Discussing telemetry, opt-in vs. opt-out and other related topics made me remember that there once was an effort to standardize opt-out via a globally recognized DO_NOT_TRACK variable instead of having each app provide their own. I like that idea. But "do not track" assumes the purpose is tracking, and that doesn't seem right to me. DISABLE_TELEMETRY would have been much better? Oh well, I guess it's moot at this point, as it never seemed to take off.
@anderseknert exactly. It would be great to use it to decide how much telemetry to send. Keeping anonymous vs reasonable tracking.
I’m frustrated by Apple’s tracking because if a user has opted out of tracking, the monitoring SDK I use won’t send anything. Not even crash reports.
Which means if I play by the rules and ask the user for diagnostic tracking, I’m just less likely to get data I need. I’m better off not implementing their tracking transparency.
In one of my apps, I'd like to ensure that users are running with the latest version, and recommend an upgrade if they aren't. Is that type of "telemetry" OK? Code is #OpenSource so it'll be easy to verify that it only calls GitHub to compare current version vs last released. Conditions:
Nothing is "sent" other than a GET request to GH to get latest version.
No data is collected — I don't even run the server.
On Swedish Mac keyboard layout, press shift + 0 to type "="
Just left of that is of course number 9, and shift + 9 is ")"
And in both #golang and #rego you happen to type ":=" all the time to express assignment. And I always fat finger press shift 9 instead of shift 0, meaning I end up with:
:)
instead of
:=
Which while annoying, never fails to put a smile on my face.
Spent some extra time today responding to an issue a user reported, as I noticed it was their first interaction with the project. Life as an #OpenSource maintainer may not be glamorous, but when that extra effort results in feedback like this later... incredibly rewarding.
I know I'm not in the majority, but one of my main gripes with #kubernetes is really more a failure of #golang, and how they never managed to get extensions/plugins right. Eventually they just gave up on the idea altogether. Which means every extension point in kubernetes involves calling a service rather than implementing an interface. And besides the performance/availability impact of that, also means you're limited in what you can extend, unless you feel like maintaining a k8s fork.
10 years since Google made the first commit in #Kubernetes. And ten years of Google mostly running something else for their own services. I don’t work that much on infra these days, and Kubernetes sure has a fuckton of warts. But I’ve gotta say, compared to what we had before, and the way it transformed the industry… “game changer” is for once not an exaggeration.
Just released #Regal v0.23.0, featuring 3 new linter rules and greatly improved completion suggestions from the language server. Try it out in the #VSCode extension for #OPA, or any other editor supporting language servers. Working with #Rego has never been easier!
Home alone with the kids this evening, and we’ve been waiting for the ice cream truck. Which back in my day meant sitting like sheep by the window waiting for hours. My kids keep track of its exact GPS position, and know exactly when it will arrive. And now daddy’s €70 poorer.
Starting the day by looking at an issue reported by a user where the error message says "unreachable code". And a narrator voice in my head saying "The code was not unreachable".
Side project reveal! For the past couple of weeks, me and two colleagues have hacked on a #OPA / #Rego extension for the #Zed editor. Using the #Regal language server we built previously, it's been a surprisingly smooth ride. And a fun one!
A few little things left to fix, but we'll be publishing this in the next couple of days, and I'll be sure to announce it here. In the meantime, I'd be very happy if you could give the project a ⭐️ to help with the launch!
@nlovsund aha, sounds like that old trick from school where you’d break a pen in the lock to have class cancelled 😄 but no, I don’t think it was that. Just the teachers having some training.
@anderseknert Found out why now: "Det var den mänskliga faktorn. Ett feltryck gjorde att flera kortläsare slutade att fungera, man kom helt enkelt inte in i flera lokaler, bland annat förskolor."
I’m not gonna break it to the kids, but it’s also most likely their last. There just aren’t that many railway light up rainbow sets coming later in life.