Only, they should have started working on it slowly one department at a time a few years ago...
They will probably end up finishing the migration to something else just before that too becomes unsupported and they're back in the same boat.

Spammers will find other ways to host their sites, like compromised boxes...
Also even if you pull boxes and domains after abuse complaints have been made, its usually too late because any victims have already fallen for the spam.

A cheap VPS plus a domain name would also be more than adequate for hosting your own email, or even a low power ARM based machine running at home on the end of your DSL assuming you have a static ip and an ISP that doesn't filter SMTP traffic.

As for spam filtering, a filter that's dedicted to you will usually be more effective too as it can learn about the email *You* typically receive... A lot of spam is sent around in languages like russian and chinese, but if you can't read these languages then chances are you will never receive any legitimate email written in these languages... A major email provider cannot block entire languages because they might have customers who speak those languages, but a mail server dedicated to one person can easily and reasonably do so.

Security is expensive, hardware is cheap.
You can buy from vendors who are used to dealing with clients holding confidential data, and expect them to handle returns or swallow the cost of replacements without returning the dead ones. It all depends on the contract between you and the supplier.

Or you can simply not return faulty drives, just replace them and then destroy the faulty ones.
Many places will stress test the drives for a while before putting live data on them, most drives that will fail during their useful lifespan will do so early on and can be returned at this point because they don't have any worthwhile data on them yet. Drives that fail after a few years are worthless anyway, and will just be replaced with newer higher capacity drives.

The overhead of encryption means inferior performance, higher cpu utilisation, overhead of key management, cost of dedicated crypto hardware etc... This will often outweigh the cost of a couple of extra drives should some fail.

The problem is you can't always be the one to dictate the contract terms, and quite often someone utterly incompetent will have come up with the terms...
There are organisations which are burdened with the requirement to encrypt *ALL* disks, even those on servers because someone writing the contract heard the encryption buzzword, or got a kickback from a company selling a disk encryption product.

When the contract stipulates that something must be done, even if that something is stupid then it's very easy to justify the expense and negative side effects of doing so. If something is not stipulated, it can be quite difficult to justify, even if there are significant benefits it can be hard to explain these to people with a poor understanding. Many don't care and are purely concerned with complying with the given spec, while others will assume that whoever wrote the spec knows more than you do and that the spec is gospel and cannot be wrong.

Some banks already sign their mails, albeit with s/mime instead of pgp... PGP requires a plugin for most mail clients, while s/mime is usually supported by default.
I work in security, and always sign my emails... The majority of our clients simply ignore the signature and have no idea what it is.

While we know that Apple were talking to Sony about 4k content, we don't know who else Apple were talking to about the same thing... It was Sony who leaked rather than Apple.

And there's nothing stopping them continuing selling the same plans at the same price, they just have to be honest and tell their customers that they aren't unlimited.

While these claims are probably ridiculous and with no merit, don't underestimate the speed at which medical research can progress under a regime which doesn't have morals holding the researchers back...
The nazis made significant progress.

You should teach general concepts that can be applied to any tools... If you teach the tools that are currently in use then chances are they will have been superseded by the time people leave education and enter the industry anyway. Even if the same tools are around, they will likely be much newer versions and could have significant changes that made your experience of the previous versions largely useless.

We learned wordperfect for dos in school, but i've never worked anywhere which used any version of wordperfect.

And this is even more ridiculous because most blank media, especially hard drives will never be used to store music...

If your booting from the SSD, chances are the machine will crash...
Would be much better to just stay in readonly mode, and give you the chance to copy data off (and yes im aware this is no substitute for a backup, but think of the use case of a travelling laptop far away from its backup server etc).

The Linux kernel will run on a 486 and upwards, i believe they dropped 386 because it was extremely crufty... It also still runs on m68k as far as i know, all the way back to mc68020 (i even have an m68k box to hand running a fairly modern kernel but its a 68060).

Mainstream distros compile to require modern hardware by default because it makes sense to do so, not making use of such features results in inferior performance when running on newer hardware, thats why many people use distros like gentoo where you can compile everything with support for your actual cpu.

There's nothing stopping you from compiling binaries to support older cpus, and there are distros out there which support them but that's no reason to hold everyone else back for the very few niche users who might want to run linux on a 486...

I'm also not sure why you'd want to run gcc3 on a 15 year old piece of hardware, you could always cross compile on new hardware and doing so is going to be MUCH faster.

Or just make your environment portable enough to run on anything vaguely unix-like... Linux, *BSD, Solaris etc will still compile and run extremely old code without problems.
There are plenty of old programs out there which still compile and run on current systems without problems. I have code which predates linux, and predates any 64bit hardware yet still compiles and runs (very quickly) on a modern amd64 linux host.

The only difference here is that the critical infrastructure is custom designed...
There are MANY cases where a piece of ageing critical infrastructure is actually a black box purchased from a vendor who are long since out of business, which is actually a much worse situation than these guys are in.

The Amiga may be a proprietary system, but it's also one that is well understood and has been well documented over the years. Several people who worked on development of the Amiga hardware are still active online (e.g. Dave Haynie). The custom control system sitting on top of it belongs entirely to these guys, they have the source code and still have access to the original developer.
Similarly, Amiga hardware is common enough that spares are easily available, and there are also a number of places which specialise in repairing Amiga hardware to extend its lifespan.

They could get the original developer to port the system to a new platform (although any platform current today will be just as obsolete as the amiga in 20 years and might not be as prevalent).