While I agree with the general points you make in your Crypto Blues, I also agree with what I am reading from others regarding paving the way for GAK. This is probably because you didn't actually deal with those criticisms individually in your rant. To say that you will draw a line somewhere down the road is ineffectual. It's not unlike the gun owners who won't join an organization, won't communicate with their elected representatives, and rely on "Well, if they come for *my* gun, they'll have to pry it from my cold, dead fingers!" It's too late by then. If and when it comes time for one or more of you to decide to quit rather than do what is wrong, it will be way too late. I'm reminded of the far-left-liberal comic and actor Dennis whatshisname who, in a standup act, made reference to the "cold, dead fingers" view of some gun owners and said, "That'll work!" Whoever it was who said that the capitalists would sell the communists the rope with which to hang them was right. It is even worse in today's brain-dead business world where policy decisions are made by bean counters who can't see more than 90 days past their bottom line and have no visibility whatsoever of the rich component mix of that bottom line. I have no doubt that if and when the time comes that the government puts out RFPs for ID chip implant systems or tattooing barcodes on our foreheads there will be no scarcity of companies lining up to get the contracts. Free enterprise doesn't automatically confer intelligence, forward view, or an understanding of freedom and the long-term consequences of undermining it. It is incumbent on everyone who favors individual liberty and privacy to do everything possible to prevent the infrastructure for GAK and other police state measures from being put into place. Adam Back is entirely correct in believing only in technical solutions, not words. It must remain technically difficult, even impractical, for politicians and bureaurats to simply prescribe how mechanisms put in place by well-meaning people will be used to invade privacy and monitor communications. The best defense against censorship is a technological structure that defeats censorship before it starts. The best defense against the destruction of privacy is a technological structure that makes invasion of privacy difficult or impossible. It is far better to have a system in which a destructive proposal fits badly and threatens to cause innumerable uncontrollable consequences than one which lays the groundwork for the easy implementation of destructive proposals. The risk-of-loss set of arguments is also in some ways a large red herring. There is nothing particularly unique about crypto that makes the potential loss of information qualitatively different than countless garden variety risks with which people deal every day. May organizations have stopped backing up desktop PCs. Most everyone tried to do those backups at one time, but as disk sizes and data volumes increased it became impractical in many organizations. This represents a far greater risk than the loss of messages or files the same employees may keep in encrypted form, but there are no interest groups or government bodies trying to propose Desktop PC Escrow systems. Many employees wittingly or unwittingly use the strongest crypto in existence, from which there is *no* recovery under any circumstances should the employee drop dead: wetware storage. Rate of technological change and increasingly full "plates" have combined to reduce documentation to historic lows. If a document becomes obsolete, even false, virtually the moment it is completed, it is less likely to be written. If it's never written because its durability is in question or there is too much else to do, it can never be recovered except from the functioning mind of the custodian of the information, but no one is proposing "brain escrow." The costs of dealing with the loss of an employee with a headful of unique knowledge are becoming an expected and largely unnoticed thing, like the expectation of a third world business that a simple action like fixing a piece of hardware may involve a large multiple of the time and money it would cost in any U.S. city. And still the bean counters haven't a clue. It is even quite common these days that when one of those custodians leaves an organization there is insufficient staffing to make even a feeble attempt to pick up what the departing person has left behind. Rate of obsolescence aggravates this by making it seem easier to just implement the next version or product than pick up the pieces of managing the old one. I've seen people depart, leaving a cube full of diskettes, manuals and, presumably, notes, only to see the organization move someone else with different job responsibilities into the cube and never, ever, not even once retrieve or look at any of the materials of the departed employee -- not their PC, their files, their desk, their file cabinets, their manuals -- nothing. Individuals can also wipe their mail files, wipe their disks, suffer disk failures, etc. They can take the only copy of something home and die in a fiery car crash. They can drop their laptops from heights more than the prescribed 2 inches. It is not possible to list the ways in which information can be lost and/or compromised, ways with which key recovery cannot possibly attempt to deal. Since many of those things happen all the time in the real world on a daily basis it goes unnoticed that the vast majority of them are dealt with by the people involved without comment and largely without noticeable adverse consequences, just as when smoking was common in offices it wasn't particularly noticed that most offices didn't burn down despite the widespread use of open flame and smoldering materials in tens, perhaps hundreds of millions of instances daily. Loss of encrypted information is almost a non-issue in reality, somewhat like the non-issue of the risk of flying by scheduled airline. Crypto, like airliners, increases the probability that a loss will be more noticeable by virtue of quantity, but has no inherent effect on the likelihood that a greater percentage of total information will be lost. An employee generally exercises the same type of diligence in this area as in the myriad other areas where some thought and action must be directed at preventing catastrophic loss. The motivation is the same, too: retention of employment and avoidance of lawsuit. An operations manager already has to figure out how to make sure that last year's backups will not only be physically safeguarded but also capable of being read and restored should they be needed, for example. There is nothing more magic about "crypto" than there is about "backup" in this sense. Crypto, like the Internet, is a mere conceptual hook on which the ignorant (strongly represented in the so-called "news" media) can hang a set of alarms and not necessarily needed solutions. It's an umbrella under which those who have nothing better to do can collect a set of issues for conceptual amplification and synergistic alarm enhancement, obscuring the fact that many functional counterparts to those issues are in the fabric of everyday business work. The mere use of a computer involves a larger quantity and more serious degree of "gotchas" than arise in the use of crypto on those computers, and has for many decades. The manpower lost in the 10+ years it took us all to move past the 640K limit and the 1960's technology of DOS on the PC platform probably exceeded what would be required to do brute force solutions of the worldwide Y2K problem. Imagine that someone had latched onto the business danger of loss of source code in 1960 and had sensitized everyone to the issue to the point where this kind of discussion and debate were being conducted over the need for national and worldwide standards and mechansisms for commercial and government protection against the loss of source code. Imagine the Congress of the U.S. effectively taking the position that businesses are too stupid to safeguard the source code that runs their enterprises and that the government must require the filing of copies of all source code in case anyone loses some. Or that all compilers and source editors have source recovery features. Would that have sounded reasonable at any time in the last 30-40 years? Would it sound reasonable today? Yes, some people and organizations have lost source code. No, most people and organizations have not lost their critical source code. Those that do tend to weed themselves out, just as those who tend to make bad decisions and take self-destructive actions in any area of life, business or personal, tend to reduce their own effectiveness and participation in the game. GAK is different, of course, in that it brings the agenda of the politics of power into the fray, and thus crypto has become the subject of an array of efforts as disingenuous as the safety arguments of unions when fighting to preserve a completely obsolete and useless job function. All of a sudden governments who never gave a rat's ass about the risks to commercial entities of losses of any other kind of information in any other context are falling all over themselves to promote GAK because they are suddenly overcome by altruism. Right. Or because they have had a revelation and have remembered that commerce is good for the country and government is supposed to be a means of providing stable legal and public safety structure and not an end in and of itself. Sure. Government never misses a chance to take opportunistic advantage, and the chance to treat crypto as something qualitatively different than thousands of other business and personal privacy questions and mechanisms is too good to pass up. The opportunity for government to gain access to information that just happens to be digital is of far greater significance than any offered justification based on the danger to businesses that they will be too stupid to assure their own access to their own information or the danger to nations that a few people or groups of people may be able to communicate in ways that preclude surveillance. The only thing that GAK is about is government power over individuals and groups -- the gradual conversion of citizens into subjects, just without all the trappings of old-fashioned royalty and notions of the divine right of kings. Governments see this as an opportunity to gain access to private communications in ways that would never have been acceptable in the days of paper letters and envelopes. Because politicobureaucrats, like ants, cooperate instinctively to concentrate power in the halls of authority we see mulitple facets to the attack on privacy, but they are only that -- facets of the same ugly stone. If the best minds fall into step to build and facilitate the mechanisms to destroy individual liberty because "someone will do it," we are lost. If "someone" does it, but the best minds are working to make a police state technologically infeasible, we may not see the light of freedom extinguished after all. In the past this battle was heavily influenced by material resources. Increasingly, the playing field is being leveled by technology. The likelihood that a small number of individuals can significantly influence the balance has passed the threshold of credible probability, witness Phil's PGP and numerous other developments. This likelihood is on an upward trend, but the forces of darkness are also making better and better use of technology. Every person who can make a difference and isn't committed to individual freedom is an effective participant in its destruction. If we have to explain to our children and grandchildren why they have microchip transponders in their asses and all their communications are archived by the government it will be a pretty weak justification to say that we brought home the paychecks without interruption. Thomas Junker tjunker@phoenix.net