Gary Nunn wrote: >Imagine this, all citizens have a federal ID card. To make it secure and >worthwhile, it must be verified by a centralized authority (if not, what's >the point, anyone could forge a fake?) You go to buy something at the >store, or renew your license, or get medical treatment, and your national ID >card won't authenticate because there are server problems / power failure / >hardware or software issues, etc - you get the point. Then you are stuck. >Disabling the authentication process defeats the purpose of the system. > >If there isn't some sort of centralized authentication, then the cards can >EASILY be forged by anyone with a good computer, and again there is no point >to the system.
I respectfully disagree that this technical point kills the idea. An efficient national ID card system can do the authentication without needing to access a central computer. You are right that this would make the system prone to failure. So let's go for parallel computing rather than centralized... And with lots of backups. IDs are always used with different levels of authentication for different purposes. For trivial use of the card, quick visual examination would be enough. For slightly more serious usage, something like a signature or a fingerprint could be quickly checked. For something more serious, the card could be checked, this time by computer, against some local state government database. For a few really serious things, centralized checking would be used. But rarely. Occasionally the centralized checking system will be down so badly that it will pull down the local databases too. Creating major trouble; but no more frequently than the times when the power grid is down. Been there, Detroit over one year ago... I still think that the good arguments against this kind of ID are about politics, privacy, and freedom. What if this ID is used to enforce unjust laws? Technical arguments, I would take as guides to improve the design. Ruben _______________________________________________ http://www.mccmedia.com/mailman/listinfo/brin-l
