jkarni

33 ms·

61.
by jkarni 8y ago
4GB is a bit of an overestimate (it's probably more like 2.5GB) but yeah, Haskell project often consume a lot of memory when building. That said, there are brew and deb packages!
62.
Radicle: A decentralized alternative to GitHub built on IPFS(radicle.xyz)
394 points by jkarni 8y ago|82 comments
63.
To Tweet a Mockingbird(turingjump.com)
3 points by jkarni 10y ago|0 comments
64.
by jkarni 10y ago
To be fair, production (vs. library) code has to usually care much less about being extensible, since whoever needs to use it can also usually change it. In servant dealing with the expression problem (albeit in a pretty different way than
65.
by jkarni 11y ago
If you're talking about the discussion of head and snoc at the beginning of the article, laziness just moved the costs around - head becomes O(n) (from O(1)), but snoc becomes O(1) (from O(n)), because you only pay for the snoc costs w
66.
Efficient Amortised and Real-Time Queues in Haskell(well-typed.com)
72 points by jkarni 11y ago|15 comments
67.
by jkarni 11y ago
SEEKING WORK Location: Berlin Remote: Yes I'm one of the original authors of the servant web framework [0]. Web services that are statically guaranteed to match the specification API; documentation (in markdown or Swagger) that never g
68.
by jkarni 11y ago
A more complete description is available here [0], but the essence of it is that the type of many functions that used to be specifically about lists has become more general, but also arguably harder to understand. [0] https://wik
69.
Content-Type Bliss(haskell-servant.github.io)
3 points by jkarni 11y ago|0 comments
70.
by jkarni 11y ago
Evaluation-order in a sense is more low-level in Haskell.
71.
by jkarni 11y ago
Funnily, there's haskell-cafe thread[0], a github issue, and even a paper , about this (and I think maybe reddit got involved too). Anyhow, the title is kind of too much. At least, given the aforementioned discussions, we're conf
72.
by jkarni 11y ago
We had QQ originally, but removed it because it wasn't extensible (you can define new expressions in the type-level DSL, but you couldn't add them to the QQ). We've figured out how to do that now, and there's an issue so
73.
by jkarni 11y ago
We came up with a way of reifying specifications of web APIs at the type level, and built a web framework around it. The basic idea was, I believe, first developed by Ralf Lämmel and Klaus Ostermann [0], as a Haskell solution to the express
74.
Show HN: Servant – A Haskell web framework with type-level API DSL(haskell-servant.github.io)
55 points by jkarni 11y ago|3 comments
75.
by jkarni 12y ago
Some more background https://lukepalmer.wordpress.com/2009/07/07/emphasizing-spec... .
76.
by jkarni 12y ago
I'd been thinking about just this recently. But to my mind, you don't even have to implement a build system - prolog can be directly used as one (modulo some library functions). You specify a goal, and a bunch of rules that ha
77.
by jkarni 12y ago
Sorry, I should have added some context: >Notice that RPN requires no parenthesis, unlike the polish notation of Lisps where the operator comes first, and RPN requires no precedence rules, unlike the infix notation used in most programmi
78.
by jkarni 12y ago
Huh? Why does Polish notation require parenthesis but Reverse Polish not? I thought the whole point of Polish notation was that it didn't require parens. Unless your operators don't have a fixed arity, anyhow (Lisp's don&#x27
79.
by jkarni 12y ago
The default keyword [0] would fix the extra-type-annotations issue, no? [0] https://www.haskell.org/haskellwiki/Keywords#default
80.
Rethinking webservice and APIs in Haskell: servant 0.2(alpmestan.com)
4 points by jkarni 12y ago|0 comments
81.
by jkarni 12y ago
Agreed. I find it hilarious that a position (logical positivism, I mean) that is often criticized for failing it's own test of meaningfulness - because it lacks verification conditions - can also be criticized for being empirically f
82.
by jkarni 12y ago
I worked on a similar project: a tool that (primarily) tees it's stdin to your tty before you decide how to continue your pipe[1], which is a little more general, though it's admittedly less polished or nice. I think there's
83.
by jkarni 12y ago
What exactly are row types? I believe Ermine also features them, but searching doesn't yield very informative results.
84.
by jkarni 13y ago
But what you think of when you say "insight into the problem" might be an antiquated notion. It may be that such a thing is a rough measure of (as you almost say) how much more likely you are to prove something else (or how much e
85.
by jkarni 13y ago
I think you're wrong, but am willing to be told otherwise. Here's why: one currency is not (generally) legal tender outside its country of origin, so there's an exchange rate determined roughly by the trade balance between co
86.
by jkarni 13y ago
This isn't obvious! See my comment to IgorPartola above.
87.
by jkarni 13y ago
It is a threat! If a there are a thousand indistinguishable block chains, it's (arguably[1]) as if the number of coins on a single blockchain were much bigger, which drives down the price. Since there's an infinite supply of new b
88.
On the future of Dogecoin, BitCoin, and other cryptocurrencies of the non-realm(marginalrevolution.com)
2 points by jkarni 13y ago|0 comments
89.
by jkarni 13y ago
If I understand this correctly, is it really that worrying an argument? One can't use covert telephone recording (i.e., without two-party consent) as evidence in most states, and it's important that the person being recorded didn&
90.
by jkarni 13y ago
Looks great (and has a great name to boot). Why Hugs though?

More ›