Slashdot Mirror
Xerox Exploits Printer Flaws To Make Pseudo-Holograms
Red Wolf writes "A chance discovery by Xerox lets printers superimpose glossy images on regular printouts, creating the possibility for document authentication along the lines of holograms on credit cards. The new technology, called Glossmark, can use ordinary office printers to superimpose a glossy image on an ordinary printed document in a way that can't be photocopied or otherwise easily reproduced."
Who wants to let me borrow his credit cards?
It's not a bug, it's a feature!
If it's easily available on a commercially available printer, how does it provide great security?
I can just imagine it now. A nice crappy background image of a lake, a school, or some other kind of normal porn background image and then superimpose a high quality, very shiny, very naked person (or animal if that's your thing).
I hope this gets used on US currency. Holograms haven't been used yet because they haven't survived the torture tests. Maybe this will fare better.
Isn't running an already-printed page through a printer a violation of the DMCA or something?
Where the value of X-Mailer: is the true measure of a man...
way that can't be photocopied or otherwise easily reproduced
Uh, except for on another Xerox printer?
how is this secure if everyone and their baby's daddy's momma can print up whatever they want. tickets to the superbowl - $0 credit card used to buy $9,000 Plasma TV - $0 fake id - $0 the fact that you made them up on a printer at work - priceless
True... who is stopping you from printing a "Glossmark" on a modified version of the "hard copy" anyway??
What? That doesn't make any sense. Did you read the article?
Any bug you can control is a feature! The big question in my mind would be... what would stop some enterprising individual from replicating this bug to forge the watermarks?
It can't be reproduced or copied. Unless you own a Xerox printer I guess:
Can be produced by existing Xerox printing solutions.
I don't really see how this works. If there's a document I want to fake I just whip out Quark and reproduce the Glossmark on my Xerox printer. Wha?
Support the First Amendment. Read at -1
And then Xerox gives up the technology, somebody else picks it up and makes a bundle.
Let's see... Mouse, GUI, Ethernet, Palm Graffiti, WYSIWYG word processors, and more
--You will rephrase your request for me to go to hell. Goto statements are not acceptable programming constructs
"Can be produced" isn't the same as "can be reproduced." Sure, I guess you could print out your own copies -- if you had access to the original images. If I understand correctly, most of the point is that you can't just scan the image and retain the glossmark effect.
From reading the article (yes I read actually read it), it would seem that only the "wax" type color laserjet printers have this ability. There was a Slashdot article a while back that dealt with color laser printers and alternatives to inkjets. The news.com.com article does specify the models or type of printers where this was discovered. Any other info on this?
I'm sure some hackers will try to do some mods on their printers to control this as well. {cough}fake holograms{/cough}
On another note, how cool a job do these "Xerox Scientists" have? I need to get a job where I can hardware hack like these guys.
MMORPG fan-boy? Prove your worth
I don't want to sound like a killjoy, but what do you think will be the odds that Xerox lets the average person get their hands on this technology? More likely that they'll take out about half a dozen patents on every known way of implementing it and then enforce ridiculously high licensing fees on any product or organisation which tries to use it.
Bash script for FP whores
Once the technique is widely spread, it's utility for authenticity will plummet as anyone including document forgers will be able to reproduce it.it's only useful as long as it stays a rare curiosity. By becoming popular it would make itself useless.
Democracy is a sheep and two wolves deciding what to have for lunch. Freedom is a well armed sheep contesting the issue
If you can see the 2 different images why can't I just do this myself ??? All i need is a photoshop'd copy of the original 2 pics (or something very similar) and a xerox .. oh ya, and time tons of time...
Companies are going to adopt this technology because they can create an "uncopyable" product (probably tickets, coupons, and other vouchers), and they already have the technology in their office.
In the mean time, some counterfeiter who has the same technology in their office or home will simply copy the main image and recreate the superimposed image in a graphics program. Then he will be able to print "authentic" tickets or whatever whenever he wants.
The number one blockade in stopping conterfeiters is the machine that produces the items they want to counterfeit, not the complexity of the artwork or image. Sure, the complex image and holograph help, but that is mainly because consumer level and most business level products can't produce images that complex. Give me a few months and I could make a damn good couterfeit $20 bill if I only had the paper and the press that makes them. It wouldn't be perfect, but the average cashier wouldn't notice.
If I drive fast enough at the red light, it'll appear green.
Unless each individual printer has some sort of unbreakable digitally signed super one-way crypto identifier in it, this sounds like a pretty easy thing to reproduce. And what's the use of mass-printed original fake documents?
I'd like to see some details about how they do this. Our office just got a Xerox color laser printer and would probably use this for some non-security stuff.
When they say "current printers," it sounds like ours would just need a driver upgrade or something. I don't know how that's possible, but I don't know much about hardware and drivers. I'm also curious whether they'll charge for this new "feature" or just include it as an upgrade. Or whether it will only be available on newer high end printers despite working on current technology.
t's a little much to expect a hologram to come out of your office printer, but scientists at Xerox think they have the next best thing.
On Thursday, the company is unveiling a new technology it calls "Glossmark," which can use ordinary office printers to superimpose a glossy image on an ordinary printed document in a way that can't be photocopied or otherwise easily reproduced.
Taking advantage of eccentricities in laser printing processes, once viewed as flaws, the Xerox scientists think they've found a way to authenticate hard copies of printed documents in much the same way that holographic stickers prove the validity of credit cards and drivers licenses.
"This does speak to something that is going to need to be addressed to ensure hard-copy security," said Dan Corsetti, an industry analyst with research firm IDC, who saw a demonstration several months ago. "There really is no reliable or affordable way of securing the content on hard-content documents, apart from putting it in a vault and locking it up."
The new Xerox process, while still a long way from market, points up a persistent demon that has dogged the technology industry's longstanding efforts to secure digital content, whether it be corporate documents or copyrighted movies and music.
Hundreds of millions of dollars have been spent on efforts to encrypt or otherwise protect content against people who might make perfect digital copies with a computer or other device. But little of this technology has been able to do anything about decidedly easy methods of reproduction such as photocopying a hard copy of a document, or taping a song as it comes out of a stereo's speakers.
In Xerox's case, the Glossmark procedure came about almost wholly as an accident.
Lab researchers had long been aware of an issue with some printers producing glossy areas in a printout, which would reflect light a little more strongly than the surrounding area. The phenomenon was an artifact of the printing process, in which plastic-like toner was melted onto the paper.
Studying a way to reduce the so-called differential gloss, researchers discovered that they could actually manipulate it, controlling where the glossy areas appeared in a printed document.
"They came back and said, 'We don't know if we can reduce it, but we sure can enhance it,'" said Rob Rolleston, the laboratory manager overseeing the Xerox husband-and-wife team that worked on the process. "They said, 'Wow, we really can control this much more than we thought we could.'"
The team worked with ways to send glossy images to ordinary color office printers and before long had figured out a way to create a consistent pattern with the glossy areas. The embedded glossy goatse image was invisible when the document was examined straight on, but would appear, hologram-like, when held at the right angle to the light, they found. However, after seeing a three dimensional rendering of the goatse man; many of the test viewers are now blind and seeking psychiatric help. "This technology certainly is not without it's potential health risks" Rob was heard to understate.
The technology isn't poised to find its way immediately into products for Xerox, which is struggling to fend off increasing competition from rivals such as Ikon and Canon in its core markets. Nor is the company wholly convinced that the discovery will add up to a new security technology.
The ability to make shiny images appear inside of printed documents could also be used in greeting cards or for artistic purposes, Rolleston said. The company ultimately will have to decide--if it is intended to be a security-enhancing process aimed at authenticating documents, having the technology widely available to would-be document forgers would be a problem, Rolleston said.
Analysts are pleased that the company is thinking about the issue--even if only by accident.
"Document security is a leading concern among IT users," IDC's Corsetti said. "The hard copy has always been the weak link in the security chain."
photocopied or otherwise easily reproduced. ...unless of course you have a Glossmark printer.
D.O.U.O.S.V.A.V.V.M.
It doesn't look really useful for preventing professional counterfieting, but for "casual" things [retail reciepts, HR files, inter-company corrospandance, etc.] It could come in handy for quick verification.
Now this might be a stupid question but using this would it be possible to have say two pages imposed on one so you could "read" a book by twisting the page...
Just a thought
Rus
Cheap UK and US VPS
The "Glossmark" thing looks like a 1-bit overlay to me. You don't need any fancy equipment to reproduce that. Just a pair of eyes and the gimp.
I suppose the next step would be ATMs that print money???
An HTML version is cached by Google here.
The new technology, called Glossmark, can use ordinary office printers to superimpose a glossy image on an ordinary printed document in a way that can't be photocopied or otherwise easily reproduced."
..
1) place document on surface
2) mount digital SLR camera on tripod
3) tilt surface or camera until image appears
4) Xerox OCR
5) reprint using Xerox Glossmark
6)
7) profit!
SCO employee? Check out the bounty
Well, I'm just passing on what I've read. But apparently the Secret Service is concerned about people using bleached ones to makes twenties and fifities. So, the paper costs a buck a note. And even at el cheapo bubble jet grade 2400 dpi it's easy to reproduce rather fine detail.
It's all good news though because it's the goverment's fucking job to make the currency and the failure to get on the stick with an e-currency is probably the single greatest hinderance to the e-economy. Credit cards my ass.
How can this be legal under the DMCA? I mean they are obviously circumventing their own protection scheme. Has anyone notified SCO, the MPAA, the RIAA, or better yet the FBI?!?!
Its just multilayer printing, its has NOTHING to do with a hologram.
Its interesting, though pretty much common sense, if you have run a sheet thru a printer 15 times.. ( and pray it doesn't jam.. the structure of paper is changed when it passes thru a fuser.. every time after that you risk paper jams. )
---- Booth was a patriot ----
Yet another sad commentary on the rampant cover-ups of the true nature of the pseudo-hologram industry.
A man's reach must exceed his grasp, or what's an erection for?
...for inkjet cartridges.
Tickets to the Superbowl: $0
Credit card to charge up $9000 in stereo equipment: $0
Same credit card, Quad-CPU, 16 gigs RAM, 1 terrabyte machine with all the latest blings: $0
A lawyer that can use the "it was a bug in the printer" defense to successfully get you off: Priceless.
*** *** You're just jealous 'cause the voices talk to me... ***
Quick, sic the DMCA on them! Off to Guantanamo the developers go!
A killer app that just happens to also be GPL'd that you can use to create PDF files on-the-fly with in Windows is called PDFCreator. Basically it is a PostScript to PDF convertor. Installed it just appears as a new printer. When you print to PDFCreator it prompts you where to save the resulting file. The file created also has selectable text.
"God fights on the side with the best artillery." - Napoleon, Marshal of France - speaking truth to power
From what I gathered reading the article... this *technology* is about controlling the glossyness of certain areas on a printed page. I don't think it involves running printed pages back through the printer... instead words/images would be differentiated by their gloss relative to the flat ink surrounding them. Looking straight at a matt photograph one would see nothing unusuall but looking at an angle one could make out shiny text, the degree of gloss is controlled, hence the "invisible to the eye" option.
just my 2c,
-ry
No wonder Xerox is struggling. While other companies are busy developing new products Xerox techs are destracted by shiny objects.
"Oooh, shiny!!!"
"Did you read the article?"
You must be new to Slashdot. Welcome!
Cannot be copied, scanned or printed using conventional equipment
Use conventional toners and papers
Doesn't this technology just make it EASIER to print these holograms?
If you don't have sufficient access to the source that produces the hard copy (for instance, a locked PDF), how are you going to get a hard copy without the Glossmark on it? 'Cause, if all you've got is the hard copy with the hologram thing which can't be effectively scanned or copied, you're a bit stuck. (Unless you wanted to re-create the document, or re-create the Glossmark and hope people will think anything with it on it is authentic.)
mrg
First, this is nothing like a hologram. (Reporter: This is shiny, holograms are shiny, this must be a hologram.)
When you print continuous tone images with specific ink colors, you have to lay down tiny dots that cover, e.g. 30% of the paper with cyan, 20% magenta, 10% yellow, 15% black. The inks are then fixed in some way: heating, rolling, burnishing or whatever--details vary based on printing technology.
If you put down the ink so that the cyan and yellow dots are: separated by a small gap; or touching each other; or piled up on top of each other; you will get different print characteristics.
It may be e.g., that when wax-based ink drops are piled on top of each other, the burnishing gives it a glossy texture, while the same amounts of inks distributed in separate dots gives a matte finish. (This is just an example based on absolutly no specific knowledge.)
Postscript and other printer control languages are sufficiently expressive that the software can control where the ink dots go. This lets the glossiness be controlled.
This posting is probably a DMCA violation.
They actually say explicitly that it's with color laserjets in the article, but I suppose reading the article would be too much to ask of the poster or the various moderators who moderated this up?
. . . can use ordinary office printers to superimpose a glossy image on an ordinary printed document in a way that can't be photocopied or otherwise easily reproduced."
Unless you have an ordinary office printer.
The point is not that this enables forging.
What it does is provide a much cheaper means for everyday users to produce gloss-watermarked documents that are much harder to forge casually.
Yes the same technology can be used to produce gloss-watermarks for forging, but would require a much harder set of steps (the gloss-watermarks claim to be unscannable). The one down side, gullible people might accept gloss-watermarked documents without question, as proof of authenticity, just because it has a watermark.
Letter To Iran
hope people will think anything with it on it is authentic
Considering the graphic they show is a ticket, and considering the care given by your typical ticket taker who is probably ripping 5000 tickets that day, my guess is that they WOULD think it authentic. I mean, it's not like they have microscopes out there at the gate...
My guess is that MOST applications that MAY want to use this as a security measure wouldn't be putting the documents under intense scrutiny.
I've had this idea rolling around in my head...though this technology is not strictly speaking a holograpm (but is some type of OVD--optical variable device.)
I would market with one little company, a special type of thin transparent paper that could go through a regular ink jet (and with a special ink jet cartridge) that could create high quality holograms. Sold obstensibly for "document security" their may purpose would be for faking the holograms on driver's licenses.
Then I would have another company selling really expensive equipment to banks and bars to detect fake licenses. The expensive equipment isn't all that sophisticated, all it would be doing is picking up a random particle that was specifically embedded into the holographic paper sold by the first company above. For some reason though, it also would reject some Lousiana driver's licenses.
Then, I change the holographic tansparent sheeting, put a new random particle in it...and then market a very expensive upgrade from the other company to the banks, bars, et cetera, so that they can now detect the new paper (but not detecting the old paper as well, Lousiana licenses work better, but now Alberta licenses are rejected.)
I would do this until someone figueres it out and requires me to take my cash and assortment of women out of the country.
Finally I can replace this candle wax and stamp sealer from the 1500's!
"The problem with socialism is eventually you run out of other people's money" - Thatcher.
What are you saying, Xerox didn't invent those, App^h^h^h MS did! Just ask their PR department, they'll set you straight, and send you a free copy of 'MS History v3.0 - This time we got it right'.
-Charlie
(Yes, once again, sarcasm, I do know my history).
This is how Xerox plans to overtake the conterfeiting industry.
Slashdot Eds Link Anonymous Posts With Logged Posts
They Are Vermin Feeding On Each Other's Feces.
I Hate \.
So... they exploit their own printers... and suddenly make a new product out of it?
Isn't that called marketing?
"Hundreds of millions of dollars have been spent on efforts to encrypt or otherwise protect content against people who might make perfect digital copies with a computer or other device. But little of this technology has been able to do anything about decidedly easy methods of reproduction such as photocopying a hard copy of a document, or taping a song as it comes out of a stereo's speakers. "
.. "Protect content against people"?
Criminal masterminds, beware! In the future, diabolical acoustic tricks inside the loudspeaker will make every attempted microphone-against-speaker based tape-recording of the song come out as an extra evil rendition of "Don't Copy that Floppy".
And what's this
Please, people, someone think of the content!
Reading this I remembered the dificult I had scanning some glossy wine labels. It's much like to scan a mirror.
I'll recreate the holograms cheaply, using this printing technology. No one will be able to tell the difference.
http://ablegray.com
It's terabyte, not "terrabyte" dammit.
And as RAM is basically always measured in binary multiples, one should use the binary prefixes according to IEC 60027-2, so tebibyte would be preferred.
Copying black on black is an old 80's xerox art technique; the effect is the same. The 2nd copy has different reflectivity properties than the 1st copy underneath. You can only see the 2nd image if you look at it skewed. Examples of this can be seen (my work) at MOMA special collections library or at the School of the Art Institute special collections library. I actually wrote a letter (1986) to xerox suggestioning the use of transparent toner for document security.
yes I used to work at Kinkos,
john r
catalyst komics
A lawyer that can use the "it was a bug in the printer" defense to successfully get you off: Priceless.
Paying that lawyer with the same card: even more priceless.
You might get the paper by bleaching one-dollar bills, but you damn sure ain't getting the press.
Yes, but then if you re-read the post you're replying to, that's exactly what he said. It's not the artwork that's stopping him from making a good counterfeit, it's the lack of ability to obtain the machinery to do it.
Making something "authentic" is relatively easy when the machinery is in every store. The Xerox machine can't make anything not easily counterfeited because everybody could get one cheaply and affordably, and then simply print out their own Glossmark crap.
- Give a man a fire and he's warm for a day, but set him on fire and he's warm for the rest of his life.
Unfortunately, thanks to 1201 (b)(1)(B)--i.e., the DMCA---you'll have to figure out the "how" part on your own.
Yeah, right.
How does this affect porn?
"tebibyte would be preferred" ... by 3 people.
Sorry, "tebibyte" is not a word.
?
The document can still be reproduced, the point is -- the pseudo-hologram can not. If the document is missing the pseudo-hologram, you know that it has been duplicated.
The security would lie in the fact that it looks different from different angles. So if you scan it/photocopy it/whatever, you only get *one* angle on it, and thus there is no easy way to get a digitized version of the watermark to feed to that other Xerox printer.
I hereby place the above post in the public domain.
Oridnary scanners can only get the color (RGB) in their scans. The Glossmark can vary the glossy-ness of the print and the specular reflections cannot be captured through a scanner and reproduced. Ever scanned a glass plate or a mirror? :D
Moral of the story: Scanners can only capture the diffuse color correctly.
It would be pretty simple to duplicate the glossmark.
1. Scan the document.
2. Then take a picture of it:
a. Put the item on the table face up.
b. Turn a desk lamp so that it shines on the paper at an angle.
c. Take picture with digital camera at an angle.
3. Load both images into the GIMP.
4. Use reverse-perspective tool in the Gimp to fix up digital camera picture.
5. Scale digital picture image to the same size as scanner image.
6. Copy digital picture into scanner image as a separate layer.
7. Set the top layer to difference.
Wallah! Instant bitmap of the pseudo-watermark. Sounds like a lot to steps do, but thety take only a few minutes to accomplish.
"It's already so easy to counterfit U.S. money, using Xerox printers. "
Modern Xerox copiers/printers will not allow you to scan/print money, gold, and other related media.
Stock is an issue as well, not much point with currency, you must not know these things.
Coupons, passes (movies, shows, etc), and IDs (non-reflective) are pretty much the only things "of street value" that can be copied easily.
Most copier mfgs make copying things such as currency impossible. Maybe you were thinking of Monopoly money.
I just finished writing a book on the subject actually, "Counterfiting for Dummies", check it out. =)
How on earth does this provide security? If its only ment to be looked at by the human eye then you dont need to copy it perfectly, you just look at the real document, draw/type what you see and hope that whoever you show the forged printout to will think its geniune. If on the other hand it can be detected by a scanner, then whats the point? and if your eyes can see it at an angle, whats to stop you pointing a digital camera at it and adjusting the angle/lighting until you could see it on the camera? I have RTFA and i have no idea what they are on about, what is the point of this? can someone explain?
This comment does not represent the views or opinions of the user.
Since I haven't seen this in person yet, I can't guarantee I could copy it. But from the available information it looks trivial to duplicate. If I can see a difference in the gloss with my eyes then I can take a photo or scan that difference. Maybe not if it were a hologram, but this doesn't look holographic at all. Once I have the gloss part scanned I can use a difference filter over the regular image, leaving me a clean channel of just the gloss. How to apply a glossy mask? Well making a conventional mask and spraying it with acrylic clear coat spray paint comes to mind. Or just loading one of my ink-jet ink cartridges with clear coat glossy ink jet ink (yes, it's available). So it looks kind of difficult for the average Joe to duplicate, but dead easy for anyone with digital printing experience.
But only because some morons didn't get that the SI-prefixes were Base 2 when it came to storage capacity and Base 10 when it came to bandwidth. I mean, of course we have subtle little secrets and speak in code words, we're computer geeks goddammit.
The kibi-, gibi-, and tebi- are the new abominations (imo) used to describe the old-school Base 2, thus a kibibyte is 1024 bytes (whereas a kilobyte was 1024 bytes in the "good old days"), and now a kilobyte it 1000 bytes.
cat
why can't I, or someone else, reproduce the psuedo-hologram?
even if it's a fairly intricate graphic reproducing on a computer something that would pass for the original is typically fairly easy
to quote Xerox: Can be produced by existing Xerox printing solutions
So anyone with the correct Xerox printer now has the ability to create a close copy of your document, complete with the pseudo-hologram
Casual Games/Downloads
They call it 'Windows'.
It's a way to print a pattern or message out of a series of shiny/non-shiny patches on the paper. But if you look at it from top-down (the same way a scanner views a document) the changes are not visible.
So, you could print "Look for the logo between these lines: ---| @@ |--- " and if the person reading it doesn't see it, they can hopefully suspect a forgery.
This would be good for coupons, low-value stuff. You wouldn't want to print lots of real money relying only on this trick, because it'll get defeated soon enough (probably by some kid with a polarizing plate on his scanner and too much time on his hands.) But it's still a neat idea.
John
"terrabyte" -- ok, I had too much coffee. So sue me.
"tebibyte" -- Not unless I acquire an adult-onset hairlip. Terabyte is just fine, thanks, and already in general use. You'd be fighting an uphill battle on that one.
*** *** You're just jealous 'cause the voices talk to me... ***
You are absolutely correct, except for one part.
;)
1000 is a grand.
And just out of curiousity, who exactly popped up and decided that a kilobyte was no longer 1024 bytes? I never heard of this.
*** *** You're just jealous 'cause the voices talk to me... ***
And just out of curiousity, who exactly popped up and decided that a kilobyte was no longer 1024 bytes? I never heard of this.
The IEC. It isn't exactly a redefinition, since AFAIK kilobyte wasn't officially defined as a unit by many standards organizations. Kibi- and friends were coined because standards bodies are by their nature incredibly pedantic, so overloading the SI prefixes was out of the question.
There was an alternative proposal to prefix binary units with 'di-', so 1024 bytes would be a dikilobyte. In writing a subscript '2' would be inserted after the prefix, giving you something like K2b. You were explicitly allowed to keep saying 'kilobyte' in conversation. This system is vastly superior for any number reasons, which is why it wasn't adopted.
As far as I know 'byte' is still undefined, so while a one KiB is definitely 1024 bytes, no one can say how many bits it is.
If you're in the right location, Xerox will sell you a bigger, 100 ppm 4 color printer.
Mebbe it's urban legend, so take it with a grain of salt... Part of the story I read [on this Internet thingy] was someone made and ethernet card that clocked at 10,485,760 bits per second, and didn't figure it out until the product was out of the lab.
cat
Yes, it is out of the question to overload the same prefix to mean both 1000 and 1024. A 2.4% error doesn't sound so bad, but once you get up to gigabytes or terabytes the gap between the two widens (there is nearly a 10% gap between terabyte and tebibyte).
Consider that even in the computing field there are many things measured in the decimal units rather than binary - for example Fast Ethernet is 100 megabits per second.
You're right about 'byte'; truly pedantic documents (like international standards) say 'octet'. On the whole, there's no good reason to keep quoting sizes in bytes; most computing devices do not have 8-bit registers or buses, and a single character does not necessarily fit in 8 bits (if it ever did; ASCII is 7-bit). It would make more sense just to use the bit as measure of information and give disk sizes in terabits, and so on.
-- Ed Avis ed@membled.com
Yes, it is out of the question to overload the same prefix to mean both 1000 and 1024. A 2.4% error doesn't sound so bad, but once you get up to gigabytes or terabytes the gap between the two widens (there is nearly a 10% gap between terabyte and tebibyte).
;-)
I don't think it's a problem in practice, though. You can usually figure out which 'kilo' is intended from the context. Right now I think 'kibi' is probably more confusing, just because no one has any idea what it is.
Anyway, one of my goals in life is to make things difficult for SI people. My system has three different 'hundreds' and I like it that way
for example Fast Ethernet is 100 megabits per second.
Units involving bits almost always use the normal SI meanings, so that's fairly unambiguous. Ditto data transmission and storage.
You're right about 'byte'; truly pedantic documents (like international standards) say 'octet'.
Or like ISO C, they punt.
On the whole, there's no good reason to keep quoting sizes in bytes; most computing devices do not have 8-bit registers or buses
Most machines are byte-addressable though, and it happens that the biggest desktop architecture around does have (logical) octet registers. I wouldn't mind going back to using the 'word' as the standard measure, but only if byte-addressing was removed.
a single character does not necessarily fit in 8 bits
Aside: A funny thing occurred to me after I wrote my post. A system using pure UCS-encoded Unicode would probably end up with 32-bit bytes in ISO C because of the way 'byte' is defined. A side-effect would be that there could be no data types smaller than 32-bits.
It would make more sense just to use the bit as measure of information and give disk sizes in terabits, and so on.
I understand that hardware people do that. The only problem is that you end up with huge numbers, and have to divide them all over the place to figure out how much memory you actually have available.
It isn't always clear from context; for example if you have a ten-gigabyte disk and one 'gigabyte' of main memory, how many memory images can you write? Even if the ambiguity is tolerable for people who've worked with computers for a long time, this doesn't mean it is a good system. And that's without even considering the mutant hybrid units as used in a '1.44 megabyte' floppy.
People may have no idea what 'kibi' is, but at least they _know_ that they do not know. That's probably better than not knowing that you don't know that 'kilo' is different to what you expect. Of course if you are talking approximate quantities you can use kilo, etc all the time: 'This computer has about a gigabyte of RAM'.
On byte-addressing: fair point.
-- Ed Avis ed@membled.com
It isn't always clear from context; for example if you have a ten-gigabyte disk and one 'gigabyte' of main memory, how many memory images can you write?
;-)
Well, discs are always SI for marketing reasons. On the other hand, most of the software I use reports disc sizes in blocks rather than bytes, which at least makes conversion easy, though it took me forever to figure out how large a block was at first.
You're right that there's a problem, since the system is practically random. I have no idea what the capacity of a CD is, for example.
People may have no idea what 'kibi' is, but at least they _know_ that they do not know. That's probably better than not knowing that you don't know that 'kilo' is different to what you expect.
Well, being the product of a modern public education system, I don't expect anyone to know what 'kilo' means either
I don't really have a problem with a new set of prefixes, I just don't like the one chosen. I'm quite happy using the 'dikilo' prefix; if nothing else, it sounds cool.
On byte-addressing: fair point.
Unfortunately.