I'm just going to address two points, because doing so is simplistic and I don't have time at the moment to correct the deeper underlying errors in parent's suppositions.
If we start to see open-source software take on the sort of role traditionally filled by Microsoft, eg partnering with Intel and other large hardware vendors to drive development of a PC (or similar) platform, I'll agree that software copyrights aren't needed to ensure technological progress.
I listened to the entire presentation. Moglen does a good job of demonstrating that the OLPC initiative is of intense interest to major hardware manufacturers precisely because it is exploring ground-breaking technologies. So this is happening. Another example is Google's interactions with hardware manufacturers to push the development of CPU cluster technologies (where the underlying driver is the Linux beowulf paradigm).
I find that I can only agree with parent post on this point to the extent that I am ignorant of what is actually going on in the world. And I find that I am not sufficiently ignorant to be able to agree with parent's point at all.
Linux users are in a sense actually free-riders, benefiting from Microsoft's platform development investment...
Well, if we ignore that Linux is actually based on Unix as simplified through Minix, and is therfore completely independent of the evolution of DOS to Windows. Of course to really appreciate how much we have to ignore, one has to know that DOS was itself built on the earlier efforts of CPM and the Apple II firmware to model, as best it could be done in the late 1970s, the file management and other key components of early Unix. And to fully appreciate the state of ignorance we would need to achieve to support parent's statement, we need to also know that the Windows GUI is not a Microsoft original: it is based on the work that Microsoft did under contract to IBM (and therefore IBM's design specifications) to recreate the experimental GUI interfaces that Xerox PARC had been developing on its Unix machines. The various GUIs that run under Linux are also derived extensively from the Palo Alto Research Center work, in fairly direct lineage. Again, there is nothing in this corner of the Linux world that is derived from Windows development.
But the power of ignorance is great indeed. By the power of ignorance alone, Microsoft is perceived as having brought computing to the masses, when in fact Microsoft has basically been a very good marketeer of derivative works that are usually of lower technical quality (but cheaper development) than the original material that had been "embraced and extended".
I guess I'll try to encourage the adoption of ODF while waiting for the first edition of "Undocumented Office Open XML" to hit the bookstores. Not much point in taking on OOXML before the critical tool to make it work is available.
It feels like deja vu all over again. "DOS ain't done until Lotus won't run"; Win Quattro Pro and Win Word Perfect that can't compete with Win Excel and WinWord because the MS developers didn't limit themselves to the authorized Windows API; etc, etc.
When a single programmer can steal the elections, it's because the electronic voting system is poorly designed.
Why is it so hard to understand that sometimes the best technology might be paper based rather than electronic? NIST has just finished saying that ANY design for an electronic voting system will be a bad design because it is not possible to meet all the critical specifications with an electronic method. Could it be that the problem in understanding this is that we've got too many people who only have hammers and therefore think every problem is a nail?
There are a number things computers will never be any good at. They make lousy bobsleds. They really aren't very good as boat anchors. And NIST has just said they aren't any good as ballot boxes, either. Seems simple enough.
...how on earth can someone code so sloppily that a WORD PROCESSOR has a serious security exploit?!
Shit happens.
The more significant question is how on earth could an exploit like this manage to get by Quality Assurance for so many years?
The answer is that the Coding For Profit paradigm necessarily imposes a limitation on quality assurance since QA is an expense that must be charged against profits.
A viable workaround is to Code For Free under one of the open source licenses where you can nurture a community of bug-hunters and developers who provide good quality assurance for free. You generate your profits from other aspects of the software business, such as service. IBM and Redhat are doing pretty well with this approach. Until recently I would have mentioned Novell here too, but now there's some doubt about whether Novell will survive what might prove to have been a fatal error.
Wake up little SUSE! The movie wasn't so hot.... but I digress.
I expect that in the next few weeks Microsoft will offer as a workaround a free plug-in that will convert all documents to its new ECMA approved standard. MS will point to Novell as an alternate supplier (therefore avoiding immediate monopolistic legal hassles). MS will point out that MS Office 2007 will be immune to this exploit, so all businesses really need to do is to install the free plug-in and begin migrating their documents to the new format. Which will be supported by Novell's version of OpenOffice, btw, no sneaky deals here, huh?
I'd rather kick in the nuts the guy who takes advantage of these 'exploits'. They cease to be exploits when there are none willing to exploit them.
So you think that upgrading to Human version 2.0 is the answer? I don't think any of us have the wetware to support that. I think we're stuck with the current ugly bags of mostly water for at least a few more years.
So this heat ray device is for use against disorganized mobs rather than organized protestors.
Why is it better than tear gas for this purpose? Is it because tear gas leaves clear signs of its usage that can be videotaped during or immediately after an event, while the heat ray leaves no evidence of its use?
If it is better than tear gas, does that allow whoever is calling the shots a wider scope of action than tear gas would? Is this a good thing, if the scope of action is expanded from dispersal of crowds that threaten the peace to dispersal of crowds that threaten to delay the Hummer from getting back to base in time for the evening movie?
How will a detail of US soldiers fair when an insurrectionist hits them with a blast from a "liberated" heat ray device? Would this leave them more vulnerable to a second punch with a machine gun or RPG? Or is this heat ray device for use in a fantasy world where the bad guys simply aren't allowed to get hold of the fancy weapons?
At this point I think the Pentagon has spent $40 billion on yet another boondoggle, and that they know it, and that is most of the reason why this thing has been developed in secret. The only strong rationale for developing this weapon is that it would allow the US forces to disperse crowds without the telltale evidence that tear gas, water cannons, and rubber bullets leave behind. In short, I think it is probably an inferior method of crowd control that is favored only because it could be used with great impunity, since it would be almost invisible to the media.
I just tried Yahoo Search. I haven't seen the ad for IEv7 in half a dozen different searches (using Firefox v2.0 as my browser).
Could this be because I have already installed IEv7? Or because I was using FF v2.0 and therefore not a likely candidate for upgrading?
In any event, I favor anything that would encourage people to migrate from IEv6 to something that was more secure and complied better with web standards, even IEv7. We'll all be better off for that.
Nope, but I do try to rejuvenate my neural nets on a regular basis by thinking young and sassy thoughts, so thanks for the compliment!
I am not naive about business operations. IBM's flocks of lawyers aren't kept in storage on the shelf in between SCO-sized lawsuits. They earn their keep doing more mundane activities. The SCO suit caused IBM to revise the priorities of the legal department for the duration. There is considerable cost in doing that kind of revision. Now that the activities related to the SCO suit are beginning to wind down, IBM can either disperse its lawyers back to the ordinary activities of quieter times, or they might look around to see if there are any other problems that this team could address before it is dissassembled. My guess is that IBM is doing the latter, since the costs associated with taking the current team apart only to put it back together in a couple of years might be avoidable by taking some actions now.
If IBM does choose to go on a legal offensive, then I think two choice targets they might be considering are the sorry state of software patents and the absurdities of EULAs. These both pose significant problems for the healthy ecosystem business model that IBM has found to be so successful in the last decade. Th e fact that clearing up the law in either of these areas would also cause major problems for a company that has betrayed them in the past and is rabid in its adherence to a business model that promotes destruction of adversaries rather than cooperation within a community would be merely an unintended consequence.
Thank you for this opportunity to further clarify my suggestions. I look forward to your next post-- I'll do my best to play Gracie Allen to your Geoge Burns' straight man lines.
You just have to trust that if some corporation (they almost always are) takes exception that you can demonstrate identify who contributed what so you can pass the blame rather than taking the rap yourself.
The concern here is probably not so much who takes the blame than for keeping the project alive. A finding that the contributor screwed it up doesn't protect the project from the loss of his code. If the project is going to survive, it should probably be the one to defend its code.
If you've got documentation that clearly asserts that the contributor had the right to make the contribution, any legal action is going to be preceded by an exchange of mails over several months:
Ink Inc: the use of Purple Prose in the commentary of crucialBigPatch violates our copyright and we demand that you either pay us ONE MILLION DOLLARS or remove the code immediately. Oooh, I mean ONE BILLION DOLLARS.
Project: we have a legitimate license to use the code in crucialBigPatch given to us by Ima Scammer and here is a copy of that. So we are not infringing on your copyright and will neither pay you nor remove the code.
Ink Inc: Ima Scammer did not have the right to give that to you because of blahblah. Please respect our copyright because we do not want to have to take you to court. We will accept your payment of ONE BILLION DOLLARS PLUS ACCRUED INTEREST FROM THE DATE OF FIRST NOTIFICATION.
Project: We have given your last letter our full attention and we find that blahblah does not apply because of yaddayadda. Please don't bother us any more. You are irritating.
Ink Inc: You are wrong in thinking that yaddayaddda can in any way cancel blahblah in this case. Unless you comply and pay our licensing fee plus accrued interest, we will take you to court.
Project: Know you this: we will have crucialBigPatch2 that uses our new Euphuistic Expression commentary in production long before the court date. That means there will be no infringement and your case will be thrown out immediately. Rather than receiving the cash you had hoped to extort from us, you will be paying attorney fees and filing costs for nothing. We fart nasty smellies in your general direction. Now go away; you are bothersome.
So that is one of the practical reasons for having project contributors sign that they own the code and are willingly contributing it to the project. It can keep the project out of court long enough so that a workaround for any code that might become a problem can be put in place.
If from the beginning the project is managed with contingency planning for this kind of modular replacement, then the whole incident can be a low stress exercise in troll - countertroll relations. You just have to accept that Ink Inc and their ilk might pull you into a mud puddle from time to time, and recognize that since they've already dirtied up your project, you might as well enjoy the opportunity to play the kid and do some splashing back at them. You can wash up after it is all done.
I am convinced that SCO figured IBM would just pay them to go away, everyone would make a quick buck, and that would be that; it would probably have been cheaper than this protracted court battle. But IBM thinks longer term. Having put up this kind of fight, do you think anyone else will ever sue them over Linux IP? This fight will make Linux largely lawsuit-proof. [emphasis added]
I couldn't agree with you more.
Looking ahead a little bit to what might happen at the end of the SCO debacle... IBM has put together a legal bulldozer at no small expense to improve their business climate for the next 25 - 50 years when they could have just bought out Darl with pocket change. I'm guessing that IBM is prolly now looking around to see if this bulldozer can be used to make other improvements before it is dismantled. And when I try to look at the world through IBM's eyes, I see three targets that might be worth taking a run at:
Software patents. IBM has been quietly putting a number of its software patents into a kind of protective escrow situation where they cannot be directly used to generate income. My guess is that IBM is no lover of current patent law and prolly sees the whole body of existing software patents as being rather like a toxic wasteland, and the patent process being a drain on their revenues. It might make sense to use the legaldozer to do some environmental clean up in this area. There are legislative avenues that could and prolly are being pursued, but since there is this shiny yellow 40,000 horsepower engine with a big old blade on the front all warmed up and ready to go, maybe attempting some change through the judicial system would be worthwhile.
EULAs and the FUD that surrounds them. I don't believe IBM is using EULAs any more (if they ever did) and IBM prolly sees these as detrimental to the business climate they are attempting to nurture. Perhaps the IBM legaldozer could be used to remove these unsightly fences from the landscape and return the fenced off areas to the commons where Linux and IBM offerings flourish.
The whole digital copyrights arena: I'd really like to see IBM take this on, but I doubt that will happen. The other two possibilities are directly within the scope of IBM's mission and core business practices. Cleaning up the digital rights garbage pit seems like much more of a stretch.
In any event, SCO encouraged IBM to develop a powerful legal team. I think it would make a lot of business sense for IBM to look at targets worthy of this team's attention before taking it apart.
There is also the distinct possibility that SCO and BSF (their lawyers) will be punished for bringing a case before the court that has zero merit. It is a frivolous case and lawyers can be debarred for that kind of conduct.
That is an outcome I would very much like to see (disbarrment of the lawyers). The lawyers involved should be disbarred and they should be charged and found guilty of felony conspiracy (as well as the corporate officers of SCO). They should never again be allowed to hold any position of public trust, not in the law, not as bank tellers, not even as a call center customer service representatives. The law firm should be broken up, its offices razed, and the rubble should be sown with salt.
If lawyers in this country were required to live up to their responsibilities as Officers of the Court, we would all be better off. This case is proving to be such an egregious abuse of the legal system that action must be taken against the lawyers involved, since to allow them to walk away would shatter the foundation of the rule of law beyond this society's ability to repair it. That would mean it would become necessary for many of America's people to dissolve the political bands which have connected them with others and to assume among the powers of the earth, the separate and equal station to which the Laws of Nature and of Nature's God entitle them. Thank you Mr Jefferson. No one in two hundred thirty years has said that better than you did.
The law was never intended to be a club that you can use in an attempt to extort money from IBM or any other company or person. The law is intended to be a set of rules that is supposed to provide some measure of fairness in the dealings we have with one another. Officers of the Court have a responsibility to uphold that concept of law; those that attempt to make a mockery of the law by participating in a sham like this one should never again be trusted in any measure. Let them earn a living as day laborers for the rest of their miserable years.
What can I say but that in the clear light of early morning I now see the error of my ways? I have learned something here and my thanks go to you and others.
I was confounding the possible values a bit stream can store with its length. Oh my.
I might respond on this thread again in a couple of days.
I'm just going to take a moment here for a couple of points.
First, I think that you are right in that the math would be the same if two dots placed side by side were used. But I'm not sure of that, and also I felt it was easier to describe one foreground dot in the center of a background than it would have been to talk about it any other way.
Second, you state that bits store states and not combinations, and I won't argue with that one way or the other (though I think it would be sort of fun to see how you would explain that).
The thing is, that's irrelevant. We aren't talking about storing the number in the computer. We're talking about painting it on a piece of paper, where bits and bytes and such don't apply. On paper, with color ink, we've got storage bins that can be in many more states than just binary. Part of the reason is because we can assign meaning to combinations of different colors, and there are a lot of possible combinations. When we bring the number back into the computer we'll want to convert it back into a binary expression, but the constraints of binary expressions don't apply to that number while it is sitting out there on the paper. It can be colorful and concise Out There, even though it becomes colorless and quite long in the computer.
That makes sense. I thought I had stripped out the whitespace, but clearly I missed the first one. I would like to hope that if it wasn't so late at night, I would have realised that 181 didn't make much sense.
779982499796 / 32640 = 23896522.7
0.7 * 32640 = 22848, so the last digit will be the 22,849th in the set
23 896 522 / 32 640 = 732.123836
0.123836 * 32 640 = 4042.00704 the digit to the left of that will be the 4,043rd in the set
the first digit will be the 733rd in the set
So it becomes a 3 digit number in the large base. Once you decide what glyphs you want to use, you can write the number out using the above directions. Just note that the conversion wasn't clean due to the limitations of my 32 bit computer, so the last digit might be off.
The more I think about this, the more I think that Sainul must either have a bunch of question marks in the steps where the conversions between binary and image need to take place, or he's cooked up some remarkably fast and efficient conversion routines. This process has got to be more limited by CPU activity than by the limits of the paper storage.
It is waay past my bedtime and I don't do all nighters very well any more.
From a simplistic point-of-view, if you could print a sheet of paper using an ordinary inkjet with that amount of information encoded on it, then logically that information has also been held in the printer RAM. Certainly my printer does not contain 29GB or whatever...
I think you are approaching the truly interesting part of what Sainul has claimed he has done. Obviously the data is being compressed before it goes to the printer buffer, and expanded after it comes back from the scanner. Converting from binary to images of numerals in some number system with a large base and back again has got to be computationally intensive. I think his breakthrough is not so much in stuffing that much data onto a page as it is in doing these conversions in a reasonable amount of time, without melting the CPU.
That has not been addressed in any of the articles I've seen so far.
Actually I was wrong; it will be the 182nd glyph. I did a one-off error by not counting zero as the first glyph. (the number you gave me converted to 181 decimal when I copy / pasted it into a conversion utility that I found with Google.)
I'll stipulate that your math is correct, and I agree that this is the way the number would need to be represented in a binary computer. But the number itself is independent of the way it is represented. And when we represent it on paper, we are not confined to the same conventions as we use in the computer. So long as we agree on a convention of glyphs for the numerals, we can use any arbitrarily large number base for the expression. It will still be the same number, and anyone who wanted to convert it back to into binary would get the same bitstream.
The number is not its representation in the computer. The number is separate from any particular representation. When it is on paper, it doesn't matter how many bits or bytes it might take to represent it in the computer-- it isn't in the computer. It is on paper. The 3 digit base 32,640 number that is one more than the largest 2 digit number in that system is the same number as when it is expressed as a 127 MB long series of bytes in a computer system.
Is it so hard to understand that non-digital media can sometimes express concepts like numbers in a way that is fundamentally different than the expression within a computer? And that sometimes there could be terrific benefits to using these non-digital expressions?
Give me a set of glyphs for the 32,640 numerals needed and I'll convert that number for you. It will be the 181st glyph in the ordered set.
If you are having trouble wrapping your mind around the concept that numbers exist independently of their representation in a computer, please keep working at it. The numbers themselves are truly independent of any method of representation. Numbers seem to be what inspired Plato to talk about Ideals, and numbers do exist, in a way, independently of any physical reality.
Getting back to the case at hand: so long as there is an agreed on convention of glyphs to express them, numbers can be represented in any arbitrary base. Numerically, there is no difference between a bitstream 125 MB long and its representation as two digits in some number system with a large base. It doesn't matter whether a computer can work directly with a particular number base or not-- that doesn't invalidate that method of representing numbers.
You really do need to climb into the box. You will find it much roomier on the inside than it appears to be from the outside.
Visualize a 3x3 matrix. Consider that it can have two states: solid black or solid white. It can represent a single binary digit. There would be 85,000 of these on the "standard page" defined earlier in this thread.
Now stay with the black and white, but allow the center of the matrix to change independently of the 8 bits that surround it. This provides a foreground and background. The same 3x3 matrix still represents a single digit, but now it is in base 4. It can be all black, all white, black foreground with white background, white foreground with black background.
Increase the number of colors possible from 2 to 16. There are 120 different ways to paint the inside of this box. It still represents a single digit, but that is now a digit in base 120.[1]
When the number of possible colors increases to 256, the number system of this digit becomes base 32,640. There are that many different ways to choose 2 colors from a set of 256 colors.
Take a text file of 10 kilobytes, treat its bitstream as a longish binary number and convert it to base 32,640. It can be represented on paper in 2 (two!) digits. A 125 MB chunk of data can also be represented as 2 digits base 32,640.[2]
So, here we have 2 3x3 matrices for a total of 18 pixels that is capable of storing more than 125 MB of data, and this without even considering the combinations of placing the dot somewhere other than the center, or using more than one dot. 125 MB of storage in 18 pixels seems fair data compression. Even if we make the pixels pretty big so we can use cheaper scanners.
So where do you think the fallacy in establishing limits on paper storage lies? I believe it has something to do with the way that introducing a foreground - background relationship brings combinatorial considerations into the picture. Start using clubs, spades, diamonds and hearts as well as color, and the data compression begins to get pretty good.
I suspect that Sainul's technique might be computation bound: I think the kinds of data compression involved in converting between binary and these high base number systems must be pretty CPU intensive. It certainly doesn't look like its paper bound.
[1]Maybe 240 different ways, and base 240? I'm not sure. But at least 120.
[2]If I can believe my calculator, the 3rd digit would kick in at just over 127 MB:
32,640**2 = 1,065,369,600 bits = 130,050 KB = 127.002 MB
I hope can help you a little bit in gaining that understanding.
There seems to be a continuing failure to recognize that this is not a simple mapping problem.
Within a 3x3 matrix, there is a background color and a foreground color. Two distinct colors chosen without replacement from a palette of 256 colors (limited by what can be easily distinguished by the scanner). There are more than 32,000 ways of selecting these two colors (Google is your friend-- see elsewhere on this thread). Each matrix can take on 32,000+ distinct states. You could think of this as each matrix representing a digit in base 32,000+.
You could take a typical business document and treat it as a long binary number and convert it to base 32,000+ and store it on paper in 2 or 3 digits. The process is easily reversible: read a couple of dozen digits from the paper, convert to binary and split up into bytes then feed the resulting spreadsheet to Excel.
And in all the above we were dealing with just one shape, say a circle of one color on top of a background of a different color. When you start combining in other easily distinguished shapes like squares, stars, etc, then you increase the base of the number system even further.
So how many unique states can you represent in a handful of digits? It all depends on the base of the number system you are using. It seems like Sainul has found an interesting application of an overlooked aspect of number theory.
You do not seem to understand how mathematics works.
How insightful of parent post for stating the obvious! IANAM, nor have I ever claimed to be one.
OTOH, many of the contributors to this thread do not seem to understand how to work the mathematics of this problem. Nor do they seem to recognize that Sainul Abideen could not have gained the recognition he now enjoys if it were so easy to show that his claims were 5 orders of magnitude greater than what they think is the upper limit of the possible. It seems much more likely that slashdot's best mathematicians were screwing it up somehow.
It seemed to me that combinatorial processes had to be involved since the articles about Sainul's work happened to mention (several times) that the mechanism was based on combinations of colors and shapes. While IANAM, I am someone who knows how to comprehend what he is reading.
Since I haven't worked with combinatorial math since my student days, I put forward an estimate that I was sure was low, but which was quite a bit higher than those being bandied about, and I sort of hoped that someone would pick up on this and come through with a description of the correct technique.
Someone did, more or less, but it was not the parent post. Parent post contains some mathematical jokes that whooshed by over my head. Or, well... there is another possibility concerning where parent post's author's head is at, but I won't go there, no sir. That would be insulting.
You calculations simply doesn't make any sense. Now let's compute an upper bound...
No let's not, since before that can be done there has to be an understanding of the underlying mathematics.
Of course my calculations don't make any sense if you haven't yet studied the applicable math. Tell you what, let's look at a couple of easier problems that are very similar. How many unique games of checkers can be played out on a 8 x 8 checkerboard? How many unique games of chess? These ARE similar problems, involving combinations and patterns.
I've talked elsewhere about this on this thread: use "parent" to get back to my first post then re-read the subsequent ones.
How can a square of 9 pixels hold more than 9 * number of colors values?
It would appear that someone has yet to discover the miracle growth of combinations...
This has nothing to do with the size of the matrix. It has to do with the number of unique selections of two colors that can be made from a set of 256 colors.
Go straight to Google and type in "256 choose 2". Then type in "combination math tutorial" and start exploring a new world.
Thanks for the link to weierstrass' Implicity blog. His diagram of the 102 unique two-color patterns is very useful in this context.
So 102 patterns using 2 colors multiplied by the number of combinations of two colors that can be drawn without replacement from a palette of 256 colors... it has been a while since I've worked combinations but I know how to use Google as a brain prosthesis:
So 32,640 * 102 patterns = 3,329,280 possible combinations within each 3x3 matrix. 85,000 such matrices on the 8.5x10 inch sheet yields 2.9*10^11 unique patterns possible on one page. That's a pretty long bit stream. Converting from bits to something understandable yields 33 GB per page.
This unsophisticated technique shows a sheet of paper can hold more than 1500 times the information that the one-bit-per-dot crowd was thinking was the max (22 MB iirc). It is still an order of magnitude lower than the reported achievement of Sainul Abideen-- but I am working as a Resource Support Assistant in a Community College and I don't profess to know much about combinatorial math or pattern recognition. I think it enlightening that Google says that "256 choose 3" gives 2,763,520 unique three-color schemes...
I do, however, know a thing or two about Google and how to use simple resources like it to make the world a little more understandable.
Slashdot Mirror
I'm just going to address two points, because doing so is simplistic and I don't have time at the moment to correct the deeper underlying errors in parent's suppositions.
If we start to see open-source software take on the sort of role traditionally filled by Microsoft, eg partnering with Intel and other large hardware vendors to drive development of a PC (or similar) platform, I'll agree that software copyrights aren't needed to ensure technological progress.
I listened to the entire presentation. Moglen does a good job of demonstrating that the OLPC initiative is of intense interest to major hardware manufacturers precisely because it is exploring ground-breaking technologies. So this is happening. Another example is Google's interactions with hardware manufacturers to push the development of CPU cluster technologies (where the underlying driver is the Linux beowulf paradigm).
I find that I can only agree with parent post on this point to the extent that I am ignorant of what is actually going on in the world. And I find that I am not sufficiently ignorant to be able to agree with parent's point at all.
Linux users are in a sense actually free-riders, benefiting from Microsoft's platform development investment...
Well, if we ignore that Linux is actually based on Unix as simplified through Minix, and is therfore completely independent of the evolution of DOS to Windows. Of course to really appreciate how much we have to ignore, one has to know that DOS was itself built on the earlier efforts of CPM and the Apple II firmware to model, as best it could be done in the late 1970s, the file management and other key components of early Unix. And to fully appreciate the state of ignorance we would need to achieve to support parent's statement, we need to also know that the Windows GUI is not a Microsoft original: it is based on the work that Microsoft did under contract to IBM (and therefore IBM's design specifications) to recreate the experimental GUI interfaces that Xerox PARC had been developing on its Unix machines. The various GUIs that run under Linux are also derived extensively from the Palo Alto Research Center work, in fairly direct lineage. Again, there is nothing in this corner of the Linux world that is derived from Windows development.
But the power of ignorance is great indeed. By the power of ignorance alone, Microsoft is perceived as having brought computing to the masses, when in fact Microsoft has basically been a very good marketeer of derivative works that are usually of lower technical quality (but cheaper development) than the original material that had been "embraced and extended".
I guess I'll try to encourage the adoption of ODF while waiting for the first edition of "Undocumented Office Open XML" to hit the bookstores. Not much point in taking on OOXML before the critical tool to make it work is available.
It feels like deja vu all over again. "DOS ain't done until Lotus won't run"; Win Quattro Pro and Win Word Perfect that can't compete with Win Excel and WinWord because the MS developers didn't limit themselves to the authorized Windows API; etc, etc.
When a single programmer can steal the elections, it's because the electronic voting system is poorly designed.
Why is it so hard to understand that sometimes the best technology might be paper based rather than electronic? NIST has just finished saying that ANY design for an electronic voting system will be a bad design because it is not possible to meet all the critical specifications with an electronic method. Could it be that the problem in understanding this is that we've got too many people who only have hammers and therefore think every problem is a nail?
There are a number things computers will never be any good at. They make lousy bobsleds. They really aren't very good as boat anchors. And NIST has just said they aren't any good as ballot boxes, either. Seems simple enough.
Shit happens.
The more significant question is how on earth could an exploit like this manage to get by Quality Assurance for so many years?
The answer is that the Coding For Profit paradigm necessarily imposes a limitation on quality assurance since QA is an expense that must be charged against profits.
A viable workaround is to Code For Free under one of the open source licenses where you can nurture a community of bug-hunters and developers who provide good quality assurance for free. You generate your profits from other aspects of the software business, such as service. IBM and Redhat are doing pretty well with this approach. Until recently I would have mentioned Novell here too, but now there's some doubt about whether Novell will survive what might prove to have been a fatal error.
Wake up little SUSE! The movie wasn't so hot.... but I digress.
I expect that in the next few weeks Microsoft will offer as a workaround a free plug-in that will convert all documents to its new ECMA approved standard. MS will point to Novell as an alternate supplier (therefore avoiding immediate monopolistic legal hassles). MS will point out that MS Office 2007 will be immune to this exploit, so all businesses really need to do is to install the free plug-in and begin migrating their documents to the new format. Which will be supported by Novell's version of OpenOffice, btw, no sneaky deals here, huh?
I'd rather kick in the nuts the guy who takes advantage of these 'exploits'. They cease to be exploits when there are none willing to exploit them.
So you think that upgrading to Human version 2.0 is the answer? I don't think any of us have the wetware to support that. I think we're stuck with the current ugly bags of mostly water for at least a few more years.
Let me fix that for you:
Yes! Great idea! Just trust all of your internal documents to a different random third party company with no privacy guarantees.
So this heat ray device is for use against disorganized mobs rather than organized protestors.
Why is it better than tear gas for this purpose? Is it because tear gas leaves clear signs of its usage that can be videotaped during or immediately after an event, while the heat ray leaves no evidence of its use?
If it is better than tear gas, does that allow whoever is calling the shots a wider scope of action than tear gas would? Is this a good thing, if the scope of action is expanded from dispersal of crowds that threaten the peace to dispersal of crowds that threaten to delay the Hummer from getting back to base in time for the evening movie?
How will a detail of US soldiers fair when an insurrectionist hits them with a blast from a "liberated" heat ray device? Would this leave them more vulnerable to a second punch with a machine gun or RPG? Or is this heat ray device for use in a fantasy world where the bad guys simply aren't allowed to get hold of the fancy weapons?
At this point I think the Pentagon has spent $40 billion on yet another boondoggle, and that they know it, and that is most of the reason why this thing has been developed in secret. The only strong rationale for developing this weapon is that it would allow the US forces to disperse crowds without the telltale evidence that tear gas, water cannons, and rubber bullets leave behind. In short, I think it is probably an inferior method of crowd control that is favored only because it could be used with great impunity, since it would be almost invisible to the media.
I think I do not like this heat ray very much.
I just tried Yahoo Search. I haven't seen the ad for IEv7 in half a dozen different searches (using Firefox v2.0 as my browser).
Could this be because I have already installed IEv7? Or because I was using FF v2.0 and therefore not a likely candidate for upgrading?
In any event, I favor anything that would encourage people to migrate from IEv6 to something that was more secure and complied better with web standards, even IEv7. We'll all be better off for that.
You must be very young...
Nope, but I do try to rejuvenate my neural nets on a regular basis by thinking young and sassy thoughts, so thanks for the compliment!
I am not naive about business operations. IBM's flocks of lawyers aren't kept in storage on the shelf in between SCO-sized lawsuits. They earn their keep doing more mundane activities. The SCO suit caused IBM to revise the priorities of the legal department for the duration. There is considerable cost in doing that kind of revision. Now that the activities related to the SCO suit are beginning to wind down, IBM can either disperse its lawyers back to the ordinary activities of quieter times, or they might look around to see if there are any other problems that this team could address before it is dissassembled. My guess is that IBM is doing the latter, since the costs associated with taking the current team apart only to put it back together in a couple of years might be avoidable by taking some actions now.
If IBM does choose to go on a legal offensive, then I think two choice targets they might be considering are the sorry state of software patents and the absurdities of EULAs. These both pose significant problems for the healthy ecosystem business model that IBM has found to be so successful in the last decade. Th e fact that clearing up the law in either of these areas would also cause major problems for a company that has betrayed them in the past and is rabid in its adherence to a business model that promotes destruction of adversaries rather than cooperation within a community would be merely an unintended consequence.
Thank you for this opportunity to further clarify my suggestions. I look forward to your next post-- I'll do my best to play Gracie Allen to your Geoge Burns' straight man lines.
You just have to trust that if some corporation (they almost always are) takes exception that you can demonstrate identify who contributed what so you can pass the blame rather than taking the rap yourself.
The concern here is probably not so much who takes the blame than for keeping the project alive. A finding that the contributor screwed it up doesn't protect the project from the loss of his code. If the project is going to survive, it should probably be the one to defend its code.
If you've got documentation that clearly asserts that the contributor had the right to make the contribution, any legal action is going to be preceded by an exchange of mails over several months:
So that is one of the practical reasons for having project contributors sign that they own the code and are willingly contributing it to the project. It can keep the project out of court long enough so that a workaround for any code that might become a problem can be put in place.
If from the beginning the project is managed with contingency planning for this kind of modular replacement, then the whole incident can be a low stress exercise in troll - countertroll relations. You just have to accept that Ink Inc and their ilk might pull you into a mud puddle from time to time, and recognize that since they've already dirtied up your project, you might as well enjoy the opportunity to play the kid and do some splashing back at them. You can wash up after it is all done.
I am convinced that SCO figured IBM would just pay them to go away, everyone would make a quick buck, and that would be that; it would probably have been cheaper than this protracted court battle. But IBM thinks longer term. Having put up this kind of fight, do you think anyone else will ever sue them over Linux IP? This fight will make Linux largely lawsuit-proof. [emphasis added]
I couldn't agree with you more.
Looking ahead a little bit to what might happen at the end of the SCO debacle... IBM has put together a legal bulldozer at no small expense to improve their business climate for the next 25 - 50 years when they could have just bought out Darl with pocket change. I'm guessing that IBM is prolly now looking around to see if this bulldozer can be used to make other improvements before it is dismantled. And when I try to look at the world through IBM's eyes, I see three targets that might be worth taking a run at:
In any event, SCO encouraged IBM to develop a powerful legal team. I think it would make a lot of business sense for IBM to look at targets worthy of this team's attention before taking it apart.
There is also the distinct possibility that SCO and BSF (their lawyers) will be punished for bringing a case before the court that has zero merit. It is a frivolous case and lawyers can be debarred for that kind of conduct.
That is an outcome I would very much like to see (disbarrment of the lawyers). The lawyers involved should be disbarred and they should be charged and found guilty of felony conspiracy (as well as the corporate officers of SCO). They should never again be allowed to hold any position of public trust, not in the law, not as bank tellers, not even as a call center customer service representatives. The law firm should be broken up, its offices razed, and the rubble should be sown with salt.
If lawyers in this country were required to live up to their responsibilities as Officers of the Court, we would all be better off. This case is proving to be such an egregious abuse of the legal system that action must be taken against the lawyers involved, since to allow them to walk away would shatter the foundation of the rule of law beyond this society's ability to repair it. That would mean it would become necessary for many of America's people to dissolve the political bands which have connected them with others and to assume among the powers of the earth, the separate and equal station to which the Laws of Nature and of Nature's God entitle them. Thank you Mr Jefferson. No one in two hundred thirty years has said that better than you did.
The law was never intended to be a club that you can use in an attempt to extort money from IBM or any other company or person. The law is intended to be a set of rules that is supposed to provide some measure of fairness in the dealings we have with one another. Officers of the Court have a responsibility to uphold that concept of law; those that attempt to make a mockery of the law by participating in a sham like this one should never again be trusted in any measure. Let them earn a living as day laborers for the rest of their miserable years.
</rant>
What can I say but that in the clear light of early morning I now see the error of my ways? I have learned something here and my thanks go to you and others.
I was confounding the possible values a bit stream can store with its length. Oh my.
I might respond on this thread again in a couple of days.
I'm just going to take a moment here for a couple of points.
First, I think that you are right in that the math would be the same if two dots placed side by side were used. But I'm not sure of that, and also I felt it was easier to describe one foreground dot in the center of a background than it would have been to talk about it any other way.
Second, you state that bits store states and not combinations, and I won't argue with that one way or the other (though I think it would be sort of fun to see how you would explain that).
The thing is, that's irrelevant. We aren't talking about storing the number in the computer. We're talking about painting it on a piece of paper, where bits and bytes and such don't apply. On paper, with color ink, we've got storage bins that can be in many more states than just binary. Part of the reason is because we can assign meaning to combinations of different colors, and there are a lot of possible combinations. When we bring the number back into the computer we'll want to convert it back into a binary expression, but the constraints of binary expressions don't apply to that number while it is sitting out there on the paper. It can be colorful and concise Out There, even though it becomes colorless and quite long in the computer.
That makes sense. I thought I had stripped out the whitespace, but clearly I missed the first one. I would like to hope that if it wasn't so late at night, I would have realised that 181 didn't make much sense.
779982499796 / 32640 = 23896522.7
0.7 * 32640 = 22848, so the last digit will be the 22,849th in the set
23 896 522 / 32 640 = 732.123836
0.123836 * 32 640 = 4042.00704 the digit to the left of that will be the 4,043rd in the set
the first digit will be the 733rd in the set
So it becomes a 3 digit number in the large base. Once you decide what glyphs you want to use, you can write the number out using the above directions. Just note that the conversion wasn't clean due to the limitations of my 32 bit computer, so the last digit might be off.
The more I think about this, the more I think that Sainul must either have a bunch of question marks in the steps where the conversions between binary and image need to take place, or he's cooked up some remarkably fast and efficient conversion routines. This process has got to be more limited by CPU activity than by the limits of the paper storage.
It is waay past my bedtime and I don't do all nighters very well any more.
From a simplistic point-of-view, if you could print a sheet of paper using an ordinary inkjet with that amount of information encoded on it, then logically that information has also been held in the printer RAM. Certainly my printer does not contain 29GB or whatever...
I think you are approaching the truly interesting part of what Sainul has claimed he has done. Obviously the data is being compressed before it goes to the printer buffer, and expanded after it comes back from the scanner. Converting from binary to images of numerals in some number system with a large base and back again has got to be computationally intensive. I think his breakthrough is not so much in stuffing that much data onto a page as it is in doing these conversions in a reasonable amount of time, without melting the CPU.
That has not been addressed in any of the articles I've seen so far.
Why will it be the 181st glyph?
Actually I was wrong; it will be the 182nd glyph. I did a one-off error by not counting zero as the first glyph. (the number you gave me converted to 181 decimal when I copy / pasted it into a conversion utility that I found with Google.)
I'll stipulate that your math is correct, and I agree that this is the way the number would need to be represented in a binary computer. But the number itself is independent of the way it is represented. And when we represent it on paper, we are not confined to the same conventions as we use in the computer. So long as we agree on a convention of glyphs for the numerals, we can use any arbitrarily large number base for the expression. It will still be the same number, and anyone who wanted to convert it back to into binary would get the same bitstream.
Now its long past my bedtime.
Nope. You just don't get it.
The number is not its representation in the computer. The number is separate from any particular representation. When it is on paper, it doesn't matter how many bits or bytes it might take to represent it in the computer-- it isn't in the computer. It is on paper. The 3 digit base 32,640 number that is one more than the largest 2 digit number in that system is the same number as when it is expressed as a 127 MB long series of bytes in a computer system.
Is it so hard to understand that non-digital media can sometimes express concepts like numbers in a way that is fundamentally different than the expression within a computer? And that sometimes there could be terrific benefits to using these non-digital expressions?
Give me a set of glyphs for the 32,640 numerals needed and I'll convert that number for you. It will be the 181st glyph in the ordered set.
If you are having trouble wrapping your mind around the concept that numbers exist independently of their representation in a computer, please keep working at it. The numbers themselves are truly independent of any method of representation. Numbers seem to be what inspired Plato to talk about Ideals, and numbers do exist, in a way, independently of any physical reality.
Getting back to the case at hand: so long as there is an agreed on convention of glyphs to express them, numbers can be represented in any arbitrary base. Numerically, there is no difference between a bitstream 125 MB long and its representation as two digits in some number system with a large base. It doesn't matter whether a computer can work directly with a particular number base or not-- that doesn't invalidate that method of representing numbers.
You really do need to climb into the box. You will find it much roomier on the inside than it appears to be from the outside.
Visualize a 3x3 matrix. Consider that it can have two states: solid black or solid white. It can represent a single binary digit. There would be 85,000 of these on the "standard page" defined earlier in this thread.
Now stay with the black and white, but allow the center of the matrix to change independently of the 8 bits that surround it. This provides a foreground and background. The same 3x3 matrix still represents a single digit, but now it is in base 4. It can be all black, all white, black foreground with white background, white foreground with black background.
Increase the number of colors possible from 2 to 16. There are 120 different ways to paint the inside of this box. It still represents a single digit, but that is now a digit in base 120.[1]
When the number of possible colors increases to 256, the number system of this digit becomes base 32,640. There are that many different ways to choose 2 colors from a set of 256 colors.
Take a text file of 10 kilobytes, treat its bitstream as a longish binary number and convert it to base 32,640. It can be represented on paper in 2 (two!) digits. A 125 MB chunk of data can also be represented as 2 digits base 32,640.[2]
So, here we have 2 3x3 matrices for a total of 18 pixels that is capable of storing more than 125 MB of data, and this without even considering the combinations of placing the dot somewhere other than the center, or using more than one dot. 125 MB of storage in 18 pixels seems fair data compression. Even if we make the pixels pretty big so we can use cheaper scanners.
So where do you think the fallacy in establishing limits on paper storage lies? I believe it has something to do with the way that introducing a foreground - background relationship brings combinatorial considerations into the picture. Start using clubs, spades, diamonds and hearts as well as color, and the data compression begins to get pretty good.
I suspect that Sainul's technique might be computation bound: I think the kinds of data compression involved in converting between binary and these high base number systems must be pretty CPU intensive. It certainly doesn't look like its paper bound. [1]Maybe 240 different ways, and base 240? I'm not sure. But at least 120. [2]If I can believe my calculator, the 3rd digit would kick in at just over 127 MB: 32,640**2 = 1,065,369,600 bits = 130,050 KB = 127.002 MB
I hope can help you a little bit in gaining that understanding.
There seems to be a continuing failure to recognize that this is not a simple mapping problem.
Within a 3x3 matrix, there is a background color and a foreground color. Two distinct colors chosen without replacement from a palette of 256 colors (limited by what can be easily distinguished by the scanner). There are more than 32,000 ways of selecting these two colors (Google is your friend-- see elsewhere on this thread). Each matrix can take on 32,000+ distinct states. You could think of this as each matrix representing a digit in base 32,000+.
You could take a typical business document and treat it as a long binary number and convert it to base 32,000+ and store it on paper in 2 or 3 digits. The process is easily reversible: read a couple of dozen digits from the paper, convert to binary and split up into bytes then feed the resulting spreadsheet to Excel.
And in all the above we were dealing with just one shape, say a circle of one color on top of a background of a different color. When you start combining in other easily distinguished shapes like squares, stars, etc, then you increase the base of the number system even further.
So how many unique states can you represent in a handful of digits? It all depends on the base of the number system you are using. It seems like Sainul has found an interesting application of an overlooked aspect of number theory.
You do not seem to understand how mathematics works.
How insightful of parent post for stating the obvious! IANAM, nor have I ever claimed to be one.
OTOH, many of the contributors to this thread do not seem to understand how to work the mathematics of this problem. Nor do they seem to recognize that Sainul Abideen could not have gained the recognition he now enjoys if it were so easy to show that his claims were 5 orders of magnitude greater than what they think is the upper limit of the possible. It seems much more likely that slashdot's best mathematicians were screwing it up somehow.
It seemed to me that combinatorial processes had to be involved since the articles about Sainul's work happened to mention (several times) that the mechanism was based on combinations of colors and shapes. While IANAM, I am someone who knows how to comprehend what he is reading.
Since I haven't worked with combinatorial math since my student days, I put forward an estimate that I was sure was low, but which was quite a bit higher than those being bandied about, and I sort of hoped that someone would pick up on this and come through with a description of the correct technique.
Someone did, more or less, but it was not the parent post. Parent post contains some mathematical jokes that whooshed by over my head. Or, well... there is another possibility concerning where parent post's author's head is at, but I won't go there, no sir. That would be insulting.
You calculations simply doesn't make any sense. Now let's compute an upper bound...
No let's not, since before that can be done there has to be an understanding of the underlying mathematics.
Of course my calculations don't make any sense if you haven't yet studied the applicable math. Tell you what, let's look at a couple of easier problems that are very similar. How many unique games of checkers can be played out on a 8 x 8 checkerboard? How many unique games of chess? These ARE similar problems, involving combinations and patterns.
I've talked elsewhere about this on this thread: use "parent" to get back to my first post then re-read the subsequent ones.
How can a square of 9 pixels hold more than 9 * number of colors values?
It would appear that someone has yet to discover the miracle growth of combinations...
This has nothing to do with the size of the matrix. It has to do with the number of unique selections of two colors that can be made from a set of 256 colors.
Go straight to Google and type in "256 choose 2". Then type in "combination math tutorial" and start exploring a new world.
Enjoy!
Thanks for the link to weierstrass' Implicity blog. His diagram of the 102 unique two-color patterns is very useful in this context.
So 102 patterns using 2 colors multiplied by the number of combinations of two colors that can be drawn without replacement from a palette of 256 colors... it has been a while since I've worked combinations but I know how to use Google as a brain prosthesis:
Google this, guys: "256 choose 2" yields 32,640 unique two-color schemes.
So 32,640 * 102 patterns = 3,329,280 possible combinations within each 3x3 matrix. 85,000 such matrices on the 8.5x10 inch sheet yields 2.9*10^11 unique patterns possible on one page. That's a pretty long bit stream. Converting from bits to something understandable yields 33 GB per page.
This unsophisticated technique shows a sheet of paper can hold more than 1500 times the information that the one-bit-per-dot crowd was thinking was the max (22 MB iirc). It is still an order of magnitude lower than the reported achievement of Sainul Abideen-- but I am working as a Resource Support Assistant in a Community College and I don't profess to know much about combinatorial math or pattern recognition. I think it enlightening that Google says that "256 choose 3" gives 2,763,520 unique three-color schemes...
I do, however, know a thing or two about Google and how to use simple resources like it to make the world a little more understandable.