Sabitlenmiş Tweet
My guest appearance on fulltimenix is out now.
fulltimenix.com/episodes/horiz…
English
Daniel Firth (Inspector GADT)
5.9K posts
Daniel Firth (Inspector GADT)
@locallycompact
Haskell, Nix, Mathematics @ https://t.co/v3jt0w7n91
Katılım Ağustos 2012
803 Takip Edilen1.4K Takipçiler
@johnzabroski Tldr stackage has a chicken and egg problem with regards to library maintenance because it's so heavily used by library authors themselves in order to supply the manifest that they would use to execute their compatibility upgrades. Nix gives you *some* options to bypass that.
English
@locallycompact Inspector GADT is such a great nickname. If I worked with you, I would feel obliged to call you that at all times. But I am having a hard time seeing where you speak directly about these limitations? Where is the actual transcript? Or do I have to listen to all 46 minutes?
English
My guest appearance on fulltimenix is out now.
fulltimenix.com/episodes/horiz…
English
But it's morally equivalent to this:
```
compiler "ghc-9.6.4"
modifiers $ do
doJailbreak True
package "lens" $ do
url "http://..."
rev "...."
package "aeson" $ do
...
```
Which is a valid free monad and naturally has any number of interpretations. So why isn't that first class data with multiple outputs?
English
because the data structure is syntactic in your example. cabal can do this because the output structure is set in stone. The type definition for cabal files is in a completely different language hidden from the user. We also cannot have functions that return arbitrary bits.
Part of the power in configuration management is the ability to yield more than one output form the same code, which means data must be first class.
Harney, CO 🇺🇸 English
Paradox is ready for its initial use case. The code is now Open Source and ready to play with.
Configuration Management!
paradox-lang.org/config/
gitlab.com/paradox_labs/p…
English
@Ngnghm @sol_plunder @satisfiesvalues @msimoni Stackage has fundamental limitations that I outlined here if you're interested.
x.com/i/status/19178…
Daniel Firth (Inspector GADT)@locallycompact
My guest appearance on fulltimenix is out now. fulltimenix.com/episodes/horiz…
English
@sol_plunder @satisfiesvalues @msimoni And never would be, especially as Stackage does not include most of the cryptocurrency ecosystem I was developing on.
English
Static types catch errors early and that's great—but they also catch non-errors early, preventing you from writing the software you want—and that's terrible.
Those who only tell you about one side of the tradeoff, or claim the other side is universally negligible—are dishonest.
English
@fresheyeball Rather than all these nested ADTs I'd like to just assume more in the interpretation.
```
compiler: "ghc-9.6.4'
modifiers:
doJailbreak: True
package "lens" ...
package "aeson" ...
```
Sort of more like cabal.project.
English
config: HorizonExport
makeOverlay
compiler: "ghc-9.6.4"
packages:
[ callHackage
name: "aeson"
version: "2.2.3.0"
, callGit
name: "my-lib"
url: "github.com/foo/my-lib"
revision: "abc123"
]
is the best I can do right now.
Harney, CO 🇺🇸 English
@fresheyeball Something like
package "lens"
hackage
name "lens"
version "5.1"
package "foo"
git
url "http://..."
Maybe
English
@locallycompact The system is extremely flexible to design changes. So I wouldn't mind making this a priority if it moves the needle for you.
English
@fresheyeball That's the dhall type definitions repo and that's what it looks like in dhall. Will call you this week.
English
@locallycompact I am sorry, I can't intuit your intentions. Give me a call.
English
QME
@locallycompact I am confused. That is not valid syntax. What does it look like in Dhall?
English
They translated this directly from dhall into paradox and it's more verbose than the dhall.
```
config: HorizonExport
HorizonExport.makeOverlay MakeOverlayArgs:
compiler: "ghc-9.6.4"
packages:
[ Package.callHackage CallHackageArgs:
name: "aeson"
version: "2.2.3.0"
, Package.callGit CallGitArgs:
name: "my-lib"
url: "github.com/foo/my-lib"
revision: "abc123"
]
```
English
I refuse to use any open source software that doxes users. Also systemd has always and forever sucked.
Lain on the Blockchain@CryptoCyberia
wow, systemd is adding an age variable for future digital ID laws. Linux users are not going to like this.
English
@fresheyeball Did they get the assessment right? The CallHackageArgs spam seems a little verbose.
English
@locallycompact Let me know any issues or if you run into anything. It's still Alpha. I expect we'll run into something and I'll need to fix it.
Harney, CO 🇺🇸 English
An old Geordie proverb: One man's chicken dinner is another man's bath.
English
English
HOLY FUK I JUST LEARNED ABOUT TLA+ AND IT'S SO GOOD FOR AGENTIC CODING
ur telling ME that i can mathematically fact check every possible scenario of my design STATE to prevent bugs and crashes
AND IF IT FINDS SOMETHING THE AGENTS GET INSTANT FEEDBACK AND LOOP FIXING IT TILL IT ALL POSSIBLE BUGS IN THE DESIGN ARE PATCHED
LOL THIS IS OP
English
@SanjayNadaraja @fresheyeball I looked up Koka and Flux and they have nice hooks but then the syntax is all not Haskelly what's with that?
English
@fresheyeball Not as a core feature, correct?
Like compared to Koka, Flix, and Unison.
English
@fresheyeball We don't need LLMs to be deterministic. We only need to stop relying on the LLM after you successfully got it to make a tool that is deterministic.
English
@kerckhove_ts How would you feel if the tests hadn't failed to establish that the code did not work this morning?
English
IME the biggest reason that communication about testing fails is that people don't understand the difference between:
"The tests verify that the code works."
and
"The tests fail to establish that the code does not work."
English
@kerckhove_ts Funny initialed name I didn't find you
English
I really like this genre of post: "Thing I believe explained better than I could explain it in a one-sentence message on slack". It lets me send a prepared explanation to someone who hasn't formed an opinion yet.
Tom Sydney Kerckhove@kerckhove_ts
CI should fail on your machine first blog.nix-ci.com/post/2026-03-0…
English