Filters That Survive Being Sent to Someone Else

Explore filters every published area of the site at once: aptitude suites, skill lessons, cognitive skill labs, puzzles, coding lessons, career families and occupations, and the editorial pages. It is generated from the same registries that build the sitemap, which has a consequence worth stating: nothing can appear in Explore that is not a real page, and nothing published can be missing from it.

That is a stronger guarantee than it sounds. A hand-maintained catalogue goes stale the first time somebody publishes without updating it, and nothing fails when it does.

• Kind: the type of thing, such as a suite, a lesson, a puzzle set or a career family.
• Activity: what you would be doing, such as reading, practising, or building something.
• Difficulty: where an item sits on its published ladder.
• Duration: roughly how long it takes, for when you have twenty minutes rather than an evening.
• Career family: learning material narrowed to the kind of work it supports.
• Progress: what this browser already knows you have started.

Set a filter and it is written into the URL. Bookmark that and you have a saved view; send it and the other person gets the same narrowed catalogue. This is the sort of thing that looks like a detail and changes how a page gets used: a teacher can hand a class a link to exactly the material for one topic at one difficulty without asking anyone to reproduce a set of clicks.

One facet cannot work that way, and it is worth knowing which. The progress filter reads this browser's own local storage, so it means nothing on anybody else's device. A link carrying it will still open; it will simply not be filtered by progress for them.

Explore narrows a large catalogue. It is the right page when you already know roughly what you are after. The Tool map does the opposite: every destination on one page, grouped by what you are trying to do, with a count beside each row read from the registry that owns it.

Use the Tool map when the question is what exists rather than which of the things that exist you want. It is also the page to send someone who has never used the product and has no vocabulary for it yet.

There is a Featured page and it deliberately publishes nothing. A curated list of topics would have to be assembled by hand and kept fresh by hand, and a topic is not suitable to promote until processing has succeeded, attribution is complete and baseline coverage has passed.

Rather than fill the page with test fixtures dressed up as editorial picks, it explains the gap and points at live search. That is the same instinct as the missing uptime number: a page that says what it does not have is more useful than a page that invents something to have.

Building Explore from the registries costs something. A hand-curated catalogue could put the best material first, write a better one-line description for each entry, and quietly omit the pages nobody should start with. A derived one cannot do any of that: it shows what exists, described the way its own registry describes it.

The trade is worth it because of what the hand-curated version costs instead. Somebody has to maintain it, and the failure mode when they stop is invisible: the catalogue keeps rendering, the descriptions keep reading well, and it simply stops corresponding to the product. There is no test that can be written against a list of opinions. There is a test that can be written against a derivation, and this repository has several.

Start broad and remove rather than add. Pick the activity you are in the mood for, then cut the duration to what you actually have, and only then narrow by difficulty. Filtering by difficulty first tends to produce an empty result and the impression that a section is thin, when the real problem is that three facets were applied at once.

If a filtered view comes back empty, drop the most specific facet rather than starting again. The URL updates as you go, so backing out one step is a single click, and the facet you dropped is the one you learned something about.

One more habit worth having: when you find a combination that produces a useful shortlist, bookmark it rather than remembering it. A saved filtered view is the cheapest possible study plan, and it stays correct as the catalogue grows underneath it.