docs(site): sequence the landing page and repair its panels - #210
Merged
Conversation
Reorder the argument, then fix what the panels were doing wrong. The comparison section carries the sentence "This is the whole argument. Every other claim on this page follows from it", and it sat fifth. The benchmark sat fourth, where a reader met the evidence before the claim it answers. The page now runs promise, filter, product, premise, mechanism, proof, limits, act: what it replaces and how it works move ahead of the measurement, so the number lands on a reader who has just been told one process does the work of five. Section styled every descendant paragraph rather than its own slotted prose. A class-plus-element selector outranks the single class a child component styles itself with, so component paragraphs inherited section margins and type — which is why the benchmark card had started defending its own copy with an id and doubled class selectors. Scoped to direct children. Starlight adds a top margin to consecutive blocks in rendered content. Inside components that lay out their own grids that margin lands between grid children, so the preview's health tiles and the contrast panes' rows did not share a baseline. Those three components take Starlight's own opt-out; the others keep content styling because they contain links. Every column of the service table was left-aligned, so the table's slack collected in the last cell and the figures huddled against the service names with half a pane empty beside them. Numerics align right and compare down their columns, and the pane closes with a count instead of stopping halfway down. The three status hues were hardcoded in three components. They are site tokens now, with light-theme and print values: the pale green and coral that read clearly on Ayu surfaces fall under 4.5:1 on white, which is where the hero's spike annotations had been sitting. Claude-Session: https://claude.ai/code/session_01FCgBTMHyJzwwbMDJ2RfzKk
The closing call to action shows the one command that installs Fanout, and selecting it by hand across a line that scrolls sideways on a phone is the worst part of the page. Analytics counts installs from the browser's own copy event, which fires for a hand-made selection and not for the Clipboard API. A button that only wrote to the clipboard would take over the common path and stop the measurement without anything looking broken, so the button announces the copy on a custom event and the analytics script counts both paths. The button removes itself where the Clipboard API is absent, because a control that silently fails is worse than selecting the text by hand. The command box gains the tabindex its own comment already claimed it had: it scrolls, and an overflow a keyboard cannot reach cannot be read. Claude-Session: https://claude.ai/code/session_01FCgBTMHyJzwwbMDJ2RfzKk
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Order
The comparison section carries the sentence "This is the whole argument. Every other claim on this page follows from it" — and it sat fifth, with the benchmark fourth, where a reader met the evidence before the claim it answers.
New sequence: promise → filter → product → premise → mechanism → proof → limits → act.
Each section now answers the question the previous one raises, and the benchmark lands on a reader who has just been told one process does the work of five.
Panels
Reviewing each visualization turned up four defects, three of them systemic:
Section.astrostyled every descendant<p>, not just its own slotted prose. A class-plus-element selector outranks the single class a child component styles itself with, so component paragraphs inherited section margins and type — which is whyBenchmarkhad started defending its copy with an id and doubled class selectors. Now scoped to direct children.not-contentopt-out; the rest keep content styling because they contain links.Verified at 1280px and 430px in both themes; no horizontal overflow.
https://claude.ai/code/session_01FCgBTMHyJzwwbMDJ2RfzKk