[jlehman]

Disclaimer: I work on Urbit full-time.

There are some good write-ups on the technical decisions here:

https://urbit.org/blog/precepts https://urbit.org/blog/precepts-discussion

Funnily enough, most of the "mystique" isn't really intentional. It's that the inevitable "muh Curtis" arguments that always show up have discouraged many of the team from bothering to engage in explaining the project (here moreso than anywhere else) and instead focus on building the thing.

Summarization: Urbit is a reaction to Unix-driven software complexity that dominates modern software development. The thesis is that cascading complexity cannot be solved without a complete rewrite of the computing stack, and that goes all the way back to the operating system. Here's an even better summarization: https://twitter.com/pcmonk/status/1201298411011629063

Alternative summarization: the internet, being built on a hodgepodge stack of tooling that wasn't made for people to communicate the way that they actually do now, has major incidental flaws. Urbit is a completely rebuilt computing stack that better maps to what we want to do with networked computers. That involves things like, but not limited to baking a non-fungible, valuable identity into the networking layer.

This is another good summarization, although the sections after "Urbit ID" are out of date: https://urbit.org/understanding-urbit

[weare138]

>Summarization: Urbit is a reaction to Unix-driven software complexity that dominates modern software development.

I keep reading explanations about Urbit along these lines but what's perplexing about this explanation is Urbit does nothing to achieve this goal. Regardless of the fact that Urbit is referred to as an operation system it still isn't one. It relies on a conventional OS, often UNIX family OS's like Linux to run. Urbit just layers a new even more complicated software and networking stack on top of the 'Unix-driven software complexity' that still underpins Urbit. Urbit does nothing to address these problems, it just makes them worse. Urbit's underlying software stack (Nock, Hoon, etc.) is just bizarre and non-nonsensical. For example, this gem from the docs:

"A loobean is a Nock boolean - Nock, for mysterious reasons, uses 0 as true (always say "yes") and 1 as false (always say "no")."

Wat? Why?!? And it doesn't end there. Urbit is littered with these WTF design choices and terminology. No one in their right mind is going to build professional software solutions on a platform like this. You might as well just be developing software in something like Brainfuck.

[tptacek]

When you write things like "muh Curtis", you engage in those arguments yourself. If you want to keep the conversation focused on the technology you're working on, do that; I agree, it's the more relevant thing for HN to talk about. People who have problems with the founder of this project are not crazy randos. Keeping HN conversations on-topic takes work, from everyone.

Beyond that, your comment doesn't say much. Computing and networking as it exists today, you say, is bad. Urbit is a ground-up rethink of all of it. OK, but there are 10th graders with the same idea (just as there were when I was a 10th grader). What makes Urbit worth taking more seriously than those ideas?

[emptysongglass]

> People who have problems with the founder of this project are not crazy randos.

The founder divested himself of the project yet the current top comment doesn't want to touch Urbit as if it is cursed by his ghost.

It's these same sorts of comments that drag down all discourse because they prohibit any other state than negative. It's these same comments that pollute any healthy attempt at discourse around crypto here with the Evil Eye that just pushes any one with a different idea to the HN hivemind out.

> What makes Urbit worth taking more seriously than those ideas?

This is explained, in technical and philosophical depth in the links provided you.

[jlehman]

Ideas are cheap, and Urbit isn't just an idea any longer. It's a working system that's been built by dozens of people over the years, and it's only picking up steam.

In Ron Garrett's words:

> The mere fact that Urbit is still a thing, that it has not yet collapsed under the weight of its own intentionally induced baggage, is worrisome to me. Something is keeping that project alive, and it's not technical merit. I don't see a lot of viable options other than some kind of fanaticism.

Looks like he and his ilk are just wrong and having a hard time believing that. It's alive because an increasing number of people want it to be, and no amount of theorizing can deny the reality of actual growth — which, in case everyone forgot, is what the OP is showing.

[ineptech]

I agree that urbit has far outlasted all of the nay-sayers in a technical sense. People have been saying "this is some elaborate joke" for years, but the project has made steady progress and really is pretty usable now. It does what it says on the side of the tin, that's quite an achievement.

But, whatever its technical merits, do you think it can ever get to a point where it reached enough users to be useful (in the "chicken and egg problem" sense that platforms need a critical mass of users to be useful platforms), given the millstone around its neck? Do threads like this (where detailed, technical explanations get downvoted to oblivion, and "I heard Yarvin is racist" is the top comment) make you want to throw in the towel?

[jlehman]

It has enough users to be useful now, which number in the thousands. It's also not a service that anyone maintains on behalf of others, so there's no one that needs to have a critical mass for it to be useful. Nearly everyone running Urbit is self-hosting locally or on some VPS somewhere, and they can do that indefinitely even if no more code was ever pushed.

If this kind of stuff made anyone want to throw in the towel Urbit wouldn't be a thing anymore. There's only a millstone around its neck on HN, but HN is gradually becoming more and more irrelevant. To our base and the people finding us, "I don't think about you at all" reflects their sentiment for HN conversations and the tenor of conversation that happens here.

I'm just here because someone told me (from within Urbit) that this post went up, it's Friday and I don't have as many meetings, and I couldn't resist myself. You seem fine though, so boot a comet and hit me up on the network if you want a planet or something to keep the conversation going.

[ineptech]

Oh, I've had a planet since 2016 or so I think, I boot it up once a year or so and check out the updates. I agree that the platform is big enough to be self-supporting in a technical sense. By "chicken and egg problem" I meant what is discussed here: https://www.joelonsoftware.com/2000/05/24/strategy-letter-ii... I can use my planet to talk to the other weirdos who use urbit, but not a ton else yet - it won't really succeed as a platform until it has an ecosystem of developers making apps and users using them.

Is there some progress there btw? Last I checked there was no way to discover apps other than hard work, and sharing them (installing an app someone else had written) was pretty arduous.

edit: and I know HN isn't the whole world, but I suspect it is representative in this case. I just can't imagine any person learning anything about urbit from any source that doesn't get the 2-3 worst things Yarvin has ever written appended to it within the first hour. It's hard to imagine how it ever gets really popular under those circumstances.

[hocwyn]

This is what my company is building on Urbit: https://uqbarnetwork.medium.com/introducing-uqbar-network-6b...

Having a large audience of pre-existing users is nice (and ditto for dev ecosystem for tooling and libraries) but ultimately what makes Urbit a good idea is, as Josh says, that we want computers to primarily be doing one set of things and they were originally designed to do a different set of things (with the things they currently do added as a complicated afterthought). A blockchain protocol is in the end a p2p application. It benefits hugely from running on a natively p2p networked operating system with a built-in cryptographic identity infrastructure. If Urbit's technology works, people are going to build on top of it because they need it, and their customers will come to use what they've built because it works better. If the tech didn't work, it wouldn't matter how many people tried it.

You're correct that there is no "Urbit Store" analogous to the Apple Store or the Android Store (this is a feature not a bug for many users, although I sympathize with both perspectives). I'm not sure what you mean about installing apps being arduous - currently you start typing the developer's name into the homepage search bar, it autocompletes, you are given a list of apps they've published, you click the one you're interested in, installation is a few seconds to a few minutes.

[jlehman]

We shipped a proper mechanism for software distribution last October. If you haven't booted up since then you might want to. Discovery still requires word of mouth (but not for long...), but sharing them is now as simple as sharing a link in chat or punching something like the following into the Dojo.

|install ~dister-norsyr-torryn %canvas

edit: yes, there's an ecosystem of developers shipping software now. Here's a screenshot of my homescreen, which I cleaned up a couple of days ago to remove some apps I don't use often: https://imgur.com/7keFr7Y

Of the apps you can see, 13 were created by distinct developers. Some people are more prolific.