(normally) - in two flavors. Used only to prioritize high-traffic pages for dev work, iff JS and tracker enabled. The former/defunct training site's id: ga('create', 'UA-52578333-1', 'auto'). The book site's id, now used for all: ga('create', 'UA-52579036-1', 'auto'). JS is also used if enabled for thumbspage galleries, Top buttons, TOCs, etc. Jun-2019: tell Google Analytics to anonymize IP address on receipt for privacy. More: https://developers.google.com/analytics/devguides/collection/analyticsjs/. Oct-2022: install a GA4 tag (i.e., script) alongside the prior UA tag. UA tags will stop collecting data in 2023. It's okay to use both until the cutover (and Google advises doing so), but the UA tag will be pointless after that and should be removed. This handles genhtml-insert clients; others get tags by running insert-analytics.py in package-publishing scripts. Also per Google: "In Google Analytics 4, IP anonymization is not necessary since IP addresses are not logged or stored". Let's hope so... -->
Shockingly, a few posts I made in 1991 on comp.lang.python can still be had at the following direct links today (the formerly used search URL on groups.google.com seems to fail on mobiles unless a desktop page is requested):
Prolog is a logic-programming language, in which computation takes the form of automated deduction. You can find a decent (if potentially biased) overview of it on Wikipedia.
I explored optimized implementations of Prolog before, during, and after grad school, and coded a complete implementation in 68000 assembler and C, which did normal and delayed tail-recursion optimization, sported variable modes and modules, resorted to CPU-cycle counts for speed, and ran quicker than the version used by the university at the time.
But please don't ask me to explain what the posts meant...