3 ms·

So how do you write fire-and-forget systems? Typewriters can't become hopelessly insecure by leaving the typewriter sit on a shelf for a year. On the other hand

So how do you write fire-and-forget systems? Typewriters can't become hopelessly insecure by leaving the typewriter sit on a shelf for a year. On the other hand, you can't leave a Linux box unattended for more than a couple months before it's hopelessly insecure. Is the solution unikernels, or what? I think we'll have to find ways to make software that stands the passage of time a bit more. I feel like it's really hard to build software that survives on its own for even a tiny bit of time. Are we in an era that will leave no usable artifact behind?

isostatic 7y ago
Typewriters can become insecure if you leave them alone -- someone could modify it to send all keypresses via wireless transmission, and probably even power from those presses. A typewriter is no more or less secure than a keyboard. A computer is far more than a typewriter, you need to compare the security of the documents and the entire system. A non-networked computer in a locked vault with encrypted files and a password is more secure from incorrect access than a pile of documents in a locked vault, although once incorrect access is gained, the scale of the damage can be far higher with a computer (it's easy to read one or two documents, but to try to smuggle 5000 documents out physically is hard. To sumggle them out from a computer is far easier. As the common poster in the 90s went - To Err is Human, to Really Foul Things Up requires a Computer.