You're by far and away my source of truth for rug knowledge. I'm sure the other Bluesky experts are good, but you're the go-to!

I make it that it's either very expensive, or sometimes very cheap (but never a middling cost for some reason), to find information that isn't really just commodity cosplaying as information.

(This thought is unrelated to any three-character acronyms in the world out there).

Does the error happen after you've done an update?

This is damn nice

Sorry, what I'm trying to say is generated code (without regular input from me on design) can tend to paint me into a corner. So I'd like some higher-level functions that can both guard against and infer when that's happening ahead of time. That'd be a nice language feature.

I'd ask what they saw on their way through their tunnel that got them to this point. What gives someone the belief to even try (no matter how surreal that "try" comes across) is the intrigue for me.

I'd want a language compiler that refuses to build, when it infers the context width of any path within the build is too narrow. Helps to prevent flattening the context of your generated code, unless you explicitly opt-in.

Then learn higher-order functions around that compiler.

Yeah that's annoying! I bet nailed your demo in front of the clients though.

Did the switch to Superforms for Svelte help the other issues or you guys just felt like it?

This was a nice read. It seemed to me like you were describing flight-recorder black boxes on commercial aircraft, with the contravariance. But later you said you wanted to shape a feedback loop out of it. I'll have to read it again in time.

I know this is a while back now, but was this in C?

I remember hitting this problem a couple of years before Swift 6 came out! I have to say UIKit's methods saved me time (and I'm not a "UIKit is better" evangelist)

I've seen you archived this since, but I think it was well worth the shot. It actually goes against an approach I was taking, where I was trying to tweak with just primitives and objects. I need to accept that FTS5 is the go-to for many reasons.

number 2 or 3 has my vote

Hahaha! Yeah the Saville Row of tech stacks, whispering to itself "any minute now" while it waits for that Beatles performance to hit its rooftops.

The Blade Runner music is already playing in my head at your last sentence!

This sounds like what I've been betting on: that you can use stores of strings/descriptions to build up personas in teams. It's nothing fancy, but I'm betting it can work if the group of people is no bigger than ~15 and the expectations they bring to the group table are not too wide of a set either.

Ah yeah I know the feeling, it happens! I'll still be on the lookout for Svelte dataviews in the meantime

Did this get published? It would be good news for the Svelte 5 dataview hunters