Tripartite Palindromic Partition of Integer (HP 50g) Challenge

[3298]

(05-26-2023 08:52 AM)Gerald H Wrote:  Now very interested in your library compiler: library 859 has size

4110.5

but after decompiling & recompiling using native 50g tools size is

9647.5

with no (noticeable?) change to time of operations.

That's the library splitter not handling library commands nested in one another very well - because it's pretty much impossible. It's seeing that e.g. pickType is in the library's link table and extracts it to a file, but the copy embedded in fromZINT stays there. Then AlgMain is embedded in both copies of pickType, but it's also in the link table, so you get a third copy in a separate file ... that snowballs quickly.
If you were to replace all the embedded copies with ROMPTRs (or IDs which CRLIB can turn into ROMPTRs) referencing the non-embedded one, you'd likely trim it back down to ... around 4200 bytes, maybe? The library splitter's (forced) shortcomings make my tool look much better than it really is.

(05-26-2023 08:52 AM)Gerald H Wrote:  Very grateful for expansive explanation of difficulties, trust elaborating to your students clarifies points for yourself - the exegesis in itself requires more effort than I can muster.

As the kids would say: ROFL!
But you have a point, my complex writing style isn't known for being easy to understand. Too long sentences, too deeply nested subordinate clauses, etc.