Michael Aufreiter
@michaelaufreiter.com
Worked on svedit.dev all summer! Now building editable.website v2 on top of it.
My homepage joined a social network → michaelaufreiter.com.
Previously: postowl.com and letsken.com.
My homepage joined a social network → michaelaufreiter.com.
Previously: postowl.com and letsken.com.
This was possible thanks to browser vendors implementing the evolving Input Events Level 2 spec.
Thanks to everyone involved for putting in the work to get from "messy" (pre 2020) to "almost there" (today).
w3.org/TR/input-eve...
Thanks to everyone involved for putting in the work to get from "messy" (pre 2020) to "almost there" (today).
w3.org/TR/input-eve...
November 7, 2025 at 10:59 PM
This was possible thanks to browser vendors implementing the evolving Input Events Level 2 spec.
Thanks to everyone involved for putting in the work to get from "messy" (pre 2020) to "almost there" (today).
w3.org/TR/input-eve...
Thanks to everyone involved for putting in the work to get from "messy" (pre 2020) to "almost there" (today).
w3.org/TR/input-eve...
We’ll be there (Linz).
October 31, 2025 at 10:39 PM
We’ll be there (Linz).
Context: github.com/michael/web-...
Impossible to cancel inputs when they are starting character composition (IME) [BLINK, WEBKIT] · Issue #6 · michael/web-editing
composition-events-leak.mov Why is this a problem? We (and possible a lot of other apps) have no way to prevent unintended DOM changes when the user presses a dead key when a contenteditable has a ...
github.com
October 30, 2025 at 3:23 PM
Context: github.com/michael/web-...
But Svelte feed is a good start, I didn't pin that one yet.
October 29, 2025 at 5:28 PM
But Svelte feed is a good start, I didn't pin that one yet.
Haha! Good point, hehe.
October 29, 2025 at 5:24 PM
Haha! Good point, hehe.
Glad you like it! :)
And yeah, why not. Wanted to visit a Svelte meetup in Austria for the longest time. Would gladly present. Thanks for the invite!
And yeah, why not. Wanted to visit a Svelte meetup in Austria for the longest time. Would gladly present. Thanks for the invite!
October 28, 2025 at 2:14 PM
Glad you like it! :)
And yeah, why not. Wanted to visit a Svelte meetup in Austria for the longest time. Would gladly present. Thanks for the invite!
And yeah, why not. Wanted to visit a Svelte meetup in Austria for the longest time. Would gladly present. Thanks for the invite!
You can install via npm into a new SvelteKit project.
Check: github.com/michael/sved...
(the github.com/michael/hell... repo is your starting point)
Check: github.com/michael/sved...
(the github.com/michael/hell... repo is your starting point)
October 27, 2025 at 7:06 PM
You can install via npm into a new SvelteKit project.
Check: github.com/michael/sved...
(the github.com/michael/hell... repo is your starting point)
Check: github.com/michael/sved...
(the github.com/michael/hell... repo is your starting point)
... but haven't studied all the engineering practices that come with types, and it can be a bit scary to look at TS Code for those folks.
So with JSDoc, users of EW can choose if they want to do full-on engineering with types and TypeScript tooling on, or go cowboy-style JS like in the old days. 🤠
So with JSDoc, users of EW can choose if they want to do full-on engineering with types and TypeScript tooling on, or go cowboy-style JS like in the old days. 🤠
October 27, 2025 at 4:20 PM
... but haven't studied all the engineering practices that come with types, and it can be a bit scary to look at TS Code for those folks.
So with JSDoc, users of EW can choose if they want to do full-on engineering with types and TypeScript tooling on, or go cowboy-style JS like in the old days. 🤠
So with JSDoc, users of EW can choose if they want to do full-on engineering with types and TypeScript tooling on, or go cowboy-style JS like in the old days. 🤠
... but I'm also working on a full-fledged Svelte app-template (editable.website v2). However, even here I think I'm going to go with "types as comments". Why? Because I want not-so-technical people to mess with the code, and I know a lot of people understand HTML and a bit of JavaScript...
October 27, 2025 at 4:20 PM
... but I'm also working on a full-fledged Svelte app-template (editable.website v2). However, even here I think I'm going to go with "types as comments". Why? Because I want not-so-technical people to mess with the code, and I know a lot of people understand HTML and a bit of JavaScript...
Thank you for your thorough response, as a heavy user of both JSDoc & TS!
Didn't mean to play down the efforts of TypeScript folks, and I acknowledge there are different preferences.
I'm maintaining a library (svedit.dev) where JSDoc feels like the right fit...
Didn't mean to play down the efforts of TypeScript folks, and I acknowledge there are different preferences.
I'm maintaining a library (svedit.dev) where JSDoc feels like the right fit...
October 27, 2025 at 4:20 PM
Thank you for your thorough response, as a heavy user of both JSDoc & TS!
Didn't mean to play down the efforts of TypeScript folks, and I acknowledge there are different preferences.
I'm maintaining a library (svedit.dev) where JSDoc feels like the right fit...
Didn't mean to play down the efforts of TypeScript folks, and I acknowledge there are different preferences.
I'm maintaining a library (svedit.dev) where JSDoc feels like the right fit...
Thank you @paolo.ricciuti.me for inviting me over!
October 27, 2025 at 4:00 PM
Thank you @paolo.ricciuti.me for inviting me over!
Those are two separate topics.
I'm just debating the syntax part, maybe there are TypeScript folks who feel the same about that choice early on (JSDoc support came later).
Was it the right choice or a mistake?
Maybe TS folks know exactly why a new syntax has to exist — I'd like to know why.
I'm just debating the syntax part, maybe there are TypeScript folks who feel the same about that choice early on (JSDoc support came later).
Was it the right choice or a mistake?
Maybe TS folks know exactly why a new syntax has to exist — I'd like to know why.
October 26, 2025 at 1:54 AM
Those are two separate topics.
I'm just debating the syntax part, maybe there are TypeScript folks who feel the same about that choice early on (JSDoc support came later).
Was it the right choice or a mistake?
Maybe TS folks know exactly why a new syntax has to exist — I'd like to know why.
I'm just debating the syntax part, maybe there are TypeScript folks who feel the same about that choice early on (JSDoc support came later).
Was it the right choice or a mistake?
Maybe TS folks know exactly why a new syntax has to exist — I'd like to know why.
Maybe. :)
And again, seeing that the TC39 propsoal suggests "Types // as comments" is another hint to me that .ts *was* a mistake, no?
tc39.es/proposal-typ...
And again, seeing that the TC39 propsoal suggests "Types // as comments" is another hint to me that .ts *was* a mistake, no?
tc39.es/proposal-typ...
TC39 Proposal: Types as Comments
Reserve a space for static type syntax inside the ECMAScript language. JavaScript engines would treat type syntax as comments.
tc39.es
October 25, 2025 at 8:39 PM
Maybe. :)
And again, seeing that the TC39 propsoal suggests "Types // as comments" is another hint to me that .ts *was* a mistake, no?
tc39.es/proposal-typ...
And again, seeing that the TC39 propsoal suggests "Types // as comments" is another hint to me that .ts *was* a mistake, no?
tc39.es/proposal-typ...
The need for TSDoc is a symptom of a bad decision, no? For what you want to express to document code, inline type definitions are not enough anyways?
October 25, 2025 at 6:21 PM
The need for TSDoc is a symptom of a bad decision, no? For what you want to express to document code, inline type definitions are not enough anyways?