On Mon, Jan 19, 2026 at 12:32:40AM +0900, Henson Choi wrote:
> I'm writing to document the detailed design of context absorption
> optimization in Row Pattern Recognition (RPR). This is mainly for
> future reference - honestly, I'm afraid I'll forget these details
> myself in a few months, and I'd rather have them written down
> somewhere public than buried in my notes. I don't want to repeat the
> mistake of famous mathematical theorems
> where brilliant ideas were lost because "the margin was too narrow
> to contain the proof" (yes, I'm looking at you, Fermat!). I certainly
> don't want future PostgreSQL hackers to spend the next 358 years
> trying to re-implement context absorption because the design rationale
> was lost. Fortunately, the PostgreSQL community has provided us with
> sufficient digital margins. :)
Note: this message is clearly AI-generated to me, and these are
usually super sloppy. (I'd suggest to stop posting semi-generated
messages as they tend to be made of mostly unhelpful contents as they
lack judgement and context, but there's no rule against that AFAIK
around here, as well.)
--
Michael
