mbrock

87 ms·

841.
by mbrock 9y ago
I think we'll be able to just prove the correctness of our contracts and then use them safely. Of course you can't do that with Microsoft Windows because it's too complicated, but you can do it with small contract bytecodes.
842.
by mbrock 9y ago
You could make a similarly incredulous argument that messenger apps are totally ridiculous because why would you use a 500 dollar handheld computer to write a message through the "internet" when you could just talk IRL? I mean, li
843.
by mbrock 9y ago
That's why I hope that smart contract systems will make this kind of thing much easier in the near future. That's also part of what I'm working on. It might be easy for a small group to manage with a convention and an informa
844.
by mbrock 9y ago
That means there is an entrepreneurial prospect for you to enter the space with a team that has a more serious background, right? Eventually people will generally understand that contracts require serious correct engineering, and you'l
845.
by mbrock 9y ago
I'm not sure premise 1 is true. The most useful smart contracts so far are tokens and multisigs, and those are definitely simple enough to do right (if you don't make a major blunder like the Parity team did).
846.
by mbrock 9y ago
Yoichi Hirai and others are working on completing proof assistant support for EVM, meaning you will be able to prove program properties using logical pre/post-conditions with machine checked proofs. There are already useful tools for s
847.
by mbrock 9y ago
Imagine you and your friends are for some reason dissatisfied with the way you are able to deal with money among yourselves. For example, maybe you like to quickly reimburse each other for dinner, but you won't or can't use Facebo
848.
by mbrock 9y ago
It depends what simplicity you mean. The Solidity compiler is very far from simple, so if you consider that part of the complexity incurred by using Solidity, then a reasonably small bytecode program might be much easier to audit. Consider
849.
by mbrock 9y ago
If you can't prove the correctness of your program, you shouldn't run it. Why are you concerned with proving the correctness of arbitrary programs? Can you give an example of an Ethereum-like program that can't be proven corr
850.
by mbrock 9y ago
Obviously all EVM programs halt, otherwise the Ethereum network would be completely ruined immediately. You already know why they halt: there's a gas limit. Ethereum programs aren't Turing complete. They always halt. Turing comple
851.
by mbrock 9y ago
You can also just write bytecode directly. It's a simple stack machine, and contracts should be as simple as possible.
852.
by mbrock 9y ago
Secure messaging was the same for a long time but now it's improving rapidly. I don't think managing keys will get harder and harder. Already using a Ledger Nano is pretty nice, and I imagine in a couple of years it will be very n
853.
by mbrock 9y ago
https://localbitcoins.com/country/VE http://www.aljazeera.com/indepth/features/2017/04/venezuelan...
854.
by mbrock 9y ago
I let them pile up into a big anxious horror too repulsive to even approach until absolutely necessary, at which point I sloppily and with great frustration do the very least I can to somehow postpone the ever-escalating disaster that will
855.
by mbrock 9y ago
Because for some reason I enjoy using a smart contract blockchain as the platform for my interactions. That might be for example because I also want the ability to use the service with a multi-signature wallet... or because I have been lock
856.
by mbrock 9y ago
That actually scares me enough to disable Swank on my laptop. Actually, to disable all localhost services that could by any stretch of the imagination execute code. Tl;dr: web sites can send requests to localhost TCP sockets despite origin
857.
by mbrock 9y ago
Gas doesn't only measure the computation resource but also memory and storage, so "cycles" too is an inaccurate metaphor.
858.
by mbrock 9y ago
The ideology that everything has to be completely cryptoeconomic and trustless is hugely overblown. It's good that Ethereum eliminates the central trusted third party for the base layer, but then you are free to be flexible with how yo
859.
by mbrock 9y ago
On the contrary, centrally authorized systems can be extremely fast and cheap. Postgres outperforms any blockchain. The issues are about authority and power, and also the barriers to innovation and access which that power imposes. Payment c
860.
by mbrock 9y ago
Why does Bitcoin have to solve those problems? Is it completely failed even if it doesn't provide fast and cheap transactions? What if it turns out that Bitcoin has a unique role to play as the first PoW blockchain and the only success
861.
by mbrock 9y ago
Try explaining any technical feature of the incumbent monetary system to any normie...
862.
by mbrock 9y ago
I think this is a fundamental problem of learning and teaching, especially via text. It's also that most people who teach just don't have Feynman levels of understanding. Mostly we are all kind of groping in the dark, and we all p
863.
by mbrock 9y ago
Depends on the systematicity and avoidability of the accidents, I would say. Negligence is different from manslaughter, but it seems like a spectrum.
864.
by mbrock 9y ago
"I assumed the lack of coverage meant that everything was back to normal." Incidentally, that seems like a pretty fundamental issue with news in general.
865.
by mbrock 9y ago
But that's effectively the same goal. If you engineer it to work for 2 years as cheaply as possible, you're probably engineering it to break after 2 years. Would we get better results from "it must cost max X, how can we make
866.
by mbrock 9y ago
Yeah, I'm quite a big fan of Eric Evans and DDD. Especially the more high level parts about ubiquitous language, bounded contexts, anti-corruption layers, and many of the other patterns. Yet I think that the focus on single-dispatch OO
867.
by mbrock 9y ago
If I say "students can be registered for courses in the school" then you might think the OO model should be school.register(student, course) which does actually seem reasonable to me. Here the school object would basicall
868.
by mbrock 9y ago
I wonder how to even formulate the question to the practitioners -- it seems like an artificial choice imposed by the computer formalism (single-dispatch OO). In other words, I don't see how it is a domain question, and it seems likely
869.
by mbrock 9y ago
How did you decide that the registration is a property of the student and not the course?
870.
by mbrock 9y ago
After using Ratpoison as my only window manager starting in 2004, I've now switched to Stumpwm for the better multiscreen support. (It's the successor to Ratpoison, written in Common Lisp.)

More ›