If the electricity is out for a very long time I imagine that's not isolated to the library and people will have much larger problems to worry about.
If a hacker manages to destroy the database, well it's a good thing backups are cheap these days and it doesn't take a trained IT person to implement them. Even Crashplan or Carbonite, or just periodically copying files to an external drive and then disconnecting it would probably do the job. If libraries have the money for hundreds of pounds of paper cards and the cabinets to store them in, plus labor to maintain all of that, they surely have the money for functional-if-basic backup infrastructure.
I'm all for securing critical infrastructure with analog systems/subsystems, but the only argument I could see for doing so with libraries is if said library is in the developing world where electricity is intermittent.
>If the electricity is out for a very long time I imagine that's not isolated to the library and people will have much larger problems to worry about.
Well, yes. But a library rendered unusable is also a major problem, in the long run. And especially if it's not just the local library, but all of them in the country?
I would probably need to read the book in question to be able to do Fermi estimates on what kind of amounts of paper and space the old system would need. But if one already has the "this is a backup system for very rare, very worst-case events, not for active use" viewpoint, one is free to do changes and optimizations. For example, miniaturized analogue backup from which the full-blown card system can be reproduced if the need arises.
What do you mean by "in the long run"? Because a library being rendered unusable for a definite length of time wouldn't be a problem in the long run unless we're talking about a system with path dependence. And even if there is some path dependence, it's not quite obvious whether or not that's a major problem. For example, not being able to lend a book during a few hours could be only thing preventing some terminally-ill genius from solving some elusive mathematical problem before he dies, while everyone else can just try to check their books some other day when the power is back on. But so long as the problem can be solved, the solution might be found out later on by someone else.
On the other hand, if libraries are rendered unusable for an indefinite amount of time, or if they become so unreliable that they aren't even worth the effort, those would always be an issue regardless of whether there's path-dependence.
But it sounds somewhat far-fetched to create system to ensure the reliability of libraries in the event of say, a massive solar flare. Plus, it misses the point: in the event of something able to take out something crucial to the entire US library system for any appreciable length of time, the librarians themselves would most likely have far more pressing concerns in their hands. And that would be our new single point of failure.
I pretty much agree with scottLobster. Unless we're talking about developing countries, where electric power is not so much intermitent as unreliable - whereas it might take a hurricane to cause power outages in other countries, a gush of wind here will do - it makes sense to have a system like this in place.
>Plus, it misses the point: in the event of something able to take out something crucial to the entire US library system for any appreciable length of time, the librarians themselves would most likely have far more pressing concerns in their hands. And that would be our new single point of failure.
I'm not exactly sure if I'm following you, but the general idea of this kind of "resilience" is to introduce enough safety / backup measures at all points in the society so that there is less things that hinge on "single point of failure".
The scenario I had in mind was a national grid of a much smaller country going down. (Apparently bringing up all the power plants back online is not as straightforward operation as flipping some switches 'on'.) For example, in conjunction with other crisis (say, conventional war with cyberattack component) that would turn out to taking much longer than anticipated (because aren't they all "over by Christmas"), so there might be an adversary interested in keeping it down.
And the libraries are not the only institutions handling critical data that used to be stored on paper but now exist on computers. The particular solution of retaining backup paper card system might be silly or otherwise prove infeasible at closer inspection, but what kind of backup systems the current institutions have?
edit. Looking at the rest of the comments, smkellat might have provided a splendidly simple answer upthread. At least, concerning libraries.
If a hacker manages to destroy the database, well it's a good thing backups are cheap these days and it doesn't take a trained IT person to implement them. Even Crashplan or Carbonite, or just periodically copying files to an external drive and then disconnecting it would probably do the job. If libraries have the money for hundreds of pounds of paper cards and the cabinets to store them in, plus labor to maintain all of that, they surely have the money for functional-if-basic backup infrastructure.
I'm all for securing critical infrastructure with analog systems/subsystems, but the only argument I could see for doing so with libraries is if said library is in the developing world where electricity is intermittent.