Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Urbit: an operating function

urbit.org

Post reply on HN
Print view View original

121–130 of 177 posts

Re: Urbit: an operating function

#121

Post by throwawayyay123 »

yarvin9 wrote: post #119
throwawayyay123 wrote: post #116

I think it's important to note that Urbit - a very interesting piece of work that deserves more attention - would've been presented at Strange Loop if not the PC crowd made Strange Loop disinvite Curtis Yarvin (the creator of Urbit) for reasons unrelated to Urbit. Using throwaway for obviously reasons.

I obviously find this a very interesting subject. However, I suspect there are better times and places to discuss it...

I don't mean to take attention away from the real conversation here. However, there are a lot of people who are unaware of the back story, and this seems like a good place to make people aware of it.

Re: Urbit: an operating function

#122

Post by agumonkey »

yarvin9 wrote: post #95
mundo wrote: post #92

Earlier quoted context omitted.

Is the list in the whitepaper correct? It shows in part: nap [ pan ] lep ( pel } If that's not a typo, using mirrored words for ( and } is downright malicious.

Doh! That serves me right for tweaking these right before shipping the whitepaper. Yes, "pel" is ).

but isn't pel a rip ?

Re: Urbit: an operating function

#123

Post by yarvin9 »

lisper wrote: post #115
fiatjaf wrote: post #113

Earlier quoted context omitted.

Hey, don't delete your comment like that!

Why not? You were right, I didn't read the paper, I just skimmed it, so my original comment was completely bogus. Deleting it seemed to be the best way to clean up the mess.

And... this is why I keep coming back to HN.

We should get lunch again sometime, Ron...

Re: Urbit: an operating function

#124

Post by invertedohm »

yarvin9 wrote: post #103
david-given wrote: post #98

Earlier quoted context omitted.

> It does mean that your brain thinks the audio. I'm sorry, that's simply not true. At least for me. It may be true for you. From the things you've said, I suspect you're a word-at-a-time reader, treating words as a sequential symbol stream, processing them as if they were speech. This is just one of the several different styles of reading. Others exist. I am, as I mentioned above, a phrase-at-a-time reader. I take i…

I read more like a page at a time. Some might accuse me of being a page-at-a-time writer. Compare the way you perceive "~~" to the way you perceive "++". I hear these as "sig sig" and "lus lus". (Or rather, as "slus," because that's a further Hoon abbreviation, but never mind.) You don't hear the former at all; but you hear the latter as "plus plus," don't you? This is because "tilde tilde" is so heavy your brain doe…

> You don't hear the former at all; but you hear the latter as "plus plus," don't you?

I really do not. A short silence, both of them. I suppose it depends on how you normally code. I have never had a need for speaking out code and am not well versed in it.

Re: Urbit: an operating function

#125

Post by lisper »

yarvin9 wrote: post #123
lisper wrote: post #115

Earlier quoted context omitted.

Why not? You were right, I didn't read the paper, I just skimmed it, so my original comment was completely bogus. Deleting it seemed to be the best way to clean up the mess.

And... this is why I keep coming back to HN. We should get lunch again sometime, Ron...

Sure, as soon as I've had a chance to read the white paper. :-)

(Send me an email if you're serious.)

Re: Urbit: an operating function

#126

Post by JulianMorrison »

csense wrote: post #31

The architecture is exciting. The problem is that the aesthetics of the tooling seems deliberately designed to alienate new users. For example the whitepaper gives new names to every punctuation symbol! And poking around the repo, there are gems like this: "A...variable name is a random pronounceable three-letter string, sometimes with some vague relationship to its meaning, but usually not..." "Nock, for mysterious…

Don't blame Nock, Bash does that too. Return code 0 = true.

Re: Urbit: an operating function

#127

Post by tptacek »

yarvin9 wrote: post #118
tptacek wrote: post #107

Earlier quoted context omitted.

I mean, this is pithy and all, but it obviously doesn't rebut my point. You could apply the same logic to any disparate collection of engineering problems.

No, you're right, it's a highbrow dismissal. Let me be a little more detailed. Once you've built a system that's both a protocol and an OS, you realize that putting an abstraction barrier between them is like building a cow by building two halves of the cow, then sewing them together. It's incredibly hard and the resulting cow doesn't work very well. Take a problem like identity. Is this an OS problem, or a protocol…

The notion that the OS and the network are essentially the same thing, a series of procedure calls somehow knitted together either with a stack in memory or as a series of serialized frames on a network, strikes me as a very 1980s way of framing computer science. I feel like the last 20 years of network software development have been a repudiation of that idea.

So yeah, I find it frustrating that Urbit insists that an overlay network is "half the cow", and that to have a "whole cow", we must also adopt an excruciatingly idiosyncratic bespoke programming language.

Mostly, for me, this is coming from a place of deep respect for the concept and potential of overlay networks, and of frustration with designs that seem to sabotage that potential in order to make philosophical/political statements.

(That's my only attachment to Urbit as a thing worth discussing; by way of bona fides:

https://hn.algolia.com/?query=author:tptacek%20overlay%20net...)

Re: Urbit: an operating function

#128

Post by JulianMorrison »

"Do users want their own general-purpose personal cloud computer?"

No, they do not.

General purpose computers, as such, are useless and boring. They are only useful when converted into one or more appliances. Modern OSes are very good at this. Cloud general purpose compute exists. Its niche to the community of appliance builders, because nobody has any use for it directly.

Re: Urbit: an operating function

#129

Post by fiatjaf »

JulianMorrison wrote: post #128

"Do users want their own general-purpose personal cloud computer?" No, they do not. General purpose computers, as such, are useless and boring. They are only useful when converted into one or more appliances. Modern OSes are very good at this. Cloud general purpose compute exists. Its niche to the community of appliance builders, because nobody has any use for it directly.

"in the cloud, all users have is a herd of special-purpose appliances".

The argument of the whitepaper is that our current internet and Unix environment does not allow for anything else than a cloud with lots of special-purpose appliances, but that does not mean this is the desirable situation, neither for the appliance builders nor for their users.

Re: Urbit: an operating function

#130

Post by galaxyLogic »

KingMob wrote: post #79
yarvin9 wrote: post #54

Earlier quoted context omitted.

At a certain scale, it's often a lot easier to build one system that solves all the problems . Building a building: hard problem. Building 1/3 of a building: impossible problem. Building a cow: hard problem. Building 1/3 of a cow: impossible problem.

Without addressing the merits of Urbit, these examples aren't valid. A third of a building is just a building in progress (foundations laid, i-beams installed, etc.) A third of a cow is a calf. It might be the case that you will get synergistic effects and/or Urbit's vision isn't possible without solving multiple problems. But it could equally be you've bitten off more than you could chew by choosing to work on multi…

Thanks. Dividera et Impera. This is the whole idea behind modularity. Solve the network-abstraction problem the best way you can. Then solve the programming language problem the best you can. Do we need a "network programming language"? Maybe, but general purpose programming languages are better for other things, and they too should be able to talk to the network "layer". I probably missed a lot on the article, it was so long to read.

Of course if you have new concepts they need a name of their own. But that only applies if all the new concepts really are new. If they are the same old concepts, use the same old commonly understood name. If they are somewhat different, use a derived name, like "Big Potatoes". They are LIKE potatoes, but they are really big. If they are totally different, try to invent a name that EXPLAINS THEM BY COMBINING the names of previously existing concepts. How many people program in "BrainFuck" and get paid for it? There must be a reason that there are not so many professional brainfuck programmers around. Nevertheless I applaud this project for thinking outside the box, if that is possible they have done it.

Post reply on HN