The problem with contracts is that they (mostly) implement the function again => formal verification is hard. Let's see if it is useful to constrain stubs to sideeffectless functions over sample datapoints, with the ability to define side effects just in the contract.

I might be naive, but so far I thought everyone in programming knows what is the mathematical meaning of Boolean algebra.

I have reported an error for a big project. They stated that it is a duplicate of a bug, of which it is not. They closed the bug. I asked to reopen the bug, as it is not a duplicate. After a long while, I made a new bug. Got marked as duplicate of previous

I reached 100% mutation coverage for my current project. The last three zombies were the hardest to kill, but each of them have uncovered hard-to-reproduce sync and state management issues which otherwise would cause lot of bad times, but probably would have gone unreported ever

I see nowadays even the Gartner Hype Cycle is much faster. We are already at the through of disillusion wit LLMs.

Reading this I had the urge to make cyberpunk ICs. Cut the sources and drains into the back of a mirror, deposit CuS, and somehow place the gate electrodes on top. On the wall near a CRT, play pong on it. Too much "somehow"s in the process though. patentimages.storage.googleapis.com/fa/5d/33/ed276…

Very s(l)ick. A neural network implemented in type level typescript to distinguish numbers.

Did I mention yet that not fixing the 'this' issue (at least with classes) is the greatest sin of Javascript? And Typescript as well. They had the opportunity. Just introduce a 'self' keyword if you want to be bug-compatible, please.

What is better? Explicit dependency management, with no duplicates, ability to change implementation deep down, but verbose, or Implicit, with duplicates, less control, but more succinct?

Gödel-blaming is a sign of weakness. #stopBlamingGödel

Ok, the embrio of the UI is working on web and Android, time to add a command line widget.

The source of so much suffering.

If you also think that generating code which the compiler knows will fail at runtime is a bug, please upwote this bugreport: github.com/microsoft/Type…

In the last few days I had terrible experiences in the TS community. Lack of attempt to understand arguments and address them. And now "you should treat your unit test addiction". I hope it is just bad luck, and not the whole community is that toxic.