typeholes.bsky.social
Interested in all things programming, cooking, cats.
Love to explore new programming languages and build dev tools.
467 posts
101 followers
276 following
Getting Started
Active Commenter
comment in response to
post
Well, you didn't specify it had to be sound or even mostly sound.
comment in response to
post
Have you tried gimp? There's some setup to do, but it is supposed to be workable. I don't really have the means to test it myself.
www.youtube.com/live/HMrpwta...
comment in response to
post
Mint and coconut chutneys are nice and cooling. Mango chutney is sweet. Tamarind chutney is great - sweet and sour. I'm sure there are many more I don't know about
comment in response to
post
Technically Westerner, but I grew up in Saudi Arabia. Condiments weren't used much there. Vinegary hot sauce, tahini dressing, and garlicky yogurt.
Indian cuisine has a huge variety of condiments. Many different oil based pickles, often quite spicy.
comment in response to
post
True. I've had to enact exponentially increasing ban times requiring CTO approval for override because of constant deployment requests that didn't compile. My big fear is that AI will improve to getting to a compiled and test passed but wrong state easily. AI is making people problems more painful
comment in response to
post
Unscientificly, I've tracked buzz from the trenches and it correlated strongly with breakages in the following 1-3 months
comment in response to
post
Tough questions that I don't have good answers to. I usually get shut down even trying to start the conversation and have never gotten buy in on gathering data. Self ratings of confidence in their work and a list of negative factors would be huge if we could get honest answers.
comment in response to
post
I make pancakes in a panini press. That probably doesn't help you figure out the joke, but it may be laughable on its own
comment in response to
post
Sadly these risks are the least likely to be acknowledged. Your actually cold take on change freezes is a good example. They often create compressed deadlines, bundle multiple changes into a giant pre-freeze nightmare, or require near magical solutions to respond to external change during the freeze
comment in response to
post
Generally true, but I think that the meta factor is a reasonable basis for priority. Risks that have a high probability of creating other risks should be tackled first. You won't stop your next breakage, but you can curb the increase of breakages over time.
comment in response to
post
We have those already(i and j for example)
comment in response to
post
Anecdotally from my recent interviews, junior to mid are more likely to be able to calculate big O. Sample size isn't large and probably has odd bias.
Very few at any experience could answer "why might we pick an algorithm with worse I"?
A couple got small input., none got real world factors
comment in response to
post
We should all be using programming languages designed by linguists. The P of PL has just created a mess and our only hope is a big L
comment in response to
post
These are corrective actions. Read my lips "corrective". By definition, they can't be root causes. Do another RCA and give me something I can easily make a new corrective actions for
comment in response to
post
IMO, compilers should report all errors and as a rich data structure. Which and how to display them should be the responsibility of the interface (ide, cli, etc) and customizable by the user.
In TsErr I collect diagnostics to show 1 error with multiple locations. It's really handy for refactors
comment in response to
post
Yes please. I do love being able to write code "up to isomorphism" and have it just work
comment in response to
post
I don't think that's actually evil. I'm pretty sure I've seen a language that has it and I'll remember the name several months from now
comment in response to
post
True, but I've never heard anyone advocate for I_O.
comment in response to
post
I expect the left to be more popular but I think the right is much better. I wouldn't even call it subtle. The left is almost grossly exaggerated
comment in response to
post
And snake case doesn't have rough edge cases with acronyms.
comment in response to
post
This perfectly describes KPIs vs actual employee performance. It's so common in business intelligence and very tiring to fight against
comment in response to
post
Of course, I also went down a rabbit hole of trying to figure out what I'd need from a type system to type compositions of simple structured type systems. So yeah, take my wild thoughts with a grain of salt
comment in response to
post
Yup. I was thinking that you could probably get more accurate results more easily by collecting the type annotations and control flow and punting the logic to the solver. In particular, some of the correspondence problems and indexed state issues might be solvable (pun?)
Probably not practical
comment in response to
post
Not that I know of but I thought "this really just needs to be a front end to a sat solver" many times while playing around with tsc
comment in response to
post
Stress that you are concerned about non technical issues too. Ask about things in their process that just don't make sense or seem wasteful. New users and the most veteran users are the most valuable for different reasons. Make no promises, but convey how much you care about all pain points
comment in response to
post
It's also very important to start learning before you start coding, especially for teams that serve internal users. Day in the life observation and ask tons of questions. Focus on the issues they have, especially the ones they've given up complaining about.
comment in response to
post
I truly hope it does and would be happy to do a draft review. IMO, the central concept is respecting the isomorphism between a set of binary relations and their horizontal composition. You're raising some great points in this thread. It would be nice to show they're consequences of that isomorphism
comment in response to
post
Would a query-invariant indicate that we need a union of relations with their individual constraints?
comment in response to
post
I think there is but the objects in that logic are the functional dependencies.
comment in response to
post
Writing it is often trivialized by the IDE. The killer for me is having to read through all the noise.
comment in response to
post
Gotcha. So at the very least we'd need the ordering preventing that for applications but allowing us to ignore the naming in contravariant positions. I'm sure that has gotchas I haven't thought of.
comment in response to
post
Would there be any downside to using an equivalence relation instead of equality? In other words, just ignore the name when checking?
Or maybe you need an ordering so you don't lose names on composition.
comment in response to
post
I wish we'd actually gotten sheet valued functions. Excel can actually be pretty useful.
comment in response to
post
Me desperately clinging to that one side project I actually finished and now hoping the ts-go rewrite doesn't mean I have to do an update. For now, I can ,by the thinnest of technicalities, offer this advice non-hypocritically.
comment in response to
post
That's the way. Experiment, improve, try bigger experiments. Trust your gut. Use your nose. You can learn a lot just by smelling ingredients separately and together
comment in response to
post
I guess I am into molecular gastronomy, I just never thought of it that way
comment in response to
post
It breaks down the starches and helps the proteins give elasticity when whipped. Mung beans would have been a better choice for creating a gel, but I wanted more body since I couldn't use cream. I did need to add some tapioca starch to get the texture right
comment in response to
post
I've posted this before, but this was the result. Nearly broke my arm whipping the mixture.
comment in response to
post
Reckless disregard for normal culinary wisdom, unusual flavor combinations, willy nilly fusion of different cuisines, and generally just bizarre ideas. For example, I wanted to make a vegan gelato and my first thought was that I needed to ferment some pigeon peas
comment in response to
post
It's surprising that there aren't any online communities that really get into talking about cooking. Or I suck at finding them . The reactiflux discord food channel is decent. Feel free to hit me up with questions or ideas. Fair warning, my cooking style is best classified as insanity.
comment in response to
post
Excuse me?
comment in response to
post
You can npm install a package with a file path on your local
comment in response to
post
Sadly you can substitute math with almost any subject and the answer is still no. Everything online suffers from the prevalence of the fringe. You just have to get used to filtering and ignoring.
Ideally we'd all contradict the nonsense but then there'd be no time or space for the actual subject
comment in response to
post
Thanks. Yes there's a similar vibe and I dig having a unifying theory.
Still, it builds up from prior knowledge. I want to focus on the mechanical aspects and only give names to things or actions so the reader can relate it to other works per their interests. I'm not sure it's a useful approach yet