[10:13:47] Are categorizes and templates low priority? Currently, it's difficult to search for existing pages of a particular category. [11:32:42] T433157 "no current plans" (re @jsamwrites: Are categorizes and templates low priority? Currently, it's difficult to search for existing pages of a particular category.) [12:36:05] It sounds like we should find a path that does not send all the functions into code. Maybe Z40008 should not have used actual functions, but instead the new Z96 for that function? Same for Z14310 acting on Z14294? Do we need a new Type for the latter? (re @Al: There are issues with object equality but I made Z40578 with the aid of Z40560. The performance isn’t great, [12:36:05] mainly because all ...) [12:40:01] Also, while I'm in the weeds. Is this related to the reason that they require so much memory. For example Z40617 uses 2.4 Gb of memory!?! (re @u99of9: It sounds like we should find a path that does not send all the functions into code. Maybe Z40008 should not have used actual fu...) [13:13:37] Yes, I’m pretty sure that would perform better, but it’s not so easy to maintain the associations, because you’d have to look up the function’s ZID. We’d need a UI component like we have for QIDs. Memoized intermediate calls might help, but that will be quite limited initially T433272, *T436125*. Lightweight enumerations are another option, but adding a new function to [13:13:37] [13:13:38] the enumeration would currently be a “staff” action. (re @u99of9: It sounds like we should find a path that does not send all the functions into code. Maybe Z40008 should not have used actual fu...) [13:43:51] I believe so. I tried it with a 24-item (QIDs, ZID) list: [13:43:51] *Duration [13:43:53] *52 ms [13:43:54] *CPU usage [13:43:56] *126.1 ms [13:43:57] *Memory usage [13:43:59] *1581 MiB [13:44:00] Orchestration: 1186.79 MiB [13:44:02] Evaluation: 393.38 MiB [13:44:03] Execution: 0.63 MiB [13:44:05] That’s in testing, and the Orchestration looks a little on the high side, possibly from the generic function-call baggage, which memoization should address. (re @u99of9: Also, while I'm in the weeds: Is this related to the reason that they require so much memory. For example Z40617 uses 2.4 Gb of ...)