How it works

Why the Featured Page Is Deliberately Empty

There is a page called Featured and it publishes nothing. Filling it would have been half an hour's work, and every option for filling it was worse than the gap.

4 min read

Applies to Novus Learn 0.1.0

Illustration: Candidate topic rows meeting a barrier, with one passing through to a published panel and the rest crossed out.

An empty page is a decision

Copy link

Most sites have a featured row. It signals activity, it gives a new visitor somewhere to click, and it is the single easiest piece of a product to fill. This one is empty, and the page explains why rather than quietly not existing.

The short version: a topic is not suitable to promote until processing has succeeded, attribution is complete, and baseline coverage has passed. Until a set of topics has cleared those checks there is nothing honest to feature, and a page that says so is doing more work than a page of plausible tiles.

The three ways it could have been filled

Copy link
  • Test fixtures presented as editorial picks, which is the version that ships most often and is the least defensible.
  • A hand-typed list of popular topics, which is correct on the day it is written and stale thereafter.
  • An automatic list of whatever was opened most, which promotes whatever a crawler touched.

Why the fixture version is the worst one

Copy link

Development fixtures exist to exercise a pipeline, and they look exactly like real content to a visitor. Shipping them as featured picks means a new reader's first impression of the product is a page of material nobody chose, presented as material somebody chose.

It is also invisible in every check. No test fails, the build is green, and the page looks finished. That combination is precisely why it needs a stated rule rather than an intention.

What to use instead

Copy link

Live search is the real entry point: any valid Wikipedia topic can be opened, in any supported language edition, and the coverage state of what you get is shown on the page rather than promised in advance. Explore is the other one, filtering everything the site actually publishes.

Between them they answer the question a featured row is pretending to answer, which is where do I start, with the difference that neither of them is asserting that a particular topic is worth your time on no evidence.

If what you want is somebody else's judgement about where to begin, the tutorials are the honest version of it. Each one walks a complete journey that works today, which is a recommendation somebody actually stands behind, rather than a tile chosen because the row needed six of them.

Illustration: A gate with a real rejection rate, rather than a featured row assembled from whatever was available.
A gate with a real rejection rate, rather than a featured row assembled from whatever was available.

The same instinct, elsewhere in the product

Copy link

The status page publishes no uptime figure, because nothing measures one. The country pages link to official wage statistics rather than quoting numbers. The Reasoning Challenge reports a percentage rather than an IQ. None of those is modesty: each is a place where the available options were to invent something or to say what is missing.

A page that says what it does not have is more useful than a page that manufactures something to have, and it is considerably easier to keep true.

The cost of shipping the appearance of a feature

Copy link

A half-built feature is not half a feature. It is a claim, and the claim is consumed whether or not the thing behind it works. A visitor who clicks a featured topic and finds a fixture concludes something about the whole product, and they are not wrong to.

Worse, nothing catches it. A page of placeholder content type-checks, lints, renders, passes every test and looks finished in a screenshot. The estate has a name for this shape of defect, because it has shipped it before: a module that is wired up to nothing at all passes every gate there is. The only defence is a rule stated in advance about what may be published, which is exactly what the empty page is.

What would change it

Copy link

A curated collection becomes publishable when there is a recorded basis for each entry: processing succeeded, attribution is complete, coverage passed, and somebody can say why this topic and not another. That is a higher bar than a featured row usually clears, and it is the bar that makes the page worth reading when it exists.

Until then the page stays empty and says so. If you want to know what the project is and who publishes it in the meantime, the about page is the honest version of a featured row.

There is a reasonable objection to all of this: if the page has nothing, why publish it at all? Because it is linked from the navigation and has been for a long time, and because removing it would answer the question by hiding it. A stated absence is a position somebody can disagree with. A missing page is not.

Consent version 2026-08-21.1

Cookie preferences