I’ve written code with SWI-Prolog since 1994, so it’s certainly over 30 years. Jan came over to our Uni when I had invited him to give a talk. Glad he also added multi-precision integers later, like I added myself using the GNU MP lib, stored as atoms. The foreign interface was very useful.

SWI-Prolog is excellent. The author, Jan Wielemaker, keeps expanding it to a does-it-all Prolog platform.

The author admits in the article that numerical weather forecast models are still needed. Numerical models use physics (PDE math etc). But AI is a black box, trained with historical data to extrapolate a prediction. AI is not robust and will make mistakes when measurements do not align with history.

Weather balloons are largely unnecessary. We have excellent weather models that are fed land-based data and sea-based data. So I’m so disappointed in IEEE to publish this unscientific article claiming that AI and weather balloons are better. spectrum.ieee.org/amp/ai-weath...

💔

Naïve implementation that runs in quadratic time, so won’t scale. Instead of that, use a sliding window and update the sum incrementally (subtract outgoing, add incoming). Use C++ for speed.

Addicted? Nah. Just always close to grab, like these on our counter right now.

Philips MSX New Media System

-1 1 RSHIFT . since RSHIFT is a logical shift, not arithmetic, it shifts in a 0 so to get max positive value of a signed integer. Minor caveat: Assumes 2s complement, but which is universal these days.

Hope this helps: I wrote a Forth introduction and manual for newbies to Forth, for a 2012 standard Forth implementation which is more up to date than the vintage textbook introductions github.com/Robert-van-E...

Combinations of words are included for performance, such as 1+ 2* 2/ 0= to perform two sequential operations with one word, but there are also a few to perform two operations simultaneously like /MOD (div mod) to return two results. Because most vintage Forth do not optimize code, these are useful.