That's not true. The sources for hpcarm (the port for ARM-based handheld PCs) compile and run just fine. Every version of NetBSD has hpcarm binaries (6.1.5, NetBSD-7_BETA, sources compiled from -current). I think you just don't understand how NetBSD works.

It's simple. C is self hosting and extremely portable, and the amount of extra stuff needed to run C can be none (in the case of kernels, everything can come with the binary), little (shared libraries) or lots (a whole OS). There aren't many other languages which can self-host and can create binaries which can run on bare metal.

Other supposedly portable languages like Java, Perl and PHP require an OS and environment to do anything, which makes them unsuitable for running on small embedded systems, for high performance applications, or for talking intimately to hardware.

The languages also change too much over time. You can't just take old code and run it in a new interpreter. If Java wasn't in the hands of a megacompany, it MIGHT be more portable and less bug-ridden, but right now it's write once, run only in certain places, deal with a zillion security issues. Many companies I support have to keep around a VM or an older machine to run an older JVM because new Java is not compatible with old Java.

C has changed, but not so much that K&R C is unrecognizable to someone learning C now. A program written in the 1970s can be compiled by a CS student today without much more than, perhaps, changing a few #includes. This is what makes it lasting and worth learning.

Like Columbus? Does that mean they're planning genocide and exploitation of a race of people as slave labor?

I don't cling to the old because I'm unafraid of change - I keep using it because nothing better has come along.

Shell-based email is still the quickest and easiest way to keep email in one place and have it be accessible even if I'm on a connection the speed of dialup. No matter how many times I try email clients, nothing works as quickly and as seamlessly. The same goes for ytalk instead of IM programs (luckily, many of the people with whom I want to chat have shell accounts, too).

Until someone comes up with something better, like a protocol which allows for downloading just the text of what I want to see, I'll happily ssh and do email on the server, like I've been doing for twenty years.

Sun's slogan for Java used to be, "Write once, run anywhere." Remember that? Sun didn't make JVMs for many platforms, and didn't even have an official JVM for GNU/Linux for ages. Add to that the fact that each major version of the JVM deprecated features and introduced incompatible ways to do things previously done other ways, and it's no wonder it has become the case that we (meaning IT folks) have to keep around an older (perhaps virtual) machine which has an older and certainly insecure JVM to talk to some hardware device or application which requires older Java. Qlogic switches come to mind.

Since the JVM isn't portable, Java isn't portable. Since software written for one JVM version can't necessarily run on another version, it's not very backwards compatible. Since it has so many security issues, you either have to hope that whoever makes your JVM keeps it up to date or that you're very careful about how it's deployed.

I can't personally think of anything more precarious than trying to deploy real software using Java.

I have long advocated for separating everything - the cable modem / DSL modem should JUST be an interface to the upstream provider, with no NAT and DEFINITELY with no wireless. See the issues with Xfinity and other providers who are now piggybacking their "free" Wifi on customers' connections - I bet it'll be shown in the near future that the already existing NAT table size issues, which already cause many consumer devices to be problematic, are being exacerbated by trying to maintain state entries for the "free" wireless, too.

So you have a cable / DSL modem which is in bridge mode. Then you have some sort of NAT device. If you like running your own OS, a Raspberry Pi or some other tiny StrongARM device is cheap and can run whatever GNU/Linux or BSD you like. Heck, you can even still use your WRT54GL if the CPU in it isn't limiting the speed of your upstream connection.

Then, you have your wireless device. Again, I strongly recommend something that just does bridging - you have the simplest setup because you're not using the wireless device for NAT or any other "features". With all the stories about consumer devices having poor security and intentional back doors, the less exposure, the better. Personally, I pay extra for Apple because the 802.11ac Airport Extreme does wonders with existing 802.11n clients.

The great thing about this is that you can have as many segments as you want without needing a switch which does VLANs. You can plug two USB-ethernets into a Raspberry Pi, for instance, and keep your wireless and wired networks on completely different segments. Or three, and you can have your old device provide a completely separate guest network.

The best thing about this setup is that if one device fails or is shown to be insecure and the manufacturers won't fix it, you can just replace that one device.

It's much better to assume that a server may be or is exposed to malicious traffic than it is to assume not. Even if there's only ever a direct ethernet connection between two machines, assume someone may compromise one of the machines and protect the other. Using a username and password is one thing; if you can filter based on IP address, use software firewall rules to only allow connections on certain interfaces and from certain addresses (or, better yet, localhost), et cetera, you're always better off.

Hope for the best, plan for the worst.

No. BIOS code only gets run at boot time.

People have been predicting the death of Unix and the command line for ages. Most people don't care about long term because they're accustomed to a constant cycle of upgrades to make money for large corporations - it's what they're conditioned to do. If we don't want to run browsers that can get infected, email clients that render whatever they're told to render and systems that have poorly written third party software (I'm talking about you, Flash and Java), then who's the smart one?

I keep wondering if I'm doing old school things just because, but every time I try something new, I find that there aren't enough compelling reasons to modernize and at the same time there are enough good reasons to use what works well.

If there's one thing we can generalize about truly intelligent people is that they are always curious. The geniuses can come up with questions nobody else can.

How about doing a dd of the entire drive from the current system to a virtual disk and trying to make that work? Is the Unisys hardware that special? If not, you might be able to get it working by manipulating the virtual hardware of your VM.

I don't like Firefox because they try to take Windows-isms and force them on Mac users. My user experience is one thing in 99% of the programs on my computer - why should how I select text be different for Firefox? Or why can't I launch Firefox normally by holding command-option and hitting the down arrow like I do for every other program but which sends Firefox into some special "safe" mode?

Firefox shouldn't proselytize specific OS behavior.

Isn't this exactly what happens elsewhere, but in the other direction? After all, many people think that KDE, GNOME and other large programs are written for GNU/Linux and just happen to be ported elsewhere. Try to Google something about setting up Apache or bash and you'll find Linux this, Linux that even though neither are exclusive to GNU/Linux in the least.