wow.

I didn't even realize we were having this discussion here.

Well, I don't think it would be appropriate have a discussion about the US, patriotism and the military in the cryptography chat.

(while we do accept a certain degree of off-topicism, I think this is pushing it too far)

I got potato shifted! :P

→ 6 messages moved to ­Trash

I had no idea there was a chat trash

@bdegnan it's where we push messages if there's too many (removed) s

(I also pushed all the old messages around this topic there)

(note that while the topic restrictions will be more loose there, it's not an anything-goes place, mods will check it)

oh, btw @forest as a "room owner" you can also move messages, so you don't have to wait on mods for that kind of thing

@bdegnan also it's a tool for room owners to "mass-delete" messages while retaining accountability

ah, I see

So, microsoft rejected my 34 character password because it didn't have any special characters or something. I cannot believe that's still a thing for anyone who knows anything. :/

@bdegnan well, it could be 34x 'A'

which is obviously not a secure password

I guess so. :/

eg 11x 'Aa1' would satisfy everything but the symbol rule

you'd think there's be a smarter way to determine it though. I'm sure there is, but I don't create enough passwords to care. The most depressing thing was I had to login to get a document that just a notice of a policy change, which they could have emailed. They also did it comic sans. :/

Why would a university have an official document in comic sans. :/

@bdegnan why would anyone ever do anything in comic sans?

(also this wouldn't fly here as my uni has a "corporate identity" which specifies standard fonts and layouts)

@SEJPM ha, true.

I do all of my work in LaTeX and send them rendered PDF reports. They still want a DOC file, so I made a script that pastes JPG images into a DOC file of my PDFs. They still haven't figured that one out. :P

@bdegnan that actually does sound neat

It has to be a .doc, and that's what they got. Administrators are just weird at US universities. The overseas ones that I work with are much more competent.

@bdegnan I saw a TRNG design that used 3 op-amps to implement modular doubling (times 2, compare to 0.5 Vcc, subtract 0.5 Vcc if necessary). There is a clock. Raw voltage output is fed back as input to the same circuit. I think the idea is that there will be some noise component on top of the input/output and even if the voltage of that variation is small, it will eventually contribute enough to effect the output after sufficiently many doublings.

Whether the output is less than .5 Vcc or greater determines whether the digitized output is zero or one.

My gut says that it's not a good idea to rely on an amplifier to behave like the simple/ideal model. Is there some transistor weirdness that would break something like that?

Found the link. "Infinite Noise" doesn't sound at all sketchy.

> A 40uV change on the hold capacitors results in a current of about 55,000 electrons, so the hold capacitors are able to capture about 15 bits of resolution of this noise signal, which gets amplified by K and combined with later noise samples in the hold capacitors every cycle. The capacitors s hold about 2^31 different charge levels in the range of 0.3V to 3V. This is effectively a 31-bit register which we multiply by 1.82 every cycle and add a 15 bit noise signal.

That sounds suspicious as well. :/