import-maps VS fresh

Long time huge fan of JS. I appreciate your calling out the multi-paradigm aspect; having these first class functions & prototype based inheritance has been so flexible.

TC39 has done a great job shaping the language over the years. New capabilities are usually well thought out & integrate well. Async await has been amazing.

The one major miss that makes me so sad and frustrated is modules; js has gotten better everywhere except it's near requirement for build tooling. Being able to throw some scripts on a page and go is still an unparalleled experience in the world, is so direct & tactile an experience. EcmaScript Modules was supposed to improve things, help get us back, but imports using url specifiers made the whole thing non-modular, was a miss. We're still tangled & torn. Import-maps has finally fixed but it's no where near as straightforward, and it still doesn't work in workers, which leaves us infuriatingly shirt of where the past was. https://github.com/WICG/import-maps/issues/2

makes sure your app is getting the download it expects. Adoption is probably pretty minimal though. https://developer.mozilla.org/en-US/docs/Web/Security/Subres...

I think the big thing making this unlikely though is that very few folks use cdns these days. We designed ESM as a module system for the language, but then took a good fraction of a decade to build import-maps, to let us actually use modules in a modular way. Good news, we can finally use modules modularly! https://caniuse.com/import-maps

Bad news? Oh import-maps only works for the simplest case. Doesn't work in webworkers/service workers. https://github.com/WICG/import-maps/issues/2

The point is that single page apps almost always are bundled together, as using CDNs hasn't even been technically possible.

Also, CDNs are kind of somewhat pointless, now that http caches are partitioned by origin (for security reasons). They might have better anycast infrastructure to get the content out faster, but without the caching there's no inherent advantage. The user will download the same jquery file again in each site they go to, no already having it cached anymore. Bah humbug!

https://github.com/WICG/import-maps/issues/2

Bare specifiers has been the tragedy of ESM. Nice module syntax... that is utterly u deoyable & which has had to have awful de-modularizing specifiers hard-coded into each file to make it work. Abominable sin to introduce "modules" to JS/es2015 then spend a decade dragging everyone along with no story for how to have modular modules.

Import-maps are like "here" to fix this on the web... finally... except they only are shipping to the happiest sunniest easiest case, with Web Workers being totally shit out of luck in spite of some very simple straightforward suggested paths forward. https://github.com/WICG/import-maps/issues/2

I think Deno is making pretty good tradeoffs along the way here. This looks like package.json at surface level, but there is a nightmare of complexity under the surface. Typescript, ESM, cjs all have various pressures they create & in Node it's just incredibly tight & tense dealing with packaging, where-as Deno's happy path of Typescript first does not slowly tatters one over time. It really has been super pleasant being free of the previous world, and having something much more web-platform centric, more intented, with less assembly & less building, and more doing the actual coding.

I really hope import-maps eventually get broader support. Maybe this long-dwelling webworker issue should be brought up with WinterCG.

Import maps proposal

The concept of Import Maps was born in 2018 and made its long way until it was declared a new web standard implemented by Chrome in 2021 and some other browsers.

For package managers, you can use something like import maps to let the user specify which path points to what package, and resolve it properly.

Huh. I was about to complain that this breaks with web standards, but apparently it's being proposed as a standard feature: https://github.com/WICG/import-maps

Interesting!

Deno: Deno with one of it's frameworks (like Fresh

The road to Fresh 2.0 (https://github.com/denoland/fresh/issues/2363) by Marvin Hagemeister Can't wait for seeing the end of the road! All in all great changes ahead.

Because I have secret magical powers that you probably don't, it's basically free for me. Here's the breakdown though:

The application server uses Deno and Fresh (https://fresh.deno.dev) and requires a shared-1x CPU at 512 MB of ram. That's $3.19 per month as-is. It also uses 2GB of disk volume, which would cost $0.30 per month.

As far as post generation goes: when I first set it up it used GPT-3.5 Turbo to generate prose. That cost me rounding error per month (maybe like $0.05?). At some point I upgraded it to GPT-4 Turbo for free-because-I-got-OpenAI-credits-on-the-drama-day reasons. The prose level increase wasn't significant.

With the GPU it has now, a cold load of the model and prose generation run takes about 1.5 minutes. If I didn't have reasons to keep that machine pinned to a GPU (involving other ridiculous ventures), it would probably cost about 5 minutes per day (increased the time to make the math easier) of GPU time with a 40 GB volume (I now use Nous Hermes Mixtral at Q5_K_M precision, so about 32 GB of weights), so something like $6 per month for the volume and 2.5 hours of GPU time, or about $6.25 per month on an L40s.

In total it's probably something like $15.75 per month. That's a fair bit on paper, but I have certain arrangements that make it significantly less cheap for me. I could re-architect Arsène to not have to be online 24/7, but it's frankly not worth it when the big cost is the GPU time and weights volume. I don't know of a way to make that better without sacrificing model quality more than I have to.

For a shitpost though, I think it'd totally worth it to pay that much. It's kinda hilarious and I feel like it makes for a decent display of how bad things could get if we go full "AI replaces writers" like some people seem to want for some reason I can't even begin to understand.

I still think it's funny that I have to explicitly tell people to not take financial advice from it, because if I didn't then they will.

Deno has also created a Next.js competitor, Fresh. I found it a few weeks ago and am starting to go through the docs, looks like a good overall concept. https://fresh.deno.dev/

https://fresh.deno.dev/ - Fresh embraces the tried and true design of server side rendering and progressive enhancement on the client side.

Along the way, I not only got the oppurtunity to revise old concepts that had blurred in my memory, but also learnt about new technologies like Fresh.js, a framework from Deno (a js runtime engine) that uses Preact, a React Routing library and used Chakra UI for the first time.

A whole project might be released as a server or framework. Frameworks like fresh, and astro) both have had things deep within them that I've wanted to reuse, within fresh it's the esbuild configuration, and islands functionality, and within astro it's the rendering of astro files themselves.

The Fresh framework by Deno cited an improved developer experience due to tighter feedback loops.