catern

>You know, if you'd actually attempted this and had some valid point, I could take you seriously

I do in fact run everything on the same machine as my wifi router: http://catern.com/computers.html

But most of them do have USB ports, and: https://numato.com/product/8-channel-usb-gpio-module-with-an...

Why do you have such an underpowered wifi router running proprietary software you can't modify? Because it is a general purpose computer that has been reduced to the status of an appliance.

>What you lament is that so many of our "PC" systems seem to have wasted processing potential

That's not my lament. I don't care at all about not wasting processing power.

I care about the fact that it's so difficult for people to use their existing computers in ways that were not intended by their manufacturers, that even technically adept people will just buy a new computer rather than deal with it.

It should be easy to install some new software on any of your computers (router or desktop or TV or phone or whatever) and have it do some useful new thing. But it's not.

I'm glad you mentioned PiHole because it's exactly the kind of thing I'm talking about.

Dedicating a Raspberry Pi to running a DNS server is silly.

You could run it on your wifi router instead, or you could run it on the individual nodes. Either of those should be extremely easy. But practical general purpose computing has decayed (or was never really there?) so instead there's this silly PiHole thing where you just buy yet another computer and convert it into yet another special-purpose appliance because general purpose computing is so hard.

The marginal extra energy consumption of running more stuff on your main PC is likely less than 2 amps.

They're really just a sign of how weak general purpose computing has become. Almost all uses of Raspberry Pis for hobbyist projects could and should be done just as well with the computers that you already own and already have running.

Increasing prices will increase equity. By charging rich people more, you can afford to give greater subsidies for marginalized groups and poor people, and lower the price from whatever it is currently for them. You can even use that revenue to pay them to go camping, if you really want to!

138

I feel like my programming knowledge helped, which maybe means it's a bad IQ test - being a programmer shouldn't mean your IQ tests higher. But then again, maybe the causality is the other way around?

The customize interface built into Emacs already provides this. Though it supports customizing far more things, so it might be overwhelming. Maybe a new user customization group would be good.

Can you say what specifically? Many language-specific package managers match your description and are completely handled by Nix, but maybe you're talking about something else?

Sure, that doesn't contradict what I said. They have been around for years, and they are very expensive and almost all systems do not use them.

It's up to userspace to do more than that. There are other issues which can cause processes to be spontaneously killed (OOMkiller for example) so it's something you should be tolerant of.

>I've personally had this disproven to me on multiple occasions.

That sounds like interesting stories! Can you elaborate?

>Yes but your computer will not gracefully handle CPUs randomly failing or RAM randomly failing

That's incorrect.

There are plenty of machines/OSs which are (or can be) resilient to a CPU failing; Linux, for example. From the OS point of view, you just kill the process that was running on the CPU at the time and move on.

Resilience to spontaneous RAM failures is rarer but possible.

Most distributed consensus algorithms, or distributed systems in general, are not resilient to nodes producing arbitrary wrong results. That's the realm of systems like Bitcoin, which achieve such resilience by paying big performance costs.

So it shouldn't be surprising that computers have the same lack of resilience.

And indeed it does not: Modern operating systems like Linux can perfectly well deal with all kinds of devices crashing or disappearing at runtime. Just like in larger distributed systems.

Wow! Reminds me of https://www.rifters.com/crawl/?p=6116

A hydrocephalic demo!

I've long looked for something like this for my day job, which involves implementing many different custom binary network protocols.

One new project in this space is poke:

http://www.jemarch.net/poke

It can describe arbitrary formats for both reading and writing, and it's usable programmatically, so I think it can solve the problem!

>It always troubled me that there was so much vitriol for Maisel from the exact same community of people who rightly insist that free software licences are enforceable and that non-compliance should be shamed.

That community is generally opposed to copyright and intellectual property altogether. Copyleft is just a hack to create a space where it doesn't exist. If I could "infringe" on others without being punished, there would be no need for copyleft enforcement. That's the idea, anyway.

>* The nixstore treats a filesystem as an object store.

That's a weird thing to list as a shortcoming, since that's the main innovation of Nix. ("package management as memory management" as the paper and thesis said)

Yes - same as with containers.

>Showing that it's as easy as `run_in_executor()` is nice, but not if it does the exact same thing as `run_in_executor()`.

That's fair and a good point! I suppose reading this, you have no evidence that anything but local_thread exists. I'll change that, somehow, to show that indeed there exist values other than local_thread, and that passing those will run things on remote hosts.

This article originally grew out of a tutorial on writing tests using this style, for which running over multiple hosts is not necessary and maybe undesirable... but that's not good for what the article is today.

>Can it spawn the remote processes too? With SSH, or its own system?

Yes, with rsyscall, is that not clear? It's pretty explicit in http://catern.com/integration.html#thread

>but all it does is explicitly run a Python function in a local thread...

That's not all that's happening. I think you might be pattern-matching what's happening in the article to something you've already seen, when in fact it's something novel.

>Running processes remotely? Supervising remote processes?

These things. This is mentioned in http://catern.com/integration.html#thread

Can you say what left you confused?

I'm surprised, I thought https://github.com/catern/rsyscall is pretty self-contained and self-explanatory (if unclear on what exactly rsyscall is useful for), and I thought http://catern.com/integration.html is quite detailed (perhaps excessively so).

Distributed software, which accesses and depends on remote resources, is harder to run than local-only software. I'm working on making it easier to run distributed software for both programmers and end-users.

See the main article describing my current vision: http://catern.com/integration.html

My most significant project in this area: https://github.com/catern/rsyscall

My website in general: http://catern.com/ (happy to collaborate on anything on there)

And my current plans: http://catern.com/plan.html

If you're interested, feel free to contact me using the information on my website.

I expect you already know about this, but Secure EcmaScript is a similar project: https://medium.com/agoric/ses-securing-javascript-in-the-rea...

Similarly the older Caja project: https://en.wikipedia.org/wiki/Caja_project

Yes, yes, it's hard to self-host federated services, because system administration is hard.

So we have to make it radically easier, then.

A washing machine already requires a fair amount of time spent on installation. I don't think "it doesn't work out of the box" is a good argument when talking about 200 pound appliances that have to be hooked up to your plumbing.