Some core algorithms would benefit from non-normative explanations of what they do

[domenic]

The algorithmic sections of this spec are top-notch in precision and clarity and use of spec language. But for some people skimming to understand the overall intent, the equivalent of "code comments" would be helpful. In particular:

https://patcg-individual-drafts.github.io/topics/#periodically-calculate-user-topics-header does a lot of complicated time calculations. If it'd be possible to summarize the usual impact, e.g. "this roughly schedules topic calculation every N days, unless (X) in which case (Y)" then that would be helpful. Put it in a <div class="note"> to be clear it's non-normative.

Similarly, https://patcg-individual-drafts.github.io/topics/#calculate-the-epochs-for-caller is doing something about a delay and a "switch time", which for some reason is calculated with HMACs. Summarizing the intent there would be nice.

https://patcg-individual-drafts.github.io/topics/#calculate-the-topics-for-caller does some randomization and some stuff that seems caller-dependent. Calling out how that works would be helpful.

[domenic]

https://patcg-individual-drafts.github.io/topics/#append-or-modify-a-request-sec-browsing-topics-header-header would also benefit from an example resulting Sec-Browsing-Topics header.