hours off from your place of employ in order to deliver your waste to the depot/transfer station/collection facility AND will find that location is not open on weekends or after/before work or basically any other time that would make it a convenient solution.
Don't disregard TCO, CFLs are notorious for ballast failures at a similar rate to incandescents due to power cycling. If one were to replace their frequently used lights with LEDs vs. CFLs I have a distinct feeling that the TCO would be lower given that the rate of "bulb" replacement would be much less frequent being that the LED would only need replacement once every few decades vs. every few years with the CFL.
Not sure where you are located.... But, for most of us Americans and Europeans you can bring your burnt CFLs (used up alkaline batteries too for that matter) to your local Ikea store where they will accept the waste for proper disposal free of charge.
I don't work for Ikea, I just like referring others to free resources that help people be more responsible in their consumption behaviors.
I've run both Linux and OS X on my beige box P4 desktop machine and the Darwin kernel seems much less responsive the the Linux kernel for everyday tasks AND the driver and open software availability are just crap compared to any modern Linux distro (or even cygwin on windows).
Once *bsd or Linux has a good enough ABI to run the OS X window manager minus the Darwin kernel THEN Apple will have something really nice (for the users) on their hands. In the mean time I don't understand the "mac tax" mentality where you pay more for hardware and software; and that isn't an exaggeration, a quality mac usb keyboard with mechanical key switches costs between 90-150USD while the equivalent for PC can be hand for practically nothing. The same can be said for other trivial peripheral like mice, web cams, tv capture hardware etc...
It is possible that someone out there has some great isight which shoots down my arguments, so by all means if I'm full of shit please provide me some solutions.
Since when has there ever been any expectation of anything but the most vanilla install of any Linux distro been expected to go correctly? The key to handling these things is a careful partitioning of file systems such that data is untouched by upgrade processes and a strong enough understanding of how the necessary services/programs are configured and interact with other core applications. If you find that you have a neither of these requirements handled it should be common knowledge that your experience with upgrading (or even running) linux will be troublesome in more then 50% of the cases regardless of what distro you use.
Slashdot Mirror
Okay, I'll bite... Lets see these bench marks.
Having been in the exact same situation I've tried all sorts of different solutions and I'd say the best current solution is NTFS, which is out of the box natively supported on both OSX and Windows (natch) and also available R/O in the default linux kernel as well as having strong R/W support now via ntfs-3g. Of course fat32 still works just fine for this application, but it's getting a little long in the tooth as far as advanced features and modern storage needs go (c'mon what is up with those weak filesize limits)!?!? And I've had some limited success with using ext2/3 on windows and linux but found that the windows kernel driver for ext2 was not very stable in my config and the userspace tools to read/write ext3 in windows was far too kludgy for my tastes; I haven't had a chance to try ext2/3 on OSX.
Also the first part of your second post is going on about shooting in nature and not having any 3g coverage... Well certainly there is more 3g coverage (in my country at least) than there is area covered by open wifi access points. And if you are shooting through 2GB cards in 3 hours I'd bet dollars to donuts that you'd be best served by having multiple cards and doing a "traditional" transfer to local for postprocessing and forgoing the entire wireless transfer scheme altogether....
Your initial argument reads like one of the pro hybrid car arguments where they base the maximum range of a car on the presumption that it starts with the batteries fully charged. It is a true statement, but it is being made about a situation that is generally atypical. Using that as the crux of your argument is the straw man I was calling you out on...
This next part might come out like gobblydegook since i'm just doing it as I'm going along... ;)
So lets split the difference and say that the average digital camera user falls somewhere between your example and what I say my typical usage is. Instead of 3.5MB per photo of your DSLR or my 1.25MB point and shoots we'll call an average picture 2.4MB. And your 3.5MB pictures fill a 2GB card in 3 hours, lets call that 195 photos/hr while my highest rate, while on vacation and taking waaay more than more normal rate, I clocked something like 30/hr at most. That gives us an average of 112 photos at 2.5 MB each per hour. Now you gave an estimated speed of 3.5MB per minute for BT+3g which is 210 of your DSLR photos per hour or something like 9.7 hours to upload 3 hours worth of photography, or 6.7 extra hours of uploading after shooting. I OTOH would take 54 hours of shooting at my stated rate/resolution to shoot 2gb and would have been uploading at faster than realtime, my card would never hit capacity as long as I had a signal and/or enough memory on my phone to cache the data (coincidentally my 3g handset has 2GB of addon memory and 64MB internal). So our theoretical "average" photog would take 4.6 hours to shoot 2gb worth of photos while uploading at about 70% of realtime incurring a 30% time to upload penalty...
Well, not sure if that math holds up... maybe i'll make a spreadysheet or something... but for now I'm gonna go enjoy dinner
You are making a straw man argument. If you are using a PAN which includes a 3g handset then you are bringing the network with you. As such, photos could be uploaded on the fly. So unless you shoot in contnuous mode and hold down the shutter button to intentionally fill the 2gb of storage you wouldn't be waiting "9 hours" to transfer your photos, the photos could instead be transferred seamlessly in the background.
If you are able to get a sustained 64kBps all the way from your camera to your webhost you will not be posting in 100% realtime, although you'd be damn near it and I'd also say it is unlikely that most people would be shooting over 2GB of pictures in any given 9 hours stretch. The bottle neck in nearly all cases will be the photographer, not the transfer speed.
I'd much prefer an SD card with a bluetooth adaptor built in that could leverage the 3G wireless internet connection which is the true core of the PAN (Personal Area Network) that is always touted as being the logical goal of the bluetooth architecture. I mean really, BT chipsets are far more optimized for power than wifi and comes with far fewer limitation as to the connections it can make. Let the devices choose the path of least resistance to the internet, be it tunnel over a phone, pda, laptop, or whatever the marketplace has in store next.
honestly I think that the working group that came up with BT designed it for exactly this sort of purpose. It'd be stupid not to also use this type of integration between PAN components to further enhance the meta data richness of the content created by the camera. GPS, PDA, camera, 3gphone, and headset sounds like a pretty good recipe for being your own gargoyle. I for one wouldn't mind being able to publish video, photo, sound, and location data at a moment's notice directly to the internet. If we are bound to live in surveillance state I'd sure like to get a good grip on the technology before Big Brother does.
Sure, just show me how I'm going to get a 4 hour repair/replace onsite technician response contract for this $3000 beige box NAS device. I'll buy 12 of them. Until then I'm gonna stick to my Netapps ;)
WTF? Are you 12 years old now or something? 10GB was orders of magnitude larger than anything I could possibly hope for when I was "9 or 10" years old.
Now you know!
And knowing is some fractional portion of some sort of battle, or so I've been told.
I'm just pointing out that your comment was a pure speculation on what people would or wouldn't do with no factual basis to point to. I personally have seen systems which do appear to be working. You on the other hand just throw the suggestion out the window as something that people are obviously not going to participate in. Support and proof are not interchangeable words... I feel no burden to prove my experience to you however I can share that experience as support for my claims. You have done neither...
Not sure where you are located, but here (Oakland, CA) there are CFL recycling bins at Ikea, Home Depot, the Waste Transfer Station, and many other hardware stores and other locations where you would purchase CFls... Usually the bins have quite a few bulbs in them which would seem to indicate that, contrary to your unsupported claim, people will and do recycle and properly dispose of items which may not be covered in their normal curbside pickup.
YMMV
There are people who will build "it, whatever it is" however I think you are drastically understating the difficulty it is to get a board with the complexity that a modern motherboard requires actually made. Hobby type board manufacturers rely on the fact the most hobby electronics use 1or 2 sided boards, not 6-10 layer boards with exact specs on trace length and orientation that is needed to fit and support the huge number of very high speed components that is needed for even the most basic P4 or Athlon.
OTOH if you "opensource motherboard" is an 8088 clone built with discrete components you just might be able to get that done...
It does have 300% more blue LEDs than the last gen sun hardware though ;)
I recall reading an article in a magazine a few years back which was talking about the research being done into using Perflurocarbon based blood replacement. My memory of the article is hazy, but iirc there was some mention of some human trials being done in the poorer parts of Africa where the demand for blood cannot be met due to the lack of refrigeration and the abundance of disease in the population.
The situations where a gun trumps a knife are those in open areas with distances between target and shooter over 20 feet and/or the agressor is the one with the firearm and the edged weapon is used for defense.
The fact of the matter is that regardless of the weapons employed, the agressor has many many cards stacked in their favor and the defender has very little in the way of reaction time in which to change the circumstances of the situation.
This training video shows the effectiveness of a knife vs. a holstered weapon at 12 feet. Clearly even a trained police officer who has fore knowledge of the impending attack cannot sufficiently defend themselves.
The Force Science Research Center, found the following while studying this exact encounter situation:
And keep in mind, the officers in this study were both aware of the impending attack and were mentally prepared to draw down and fire on the agressor.
AND VHS has a strong history of legal precedent allowing consumer recording and place/time shifting.... The same doesn't exist for digital media and the legal world seems to be swaying in the direction of keeping that from ever happening.
Too bad that the software release they've announced won't run on any of that equipment and the hardware is only vapor...
It would be insanely trivial for someone with an EE background to make a jumper emulator which is operated via parallel port that could switch the drive (or any other jumper for that matter), the issue is whether or not the drive's firmware honers the jumper setting at anytime or if it is a "check once" type of setting. I suspect that the firmware only cares about the jumper setting when the disk is first spun up, but I could be wrong.
While this is probably sound advice in theory, I'm sure that anyone following it will quickly find that the answer they receive from their local FD will be of virtually no value. The reason being that most localities spend very little money on staffing for their hazardous waste collection facilities which leaves them available to the citizenry only during a few days a week and usually for only a small subset of normal business hours. If my experience is anything like the majority of people on Slashdot, you generally will not want to take hours off from your place of employ in order to deliver your waste to the depot/transfer station/collection facility AND will find that location is not open on weekends or after/before work or basically any other time that would make it a convenient solution.
Don't disregard TCO, CFLs are notorious for ballast failures at a similar rate to incandescents due to power cycling. If one were to replace their frequently used lights with LEDs vs. CFLs I have a distinct feeling that the TCO would be lower given that the rate of "bulb" replacement would be much less frequent being that the LED would only need replacement once every few decades vs. every few years with the CFL.
I don't work for Ikea, I just like referring others to free resources that help people be more responsible in their consumption behaviors.
I've run both Linux and OS X on my beige box P4 desktop machine and the Darwin kernel seems much less responsive the the Linux kernel for everyday tasks AND the driver and open software availability are just crap compared to any modern Linux distro (or even cygwin on windows). Once *bsd or Linux has a good enough ABI to run the OS X window manager minus the Darwin kernel THEN Apple will have something really nice (for the users) on their hands. In the mean time I don't understand the "mac tax" mentality where you pay more for hardware and software; and that isn't an exaggeration, a quality mac usb keyboard with mechanical key switches costs between 90-150USD while the equivalent for PC can be hand for practically nothing. The same can be said for other trivial peripheral like mice, web cams, tv capture hardware etc... It is possible that someone out there has some great isight which shoots down my arguments, so by all means if I'm full of shit please provide me some solutions.
Since when has there ever been any expectation of anything but the most vanilla install of any Linux distro been expected to go correctly? The key to handling these things is a careful partitioning of file systems such that data is untouched by upgrade processes and a strong enough understanding of how the necessary services/programs are configured and interact with other core applications. If you find that you have a neither of these requirements handled it should be common knowledge that your experience with upgrading (or even running) linux will be troublesome in more then 50% of the cases regardless of what distro you use.