webp is a fine format, blame the websites that disallow webp upload, but then proceed to convert the image to webp anyway

pasting the quote here because the post title is confusing

“If we automate the whole world and only a few people have the means with which to more than just survive, but to live well, have meaningful lives, there’s a big problem, a huge problem coming down the line,” the Pope said per the National Catholic Reporter.

I’ll just add that a white paper with technical information about how the product works is actually valuable. A white paper that reads like an ad in the form of an infographic is a waste of time.

glad to know it’s not just me

maybe they could be sorted alphabetically to give you an idea, but yeah, it’d be harder to know for sure without a mixed format like

worda:wordb::f1

exactly, I also like this peace of mind for my home network and see no benefit in using ipv6 there. Similarly for any VPC I deploy to an IaaS.

I thought they only banned talis

refusal to allow useful endpoints that aren’t sufficiently restful

And there are good reasons for that, GraphQL-like endpoints seem great to use, but are often a bad idea. The more freedom is given through an API, the less guarantees one can deliver. Security, scalability, and maintainability all become more difficult for APIs with endpoints that attempt to do several things at once.

But most importantly, REST doesn’t tell you exactly how to build your endpoints, as long as they’re stateless, cacheable, and refer to system resources with enough context to allow their direct manipulation.

These are good principles for older and modern web apps, that hasn’t changed. In fact, one can argue that the larger and more complex the system the more important it is to simplify its endpoints. And you can build pretty complex systems while following these criteria.

And what’s inherently new in modern applications? We’re transferring state and operating on resources just like we used to do. Most web apps are variations of CRUD.

Instead you could have a function, say t(“Ciao”) that kinda runs something like (of course loading all the translations in ram at startup and referencing that would be better than running a query for each and every string).

this backfires when the same text translates to different strings depending on the context

e.g. EN “Play” may translate to PT “Jogar” (as in play a game) or “Reproduzir” (as in play a video)

for reference, that’s usually a ISO 639-1 combined with ISO 3166-1 alpha 2

and if an exact locale match is not available, it makes more sense to return another language match than the default language fallback

you’d rather have no responses following a standard rather than only some doing that? No, thanks.

returning a 400 never prevented me from adding more info to the response

you must find it really annoying to learn Spanish, German, Portuguese, Italian, and a bunch of other languages that have gendered nouns.

or “oh yup, just need to give it a kick”

used OpenSearch in a recent project, but the number of annoyances with it are through the roof. From SSL certs setup to bad defaults in settings, and the fact it does type inference for indices requiring you to manually recreate the index, and the docker container that takes 30s to start every time…

If you can use mongo, just use that. Or pick something other than OpenSearch if that’s overkill for you.

but from the floor, yummy

I’ll fix it in an hour. When I get to it in a couple of weeks.

yeah, it’s unfortunate the general population has no awareness of how insecure email is

Distros may not update software versions when backporting some things, meaning they add a suffix they control to the version e.g. 2.4.57-ubuntu1.2 whatever, but the version reported by the software itself might still be 2.4.57.

It depends on the release process. I was also confused once I was asking myself why the repo was reporting a CVE as fixed when it still showed the old version.