1) people are storing data now that needs to be protected past the time that "quantum" hits. So safe algorithms are needed even now.
2) large scale systems with lots of parties etc. take forever to effect change. You have to get everyone to agree that something needs to be done, then get them to agree on what to do, then get them to actually do it. Barring a worldwide disaster/alien attack/etc., this just won't happen in a matter of weeks, or months, or even years in some cases. It can take decades to get industries to move off of unsafe algorithms even when you can demonstrate an actual danger. If something's going to be a problem in 8 years in those industries, and you aren't actively trying to solve it now, you are not gonna fix it in time.

https://www.iana.org/dnssec/ce...

Here is a link. The ceremonies are performed to do any cryptographic operations which require a Root Signing Key. When you need to use such a key, you usually have to get a number of people called "key custodians" who each have independent physical access to one part of the cryptographic key, usually stored on a smart card or other secure token device. You will usually have an overall number of custodians and a certain quorum of them will need to be there for a given operation. Like, six of ten, three of seven, etc.

They all have to get their fragment of the key (their assigned device) which is usually stored in a safe which only they have access to. Then they all need to be in the same room, usually a SCIF (think a bank vault with a data center inside it). Whatever process they run will ask for their components individually, and then once the required number of components have been entered, the system will reassemble the master crypto keys and do whatever it needs to do.

The process is designed to make sure that fraud is very difficult and cannot happen without being detected. All the systems and physical access along the way will typically be monitored, controlled with biometrics and other secure mechanisms, and easily auditable. Any activity requires an intentional quorum of people to agree to do it, so you can't just get one guy to go do something bad.

It is kind of like nuclear missile launching, the root of a certificate authority, the root of a financial processing crypto scheme, etc.

In this case, sounds like something broke down and they can't get into a safe or some other secure location to retrieve key components. Usually these systems are designed to fail secure except in the case of life safety (i.e. you can get out if there's a fire, it just creates a huge audit nightmare).

I like Silverlight in that it sucks and makes DRM hard to use, which helps hasten the demise of digital restrictions management by pissing off users and causing the bastards pushing it to lose revenue. Free Software based companies have to resist -- who else will? We've taken over the entire computing world, and now we should use our power for good by refusing to support DRM or anti-features of any kind. If Linux doesn't support RestrictedBoot for example, Dell couldn't sell any servers with it enabled.

If all of us good programmers refuse to participate in the DRM culture, then it will die from a lack of anyone with the skill required to work on it. If everyone on the street refused to accept DRM, market forces would have to change. It worked for music (but seems to be coming back with Spotify and the RIAA's amazing nearly billion dollar judgment against the only competitor...).

This is the last grasp for profit and power by a dying industry. They should just have the decency to go ahead and die.

In the mean time, the pirate bay exists.

I think you meant Digital Restrictions Management. It's a sad day for Mozilla, the w3c, the web as a whole, and open culture. At least there's still the iceweasel fork that doesn't come with this shit.

I have taken apart a black endura pro and ... unless there's supposed to be a second steel plate to add weight, the entire keyboard is just a pair of metal plates with the membrane mechanism sandwiched between them. Don't have it handy so I can't tell if it is steel or aluminum.

What I'm saying is ... fn n m, and f (n, m) being different is good. You have more flexibility since automatic currying is available when sensible.

The author has worked on SML/NJ internals, has a strong background in type theory, and has written similarish software in the past. I'd say he was able to solve those problems. Those problems are not even particularly complicated to solve once you have an expressive type system.

It would have killed them, because (n) is a tuple of one element.

The current syntax has a few benefits. A big one is that you can write things like fun mul n m = n * m; fun double = mul 2; double 4 (* = 8 *);. Automagic currying is good.

Back at HOPE9, there was a really awesome presentation on semiconductor manufacturing. It's worth the entire 90 minutes and IMHO was the best part of the conference. I've ended up showing the video of it to a few folks now, and it never becomes less awesome.