smaddox

50 ms·

241.
by smaddox 5y ago
Large scale climate models are more like modeling the possible energy distributions of such pendulums into the future. That can be done analytically for pendulums. You can do it analytically for very simple models of climate, too, but more
242.
by smaddox 5y ago
Seems pretty straightforward and accurate to me. Helion is already re-capturing most of the energy from each pulse. I'm not aware of a single other approach that has demonstrated direct conversion from plasma energy to electric energy.
243.
by smaddox 5y ago
Helion is projecting a small amount of net electrical energy out of their next reactor, which they project to come online in 2024. And that reactor is designed primarily for producing He3, which they will use in their follow-up reactors des
244.
by smaddox 5y ago
So only executives and rich people should be held accountable? Exactly how rich must one be?
245.
by smaddox 5y ago
> It should ultimately be possible to compile Dawn to concurrent nets and therefore to evaluate concurrent sections concurrently Correction: to interaction nets: https://en.wikipedia.org/wiki/Interaction_nets
246.
by smaddox 5y ago
Considering it took them roughly 2 years to develope Starship from scratch, even if that's the case, it would still be an order of magnitude lower cost than anything else, wouldn't it?
247.
by smaddox 5y ago
There are a few more F-150 drivers than there are Shell board members or rich conservative space tourists.
248.
by smaddox 5y ago
Is there something particularly far fetched about in orbit refueling? In-orbit docking and pressure equalization is a long solved problem. Is there something I don't know about?
249.
by smaddox 5y ago
Yes, essentially. Things are slightly different for the concatentive calculus, though. Traditional combinator calculus is a restriction of the lambda calculus. SK calculus does not operate on a stack.
250.
by smaddox 5y ago
These toy languages are not practical for general programming. Their purpose is to test and validate the formalized foundations of Dawn. Once we build up functionality nearing that of Dawn, it should be more feasible to provide practical ex
251.
by smaddox 5y ago
Joy is definitely an influence. In my understanding, it's the first strictly concatentive language, whereas Forth has several non-concatenative terms.
252.
by smaddox 5y ago
I do hope to provide a derived term or syntax for something like a for loop with support for continue and break, since it's hard to match them in terms of productivity and ergonomics. I haven't tried to prototype anything like tha
253.
by smaddox 5y ago
I'm not entirely sure I understand. The formal semantics hopefully speak for themselves. Ultimately, when compiling Dawn to C or machine code, we'll want to use a stack of instruction pointers rather than maintaining an explicit q
254.
by smaddox 5y ago
I'm not sure if such a reduction is possible, in general, without explicitly modeling a map from stack names to stacks (which you could certainly do in the UCC, but it would be quite verbose). For specific programs you can always do a
255.
by smaddox 5y ago
> If you only ever push "effects" onto a stack, you're just building up a trace. That's an interesting concept, but not my plan for capabilities/effects. Instead, I'll be relying on linearity and qualified t
256.
by smaddox 5y ago
My next post will cover this :-)
257.
by smaddox 5y ago
> Why not base this project on Factor, which has a pretty decent existing implementation and compiler? I have some ambitious ideas that aren't really compatible with Factor, or with any other existing language.
258.
by smaddox 5y ago
Joy is essentially a dynamically typed (single-stack) concatentive calculus, as far as I understand. So it's essentially an extension of this untyped concatentive calculus, with different names for some of the intrinsics. Dawn itself w
259.
by smaddox 5y ago
Yes, that is definitely a possibility I hope to explore. Linearity has some interesting consequences for co-routines or async/await style code that I don't fully understand.
260.
by smaddox 5y ago
If you take a close look at that sentence, I was careful to limit the compactness claim to this pair of languages :-) I'm not surprised that there are pairs of languages where one is uniformly more compact or efficient than the other.
261.
by smaddox 5y ago
> [...] TBH I don't see the advantage of his multi-stack proposal compared to local variables as in Factor [...] Local variables effectively turn the language into the lambda calculus. They're equivalent in the sense of any two
262.
by smaddox 5y ago
In the existing prototypes, only expression stacks are nested, not value stacks. The value multistack is flat and is modeled by a map from stack name to value stack. Eventually, when compiling to C (and potentially to machine code), values
263.
by smaddox 5y ago
1. There's really no main stack, there's just the current stack and the enclosing stack (and the stack enclosing that, etc). It seems to work well in practice, but there might be some subtleties that I've missed, which is a m
264.
by smaddox 5y ago
The formal verification aspect doesn't really have anything to do with Forth, but I do expect the multistack concatentive calculus to be much more ergonomic than applicative languages for linearity, which will play an important role in
265.
by smaddox 5y ago
;-) There's a lot of work left to do to demonstrate the value of this, but it's my primary motivation for basing Dawn on the concatentive calculus.
266.
by smaddox 5y ago
Author here. Ask me anything :-)
267.
Foundations of Dawn: The Untyped Concatenative Calculus(dawn-lang.org)
94 points by smaddox 5y ago|45 comments
268.
by smaddox 5y ago
If I remember correctly, the silicon used to manufacture transistor-grade wafers is purified to roughly 10 impurities for every billion silicon atoms. Less pure starting stock might increase the cost of getting there, but probably not by to
269.
by smaddox 5y ago
> The radiation risks from fusion are orders of magnitude smaller than from nuclear fission power. We still need to think about them and make sure that plants are safe, but it should be significantly easier to manage. Strictly speaking,
270.
by smaddox 5y ago
I think that's the debug allocator, but I don't know if that's even implemented yet.

More ›