iiinova — where your digital self keeps evolving

Derk-Jan Karrenbeld @derk-jan.com · 15/11/2025

It's okay. There's much wrong with me too ;)

Derk-Jan Karrenbeld @derk-jan.com · 15/11/2025

If they actually cared deeply they'd all resign. This group of people, their actions, and everything that has come after is about the worst thing that has happened for our community in quite a while. As far as I am concerned this can burn to the ground and die.

010

Derk-Jan Karrenbeld @derk-jan.com · 02/10/2025

@joel.drapper.me Yooooo. Sorry I rolled away *if you recalled* at EuRuKo. I did wanna chat, butttt we had reservations for dinner! Thank you for all the well pointed comments on the current issue in our community, and thanks for the workshop at the conf!

020

Derk-Jan Karrenbeld @derk-jan.com · 19/09/2025

bsky.app/profile/mike... To continue reading

010

Reposted by Derk-Jan Karrenbeld

Erika @erika.omg.lol · 19/09/2025

“With the recent increase of supply chain attacks we decided to do one of our own”

21019

Derk-Jan Karrenbeld @derk-jan.com · 19/09/2025

There isn't a single apology or even recognition of bad decisions. Once again the generic "we need to move forward" bullshit. So yeah, that taste will remain for the foreseeable future.

260

Derk-Jan Karrenbeld @derk-jan.com · 22/07/2025

Accessibility isn't a feature. Something that's not accessible is unusable and that should be a major programming mistake, also known as bug.

030

Derk-Jan Karrenbeld @derk-jan.com · 22/07/2025

I stopped using BlueSky because of this. 13/15 issues (I stopped reporting after that) are still open, nothing marked as planned. One of the closed issues was silently reverted without any response from the developer who did that why they did that. Sure-fire way to make people feel second class.

120

Derk-Jan Karrenbeld @derk-jan.com · 27/05/2025

github.com/bluesky-soci... In the past 6 months, 2 issues have been closed (resolved). One by someone from the staff. The other one has silently been reverted by @hailey.at with no explanation or follow up. The fix is no longer live. See also: github.com/bluesky-soci...

github.com

Issues · bluesky-social/social-app

The Bluesky Social application for Web, iOS, and Android - Issues · bluesky-social/social-app

020

Derk-Jan Karrenbeld @derk-jan.com · 06/03/2025

It's now part of my first use process to completely remove all copilot shit. That and the shit move where they increase everyone's office 365 subscription price to have access to copilot with a pretty difficult path to downgrade...

000

Derk-Jan Karrenbeld @derk-jan.com · 28/01/2025

PSA Using Office 365? Microsoft with automatically "upgrade" you to an expensive plan including CoPilot horseshit. Follow this guide to downgrade (yes, you can only downgrade via Cancel subscription). For family plans you will save ~30+ EUR/year. support.microsoft.com/en-gb/office...

support.microsoft.com

Switching to Microsoft 365 Personal and Family Classic Plans - Microsoft SupportYour Privacy Choices Opt-Out Icon

Learn how to switch to a Microsoft 365 subscription plan without AI features, which include Microsoft 365 Basic, Microsoft 365 Personal Classic and Microsoft 365 Family Classic.

000

Reposted by Derk-Jan Karrenbeld

Adam Argyle @nerdy.dev · 14/01/2025

there are tons of color syntax's folks dont know about, check out my #CSS Color MEGA LIST 🤘🏻 codepen.io/argyleink/pe...

codepen.io

CSS Color Syntax Mega List

...

5535

Reposted by Derk-Jan Karrenbeld

Kevin Parry @kevinbparry.bsky.social · 10/01/2025

Chair Stop-Motion 🪑💨 Made with 444 pictures taken over 10 days

443320818431

Derk-Jan Karrenbeld @derk-jan.com · 31/12/2024

en.wikipedia.org/wiki/Doughnut Origin story! Oil cake/oliebol became the donut/doughnut!

en.wikipedia.org

Doughnut - Wikipedia

000

Derk-Jan Karrenbeld @derk-jan.com · 22/12/2024

It is extremely obvious that whomever handled my support requests did not even attempt to reproduce any of the issues as I've been asked to make a screenrecording "if the issue still persists". The only issues they noted for "future improvements" have already been fixed weeks ago. What's the point?

020

Derk-Jan Karrenbeld @derk-jan.com · 22/12/2024

It's now the 22nd of December. I just received a "response" to all the other support requests I made. The response ranges from "your feedback is valuable" to a proverbial please go fuck yourself. Accessibility issues are not "nice suggestions to track". Generic responses after 1 month are not okay

120

Derk-Jan Karrenbeld @derk-jan.com · 18/12/2024

Guess what. The support requests about not being able to use the site from over 1 month ago reported via Zendesk just got a generic "can you check if this still an issue" response. Imagine relying on that support.

110

Derk-Jan Karrenbeld @derk-jan.com · 08/12/2024

That said, if github.com/ariatype/ari... is unioned with the HTMLAttributes thing, ya should be there?

github.com

120

Derk-Jan Karrenbeld @derk-jan.com · 08/12/2024

We can. It's quite a bit of work though, because you should NOT at `role` to elements that have a default role (unless you want to overwrite it). So it's not a "role should be there", but it's "the computed role should be x"

github.com

130

Derk-Jan Karrenbeld @derk-jan.com · 08/12/2024

We're almost hitting 4 weeks. None of the other 13 have been updated, touched, solved, or interacted with.

110

Derk-Jan Karrenbeld @derk-jan.com · 08/12/2024

14. It's more than I expected as European :')

000

Derk-Jan Karrenbeld @derk-jan.com · 04/12/2024

More & more politics in my discover feed, and a lot of it is opinions mixed with facts. Oh oh oh...

010

Reposted by Derk-Jan Karrenbeld

Hidde @hidde.blog · 30/11/2024

✏️ Turn off AI features by default (to reduce their climate impact) hidde.blog/please-make-... new post. sorry to oversimplify but I just hope we'd all do less overcomplexifying

hidde.blog

Turn off AI features by default (to reduce their climate impact)

If users don't use a future that requires a lot of energy and water, that's wasteful. Making such features opt-in is the right thing to do.

02511

Derk-Jan Karrenbeld @derk-jan.com · 25/11/2024

Something like date.toLocaleDateTimeString(navigator.languages[0]) would automatically solve this.

100

Derk-Jan Karrenbeld @derk-jan.com · 25/11/2024

new Intl.DateTimeFormat('en-GB').format(date) navigator.languages for the list (fallback to navigator.language for older browsers developer.mozilla.org/en-US/docs/W...)

100

Derk-Jan Karrenbeld @derk-jan.com · 25/11/2024

It should display like this (en-GB):

https://skyzoo.blue/stats/derk-jan.com

with en-GB locale

100

Derk-Jan Karrenbeld @derk-jan.com · 25/11/2024

The following matches en, but not en-GB

https://skyzoo.blue/stats/derk-jan.com

with en-US locale

100

Derk-Jan Karrenbeld @derk-jan.com · 25/11/2024

From the top, en-GB should resolve to: date.toLocaleString('en-GB') '25/11/2024, 21:38:07' ...whereas en defaults to en-US, being date.toLocaleString('en') '11/25/2024, 9:38:07 PM' I should therefor see 25/11/2024, 21:38:07.

100

Derk-Jan Karrenbeld @derk-jan.com · 25/11/2024

It uses nl instead of nl-NL because Dutch (Netherlands) just like English (Netherlands) as en-NL are not offered by chrome. Chrome injects en under en-GB probably otherwise I would see french before english if a site understood en-US and not en-GB. This will break with Chinese, but whatevs.

100

Derk-Jan Karrenbeld @derk-jan.com · 25/11/2024

Ok. So my settings are based on the languages I speak: 1. en-GB 2. nl 3. fr-FR 4. fr 5. de-DE 6. de 7. en-US 8. en Chrome now translates that to the following order of locale preferences: en-GB : 1 en : 0.9 nl : 0.8 fr-FR : 0.7 fr : 0.6 de-DE : 0.5 de : 0.4 en-US : 0.3

Chrome language settings

1. English (United Kingdom)
2. Dutch: This language is used when translating pages
3. French (France)
4. French
5. German (Germany)
6. German
7. English (United States)
8. English: This language is used to display menus, settings, alerts, and other text in Chrome

100

Derk-Jan Karrenbeld @derk-jan.com · 25/11/2024

Update: so atproto DOES do this right, using did - decentralized identifiers. My "permanent" profile link is bsky.app/profile/did:.... See: atproto.com/specs/did

bsky.app

000

Derk-Jan Karrenbeld @derk-jan.com · 25/11/2024

Just leaving a reply so I can get my did the easy way. Thanks for the blog post!

020

Derk-Jan Karrenbeld @derk-jan.com · 25/11/2024

TL;DR Not theoretical, we currently already use this behaviour (both with and without 303), outside of forms. FUN!

010

Derk-Jan Karrenbeld @derk-jan.com · 25/11/2024

...break anything. In practice, you are absolutely right that *usually* you absolute want to return 303, or you may start deleting what you redirect to! A bug? No. Expected? Probably not! Hell? Yes sir!

110

Derk-Jan Karrenbeld @derk-jan.com · 25/11/2024

...which would change each time you replace/revise/create it. That now needs to 303 (or 204, or 200) in order to show something in the browser. If we could extend our forms with DELETE/PATCH/PUT, and you 302, that is currently already undefined behaviour. So *spec-technically* it would not [...]

110

Derk-Jan Karrenbeld @derk-jan.com · 25/11/2024

that behaviour is actually used by "us" purposefully. Example is you deleting an attached file. You DELETE the permalink to the file, which stays the same if the file is replaced/revised/created/whatever. It redirects (307) to a location that is the permalink for the file "instance", [...]

110

Derk-Jan Karrenbeld @derk-jan.com · 25/11/2024

Ah. Well, as you can read in the GitHub issue I linked, that's not the issue with extension. Security (same-origin policy, exceptions) is. Ignoring that, yah, if you redirected 307 explicitly, using PATCH/PUT/DELETE, you'd be deleting, putting, or patching whatever you redirect to. That said [..]

100

Derk-Jan Karrenbeld @derk-jan.com · 24/11/2024

A compliant server would do the same (use the same verb in case of a 307 and 308, and change the verb to GET on 303, as specced). Can you elaborate further so I can (probably) be as annoyed as you are about whatever other thing we devs have to deal with?

100

Derk-Jan Karrenbeld @derk-jan.com · 24/11/2024

If that's what you meant, then I misinterpreted your post, as form submissions are never sent with fetch, unless intercepted, but alas. I don't understand your fetch-spec comment Point 12 is correct for 303, 307, and 308, as specced in HTML (separately from fetch) and the implementation matches.

100

Derk-Jan Karrenbeld @derk-jan.com · 24/11/2024

Rails specifically not working (redirecting) on a non-303 is not rails, but Turbo's shit implementation and imo incorrect handling. It can also break if you use action.haml instead of action.html.haml which is caused by an incorrect addition to support turbo streams. See: github.com/hotwired/tur...

github.com

Turbo new install - forms do not redirect · Issue #122 · hotwired/turbo-rails

I just installed Turbo in an existing Rails 6 app and all my existing forms are broken. They submit as expected but the redirect after doesn't happen. I'm able to interact with Turbo as expected - ...

100

Derk-Jan Karrenbeld @derk-jan.com · 24/11/2024

You don't need to return a 303. Forms will work fine if you 200, 201, 202, 303, 307, or 308. To redirect properly, you want 201, 303, 307, depending on your use case. 308 was introduced because 301 has ambiguities. 307 was introduced because 302 also has ambiguities.

100

Derk-Jan Karrenbeld @derk-jan.com · 24/11/2024

However, in August this was revisited as the Private Network Access and Navigation standards were updated. There is now this proposal that has recently been updated: alexanderpetros.com/triptych/for...

alexanderpetros.com

Support PUT, PATCH, and DELETE in HTML Forms

010