That's even sneakier than the "virus I would design if I were a virus-designer working for a hard drive company". Every day, this virus would randomly select one unused cluster table entry and write the "bad cluster" code on it (it's only 28 bits on a FAT32 drive). The result of this is that disk space would just transparently and nondestructively "disappear", without being easily recoverable by the average user (do any standard Windows disk utilities support a "re-test bad sectors" option anymore?), until the user notices he is nearly out of disk space, and buys a bigger hard drive (using the "helpful" free imaging software included to transfer all data + programs + Bootsector to the new drive...)
Some months back my brother bought a 120GB drive and installed it into an old computer that definitely wasn't meant to support a drive of that size. I think it was a Pentium/400 from the days when those were top of the line. He was ever so pissed that he couldn't use the full drive capacity, and went on a mission of flashing new BIOSes, fiddling with EZ-Drive and generally mucking around with low-level drive tools (not Ghost specifically, I think it was Partition Magic) until the computer would recognize his huge-ass drive as a huge-ass drive.
That accomplished, he wasted no time dumping every file imaginable onto the drive, especially a huge collection of MP3s. He came to me ecstatic, blathering something to the effect of, "DoOOoOooD! I just put 145 gigs' worth of MP3s from all my other drives onto the '120GB'!" (generally thinking he had outsmarted that dumb HD manufacturer that underpartitioned/undersized/reserved spare sectors on the drive without telling anyone). Then wasted no time in reformatting all the drives he copied the files from, to install OSes, etc. on them.
It wasn't until a few days later that I got the frantic call: "Do you know a program that will mach up pieces of files with their filenames?" It turns out that about a third of those 145 gigs of MP3s now consisted of part one song, part another...
One of my work-related projects required me to sort of crash-course myself in the FAT filesystem, and the one thing it made me notice, above all else, is just how many places in the filesystem you can safely stuff little turds of data. (From the good ole days to today, 'turds'--little bits of state-indicating data--are used by some software anti-copy routines, shareware (to prevent reinstalling for another 30 free days), etc., to store hidden information about previous installs)
If I didn't have better things to do, I'd try my hand at writing a bootsector program, just because it would be so simple. But back on that data-stuffing. Need just a bit? 'Reserved' bits all over, that official whitepapers specifically caution everyone to ignore. Reserved bytes, even, not just in the bootsector, but in the partition tables, BPBs and just about anywhere else you can think of. Want to create a *really* hidden file? Forget the Hidden/System/etc. attribute bits, put it in the root directory and mark it as a Volume ID. Yeah, there's only *supposed* to be one Volume ID file (0 bytes), but.... ("Sadly", at least Windows 2000 chkdsk will set the VolumeID file at the beginning of the root back to 0 bytes if it's changed...don't know if it will check for extras though. I think it's even less likely that it will look for a file marked VolumeID stuffed inside a subdirectory somewhere). On a big FAT32 drive, 0xFFFFF7 (just from memory here, you may want to double-check that I have enough Fs) can be an allocatable cluster number. However, it also happens to be the FAT32 code for 'bad cluster'; so this cluster (if it exists for the particular drive size/formatting) is not actually allocatable for data. So there's another place for a really big turd. Many disks are not actually bootable (basically any drive with no OS on it, floppy disks that aren't explicitly 'Boot floppies', extra non-boot HDDs for porn and MP3 storage), but they all contain an executable boot sector. (It's usually a piece of useless code that prints "Missing Operating system" or some similar message to your screen.) There's another 448(?) bytes that could be put to better use by paranoid shareware guy, although FDISK/MBR will make short work of it. (So, smarter paranoid shareware guy would stuff the turd in an unused partition table entry, where/MBR won't touch it. Who but ubergeeks has 4 primary partitions on a drive, anyway?)
With a little dye (food coloring), water and isopropyl alcohol, you can make your own ink. Now, I do NOT recommend you try this at home unless you have a printer you don't care about, because there's no guarantee that the specific coloring you use won't gum up the heads and such. Also, not recommended for quality (esp. photographic) prints, as most household food colorings/etc. are not available in exact CMYK compositions.
Try something like 1 part isopropyl to 1 part distilled water, and add dye til you feel it's entirely too much (the print will end up considerably lighter than you might expect from looking at the 'ink' itself!). If you want to do it proper, the official formulas probably call for extra ingredients such as glycol, detergents (to prevent gumming up the heads), etc.
Ah....for all of you who are going to continue jumping in with "1 in 20? more like 1 in 1..." without reading the article...
The "1 in 20" figure the researchers got was not from scanning the HDDs with Spybot/AdAware/etc....they sniffed for known packets from FOUR of the significantlymorethanfour known malwares.
So, to be detected at all, the machines had to be running and the spyware loaded and actively broadcasting packets during the sampling period. Given this lack of an exhaustive check, the 1 in 20 figure doesn't surprise me. (We all know it is 1 in 1...:-)
Re:OT - Re:Urban Legend Time.
on
See Spot Surf
·
· Score: 1
Wow, it really is funny how these things work. Where I come from, "horse dick" is not exactly an insult. A few people I know started perpetuating a rumor...a couple with firsthand experience, but more just blindly spreading rumors without checking into them, so I've heard 'em all - The Ruler, the Prudential, Vlad the Impaler, references to making "bitch-kebobs", etc.
Needle-Dick the Bug Fucker, that I can understand...but what ever gave you the notion to tease a guy by calling him "horse dick", and whatever gave him the notion to take it as such? Of course, I suppose the context matters. Saying it while knocking the books out of his hands, throwing him into lockers, and etc., probably clarified it...
Imagine placing these chips on top of light poles every 1/2 mile on big city highways. Now enable them to start photographing drivers and license plates whenever excess speeds are detected...
I am currently a firmware writer for a company that specializes in bleeding-edge datalogging and sensor products for some very large clients in the US (including, you could say, THE largest). I can't give too many details about my latest project's internals, but this Thursday, the company I hack for shipped out a prototype of the product that will be rocked, jarred, tumbled, drenched and otherwise abused on the Atlantic for the next 8 months or so. This logger (including batteries) is a completely sealed unit that won't even be able to be opened until it's time to retrieve the data. Now, forget about those wireless-router wusses who can just publish a Web-based firmware update when they find out that their shit doesn't work. MY shit needs to WORK, and work correctly, when it hits the door, because there's no way we'll be able to push a fix-it patch later.
Call me oldfashioned, but I would hope that a product would ship actually having all the features advertised on its packaging. However, I do understand the importance of time-to-market, and the sad reality of getting trounced by the competition pushing out a half-working version of the product you're busy perfecting, promising they'll fix it "real soon now".
But all that's a minor annoyance in comparison with the following: What boils my blood is 'updates' that REMOVE features, or otherwise intentionally cripple a product. Either so as not to compete with an upcoming 'DeLuxe' model, or to avoid the possibility of legal uncertainties (e.g. at least one PVR manufacturer who disabled certain time-skipping [commercial-skipping] buttons post-sale), as a form of planned obsolescence, or whatever other reason. To me, such intentional crippling or other forms of 'self-help' amounts to no more than fraud.
Bah, this is an ancient thread now by/. terms; I'm sure this won't get read:-)
The behavior you're describing sounds suspiciously like a discrete BCD-to-7-segment decoder (74LS47?) chip when it's given bit values corresponding to values above '9'. The decoder's input is 4 bits, giving up to 16 possible digits to be displayed. Since they're only 'expected' to display 0-9, however, part designers are free to choose arbitrary patterns for the remaining bit combinations. (Or are they meaningful patterns to someone, somewhere?)
Any 'LS47 designer out there know why these particular patterns were chosen?
The most recent actual "hack" I've been involved with is the single-use (Dakota) camera. So far, the 25-picture disposable camera has been made to also support time-lapse computer-controlled photography, continuous video (i.e. Webcam) modes, and been able to store (in my brief, informal test) 58 pictures.
The rest of these might not be considered hacks per se, just projects.
A project that never got finished would have put a high-power subwoofer amplifier in my car, complete with an authentic '60s fluorescing vacuum tube as a level display. Much classier than the usual LED-bargraph arrangements popular with the kiddies these days. Unfortunately, in the middle of building this I got offered a job and moved 'cross-country, but didn't have room to pack the unfinished bits+pieces and all my electrical test equipment in my little 2-door.
In my college years, I had the position of running an underground student newspaper. An issue was released 'every few weeks' when its dedicated editors were free/bored enough to put one together, but one thing everyone thought would be nice would be to commandeer the University (dorm) cable system after-hours for a student-run movie and wierd footage channel. Starting at about midnight or so, this would replace a lame "information channel" text marquee (which was always several weeks out of date and advertising events whose deadlines had come and gone), that was currently occupying a perfectly good cable channel.
We had obtained keys to the main hub room (also the cable feed room), so inserting the signal was not a problem. The student TV footage was intended to begin late at night, when university officials were guaranteed not to be watching, and would be pre-recorded. This presented a minor problem, however: everyone on the 'staff' had early classes and poor memories, and could not be counted on to get into the hub closet after hours to insert the day's programming and press 'play'. Also, while some students (volunteering for the Computer center) did legitimately have access to these areas, students going in and out of there after hours would arouse unnecessary suspicion from campus security.
It was decided that the best solution was to equip the VCR with a 'remote control' of sorts that would allow it to be controlled over the dorm network via the abundant Ethernet connections available in the room. This would allow for automated starting and stopping as well as manual intervention as necessary; footage could then be loaded during the daytime hours at the convenience of those involved.
Making a VCR Internet-ready is not has hard as it sounds. I simply built a board with eight simple Darlington transistor circuits (corresponding to 8 data pins on a parallel port) to drive the important VCR function buttons via this port. A simple Web server (disposable '386) running a perl-based CGI interface allowed Web-based control of the parallel port bits, which in turn operated the disposable VCR with wires soldered into the appropriate front-panel switches.
The tricky part then became finding controversial/interesting/non-stupid, but legal, student-produced content worth displaying, but that's another story.
Ever consider that the dinosaurs might still rule the Earth if they had MADMEN?
That's an interesting point you bring up. It makes me wonder, how would today's Earth have evolved if the dinosaurs had never been wiped? Would the planet be ruled by huge, smart reptiles? Or perhaps dumb ones?
Perhaps the occasional cataclysm is beneficial to the planet in the long run, by wiping out species that have hit (or are approaching) some sort of evolutionary wall. If humans were similarly wiped by an asteroid, would something still more advanced evolve in our absence?
Slashdot Mirror
That's even sneakier than the "virus I would design if I were a virus-designer working for a hard drive company". Every day, this virus would randomly select one unused cluster table entry and write the "bad cluster" code on it (it's only 28 bits on a FAT32 drive). The result of this is that disk space would just transparently and nondestructively "disappear", without being easily recoverable by the average user (do any standard Windows disk utilities support a "re-test bad sectors" option anymore?), until the user notices he is nearly out of disk space, and buys a bigger hard drive (using the "helpful" free imaging software included to transfer all data + programs + Bootsector to the new drive...)
Some months back my brother bought a 120GB drive and installed it into an old computer that definitely wasn't meant to support a drive of that size. I think it was a Pentium/400 from the days when those were top of the line. He was ever so pissed that he couldn't use the full drive capacity, and went on a mission of flashing new BIOSes, fiddling with EZ-Drive and generally mucking around with low-level drive tools (not Ghost specifically, I think it was Partition Magic) until the computer would recognize his huge-ass drive as a huge-ass drive.
That accomplished, he wasted no time dumping every file imaginable onto the drive, especially a huge collection of MP3s. He came to me ecstatic, blathering something to the effect of, "DoOOoOooD! I just put 145 gigs' worth of MP3s from all my other drives onto the '120GB'!" (generally thinking he had outsmarted that dumb HD manufacturer that underpartitioned/undersized/reserved spare sectors on the drive without telling anyone). Then wasted no time in reformatting all the drives he copied the files from, to install OSes, etc. on them.
It wasn't until a few days later that I got the frantic call: "Do you know a program that will mach up pieces of files with their filenames?" It turns out that about a third of those 145 gigs of MP3s now consisted of part one song, part another...
Oh, this article has to do with mobile phones. I saw the words "New Net", "battle", and a nonexistant TLD, and got the wrong idea.
One of my work-related projects required me to sort of crash-course myself in the FAT filesystem, and the one thing it made me notice, above all else, is just how many places in the filesystem you can safely stuff little turds of data. (From the good ole days to today, 'turds'--little bits of state-indicating data--are used by some software anti-copy routines, shareware (to prevent reinstalling for another 30 free days), etc., to store hidden information about previous installs)
/MBR will make short work of it. (So, smarter paranoid shareware guy would stuff the turd in an unused partition table entry, where /MBR won't touch it. Who but ubergeeks has 4 primary partitions on a drive, anyway?)
If I didn't have better things to do, I'd try my hand at writing a bootsector program, just because it would be so simple. But back on that data-stuffing. Need just a bit? 'Reserved' bits all over, that official whitepapers specifically caution everyone to ignore. Reserved bytes, even, not just in the bootsector, but in the partition tables, BPBs and just about anywhere else you can think of. Want to create a *really* hidden file? Forget the Hidden/System/etc. attribute bits, put it in the root directory and mark it as a Volume ID. Yeah, there's only *supposed* to be one Volume ID file (0 bytes), but.... ("Sadly", at least Windows 2000 chkdsk will set the VolumeID file at the beginning of the root back to 0 bytes if it's changed...don't know if it will check for extras though. I think it's even less likely that it will look for a file marked VolumeID stuffed inside a subdirectory somewhere).
On a big FAT32 drive, 0xFFFFF7 (just from memory here, you may want to double-check that I have enough Fs) can be an allocatable cluster number. However, it also happens to be the FAT32 code for 'bad cluster'; so this cluster (if it exists for the particular drive size/formatting) is not actually allocatable for data. So there's another place for a really big turd. Many disks are not actually bootable (basically any drive with no OS on it, floppy disks that aren't explicitly 'Boot floppies', extra non-boot HDDs for porn and MP3 storage), but they all contain an executable boot sector. (It's usually a piece of useless code that prints "Missing Operating system" or some similar message to your screen.) There's another 448(?) bytes that could be put to better use by paranoid shareware guy, although FDISK
If you can find a cellphone that runs on micro-amps, go for it :-)
With a little dye (food coloring), water and isopropyl alcohol, you can make your own ink. Now, I do NOT recommend you try this at home unless you have a printer you don't care about, because there's no guarantee that the specific coloring you use won't gum up the heads and such. Also, not recommended for quality (esp. photographic) prints, as most household food colorings/etc. are not available in exact CMYK compositions.
Try something like 1 part isopropyl to 1 part distilled water, and add dye til you feel it's entirely too much (the print will end up considerably lighter than you might expect from looking at the 'ink' itself!). If you want to do it proper, the official formulas probably call for extra ingredients such as glycol, detergents (to prevent gumming up the heads), etc.
You mean meet one of the many ravishing, worldly, exciting, virile Slashdot men?
Ah....for all of you who are going to continue jumping in with "1 in 20? more like 1 in 1..." without reading the article...
:-)
The "1 in 20" figure the researchers got was not from scanning the HDDs with Spybot/AdAware/etc....they sniffed for known packets from FOUR of the significantly more than four known malwares.
So, to be detected at all, the machines had to be running and the spyware loaded and actively broadcasting packets during the sampling period. Given this lack of an exhaustive check, the 1 in 20 figure doesn't surprise me. (We all know it is 1 in 1...
But... Have you considered GROWING grass?
You mean, running around at night watering all the lawns you mowed during the day?
Back at LTHS... I never saw what the sign originally said; I only saw the picture taken afterward. For an entire weekend, passers-by saw:
A GIANT POON ATE MR. RICHARDS
I have *several* of those. They have 6 sides and little dots all over...
Finding and squishing a sofware bug can be mildly satisfying.
Finding and squishing a real bug just gets your shoes all gooey.
And can be satisfying. DIEEEE! *splut*
Does this invalidate the stereotype of newbie sites on free webspace hosting "pictures of my dog" as their primary content?
I didn't, you insensitive clod!
Wow, it really is funny how these things work. Where I come from, "horse dick" is not exactly an insult. A few people I know started perpetuating a rumor...a couple with firsthand experience, but more just blindly spreading rumors without checking into them, so I've heard 'em all - The Ruler, the Prudential, Vlad the Impaler, references to making "bitch-kebobs", etc.
Needle-Dick the Bug Fucker, that I can understand...but what ever gave you the notion to tease a guy by calling him "horse dick", and whatever gave him the notion to take it as such? Of course, I suppose the context matters. Saying it while knocking the books out of his hands, throwing him into lockers, and etc., probably clarified it...
Imagine placing these chips on top of light poles every 1/2 mile on big city highways. Now enable them to start photographing drivers and license plates whenever excess speeds are detected...
I am currently a firmware writer for a company that specializes in bleeding-edge datalogging and sensor products for some very large clients in the US (including, you could say, THE largest). I can't give too many details about my latest project's internals, but this Thursday, the company I hack for shipped out a prototype of the product that will be rocked, jarred, tumbled, drenched and otherwise abused on the Atlantic for the next 8 months or so. This logger (including batteries) is a completely sealed unit that won't even be able to be opened until it's time to retrieve the data. Now, forget about those wireless-router wusses who can just publish a Web-based firmware update when they find out that their shit doesn't work. MY shit needs to WORK, and work correctly, when it hits the door, because there's no way we'll be able to push a fix-it patch later.
Call me oldfashioned, but I would hope that a product would ship actually having all the features advertised on its packaging. However, I do understand the importance of time-to-market, and the sad reality of getting trounced by the competition pushing out a half-working version of the product you're busy perfecting, promising they'll fix it "real soon now".
But all that's a minor annoyance in comparison with the following: What boils my blood is 'updates' that REMOVE features, or otherwise intentionally cripple a product. Either so as not to compete with an upcoming 'DeLuxe' model, or to avoid the possibility of legal uncertainties (e.g. at least one PVR manufacturer who disabled certain time-skipping [commercial-skipping] buttons post-sale), as a form of planned obsolescence, or whatever other reason. To me, such intentional crippling or other forms of 'self-help' amounts to no more than fraud.
Bah, this is an ancient thread now by /. terms; I'm sure this won't get read :-)
The behavior you're describing sounds suspiciously like a discrete BCD-to-7-segment decoder (74LS47?) chip when it's given bit values corresponding to values above '9'. The decoder's input is 4 bits, giving up to 16 possible digits to be displayed. Since they're only 'expected' to display 0-9, however, part designers are free to choose arbitrary patterns for the remaining bit combinations. (Or are they meaningful patterns to someone, somewhere?)
Any 'LS47 designer out there know why these particular patterns were chosen?
The most recent actual "hack" I've been involved with is the single-use (Dakota) camera. So far, the 25-picture disposable camera has been made to also support time-lapse computer-controlled photography, continuous video (i.e. Webcam) modes, and been able to store (in my brief, informal test) 58 pictures.
The rest of these might not be considered hacks per se, just projects.
A project that never got finished would have put a high-power subwoofer amplifier in my car, complete with an authentic '60s fluorescing vacuum tube as a level display. Much classier than the usual LED-bargraph arrangements popular with the kiddies these days. Unfortunately, in the middle of building this I got offered a job and moved 'cross-country, but didn't have room to pack the unfinished bits+pieces and all my electrical test equipment in my little 2-door.
In my college years, I had the position of running an underground student newspaper. An issue was released 'every few weeks' when its dedicated editors were free/bored enough to put one together, but one thing everyone thought would be nice would be to commandeer the University (dorm) cable system after-hours for a student-run movie and wierd footage channel. Starting at about midnight or so, this would replace a lame "information channel" text marquee (which was always several weeks out of date and advertising events whose deadlines had come and gone), that was currently occupying a perfectly good cable channel.
We had obtained keys to the main hub room (also the cable feed room), so inserting the signal was not a problem. The student TV footage was intended to begin late at night, when university officials were guaranteed not to be watching, and would be pre-recorded. This presented a minor problem, however: everyone on the 'staff' had early classes and poor memories, and could not be counted on to get into the hub closet after hours to insert the day's programming and press 'play'. Also, while some students (volunteering for the Computer center) did legitimately have access to these areas, students going in and out of there after hours would arouse unnecessary suspicion from campus security.
It was decided that the best solution was to equip the VCR with a 'remote control' of sorts that would allow it to be controlled over the dorm network via the abundant Ethernet connections available in the room. This would allow for automated starting and stopping as well as manual intervention as necessary; footage could then be loaded during the daytime hours at the convenience of those involved.
Making a VCR Internet-ready is not has hard as it sounds. I simply built a board with eight simple Darlington transistor circuits (corresponding to 8 data pins on a parallel port) to drive the important VCR function buttons via this port. A simple Web server (disposable '386) running a perl-based CGI interface allowed Web-based control of the parallel port bits, which in turn operated the disposable VCR with wires soldered into the appropriate front-panel switches.
The tricky part then became finding controversial/interesting/non-stupid, but legal, student-produced content worth displaying, but that's another story.
but phones are simple, and don't hold a big charge
:-)
Until someone rings you, and better than 60VAC comes down the line
My car leaks oil, you insensitive clod!
"Stop overwriting my money, you bastards!"
Ever consider that the dinosaurs might still rule the Earth if they had MADMEN?
That's an interesting point you bring up. It makes me wonder, how would today's Earth have evolved if the dinosaurs had never been wiped? Would the planet be ruled by huge, smart reptiles? Or perhaps dumb ones?
Perhaps the occasional cataclysm is beneficial to the planet in the long run, by wiping out species that have hit (or are approaching) some sort of evolutionary wall. If humans were similarly wiped by an asteroid, would something still more advanced evolve in our absence?
...Humans throw little rocks from big r...oh.
Which way do you fire the flares?