Taking a 10,000' view of modern OS design. (Warning: extended rant.)

[liam_on_linux]

Frankly, coming from a background in 1980s and 1990s OSes, I think modern ones are appalling shite. They're huge, baggy, flabby sacks of crap that drag themselves around leaving a trail of slime and viscera - but like some blasphemous shoggoth, they have organs to spare, and the computers they run on are so powerful and have so much storage that

Comments

[reddragdiva]

Be careful what you wish for. Go look up Urbit, which is designed to solve precisely this problem. Read at length. I'll be here when you get back.

[liam_on_linux]

I have, I read about it at great length. It is, literally, brilliant stuff. But he doesn't know Lisp or LispMs, so he doesn't know his history and reinvents some wheels. Also, his raging NIH syndrome means that some parts of it are unnecessarily hard, complex, weird or even compromised.

It's still astounding, though. It's further than anyone else has ever got.

But see:

http://www.loper-os.org/?p=1309

http://www.loper-os.org/?p=1352

[bryangb]

You have punted this at The Reg, haven't you?

[liam_on_linux]

This was an intentional context-less intro-less rant. I am still working on bringing together something much larger, more all-embracing, more explanatory, possibly for the Reg or possibly for someone else...

[bryangb]

Top and tail it, and punt it in. Stop automatically over-thinking and over-engineering these things! (-;

[uon]

You say this:
I'm not a programmer.

Followed by:
Your computer should know that a street address is an address, [...] It should understand names and dates and amounts of money

Your first statement gives some indication of why you think the problems you list in the second statement should be easy to solve, when from the right (or wrong) angle they can look a lot like trying to solve the generalised DWIM problem. Quite a few seemingly easy tasks can actually be a right pain to do properly. It's dead easy to say "just calculate this bit, reformat that bit" until, in order to actually program a machine to do so, you try to get a sufficiently concrete understanding of what it actually means in every context, and discover yourself deep in the weeds of weird problems you never even knew existed. There's a reason "yak shaving" is a common jargon term amongst programmers - even just trying to answer the question "what time is it?" can be unbelievably trickyThat's not to say that I'm happy with the state of the art in software or hardware

[andrewducker]

Yeah, that's my feeling with this stuff too.

It feels a bit like Ted Nelson shouting "HTML is bad and wrong and impure, and there is a better way!"

To which my answer is "Yes, and I am in no way convinced that said better way is actually any better by the time you've implemented a global-scale system in it!"

[liam_on_linux]

That's true, and I picked that merely as a to-hand example. It's not a fanciful made-up one. It's based on Jef Raskin's research and the original concept for the Macintosh before Steve Jobs took over and made it into an inexpensive vehicle for the radical new idea of the GUI.

I need to back that up with links and citations, clearly, and the page that I read about it seems to have gone. OK, for future reference, I'll remove that section.

I'd read the real-names article before; the other 2 were new to me - ta.

I was looking for a general example of something non-programming-language or system-architecture related. This is either not it, or needs to be fleshed out much more.

So, thanks for pulling me up on it.

[ext_1399867]

Imagine a 64bit version of Apple Integer Basic with gigs of ram!! ZOOOOM!

[lostcarpark]

Broadly agree too. The whole OS landscape is a gigantic mess. The Linux side of the camp is very slightly less of a mess than the Windows side, but only slightly. We really could do with tearing the whole thing up and starting fresh. And by starting fresh, I don't just mean write a new OS, I mean design new microprocessors, architect new computer hardware, invent new languages, and write new OSs and applications that do it right this time. The trouble is we're too busy updating our Facebook profiles and tweeting what we had for breakfast to get around to any of this.

[andrewducker]

Oh, have you been following the stuff around the Mill CPU?

Fascinating stuff, and genuinely revolutionary.

http://hackaday.com/2013/08/02/the-mill-cpu-architecture/

[liam_on_linux]

I have, thanks, yes. I spent a few days trying to learn enough about it to write a non-technical article about it - and failed. I am not sure I know enough about CPU design to actually do so. I think they're really on to something; whether they'll get the money to do it, I don't know. Sadly I kinda doubt it.