# UI ## Philosophy A developer tool people use daily to browse code and manage repos. Optimise for: 1. **Speed-first** — minimise clicks, maximise information density 2. **Scannable** — find what you need in under a second 3. **Quiet confidence** — premium without flashy; the UI should disappear 4. **Code-centric** — code is the hero, everything else supports it | Principle | Meaning | |---|---| | Readable density | Compact without cramped. Maximise info per viewport. | | Clear hierarchy | Strong contrast between labels, content, and muted elements | | Functional spacing | 8px on items, 12–16px on sections. No wasted space. | | Keyboard-first | Every action reachable without a mouse | | Dark-mode primary | Developers live in dark mode. Light mode is supported, secondary. | **Do:** small type (`text-xs` / `text-sm` for most UI), small buttons, monospace for paths, SHAs, and branch names, opacity modifiers for text hierarchy, tight spacing. **Don't:** large type outside page titles, shadows in dark mode (use borders), bright backgrounds, hover animations that move or resize things. ## Where the portfolio framing bites Steid is portfolio-first, not a Gitea clone. The profile page is the product — repos are one kind of thing on it, alongside writing and projects. Any layout inherited from a GitHub-shaped forge needs checking against that before it's copied. ## The profile page Settled 2026-08-29, after looking at the built page with real data rather than imagining it. Three sections of equal weight — Repositories, Writing, Projects — two of them empty boxes, repositories sorted alphabetically so `dotfiles` outranked `steid`. A third of the page was content and two thirds advertised incompleteness. **Flat, app-style navigation. No cards, no boxes.** ``` James Gill @jamesgill ───────────────────────────────────────────── Overview Repositories 3 Writing ───────────────────────────────────────────── CURRENTLY BUILDING steid Rust A self-hostable gitforge where the profile is the product. 421 tests · updated 2 hours ago REPOSITORIES All 3 → ────────────────────────────────────────── topcoat-notes 3d Notes on building with Topcoat 0.5. ────────────────────────────────────────── dotfiles 2w ``` - **A tab bar, not stacked sections.** Empty sections stop being visible failures and become destinations that simply have nothing in them yet. It is also what makes Writing shippable without redesigning the page around it. - **Tabs are links.** No JavaScript, and each tab is a real URL that can be shared. - **One lead item with genuine typographic weight**, then a compact list. This is the editorial layer portfolio-first requires: without it the page has no opinion about what matters, and ordering alone cannot carry that. - **Hairlines and hover states, never bordered containers.** The row is the unit. - **Recency, not alphabetical.** Alphabetical is a filing rule; a portfolio needs an editorial one. - **Sections with nothing in them are hidden from visitors** and shown to the owner as an affordance. ### One vibe. Hierarchy without scale. The first version of this design gave the lead item large type and justified it by splitting the product into a "portfolio surface" and a "tool surface" with different rules. **That was wrong and is not what ships.** Two design languages in one product is how a product stops feeling like one thing, and it contradicted this file's own philosophy — quiet confidence, the UI disappears, small type throughout. Nothing exceeds 14px except the page title. The lead item is distinguished by **everything except size**: | Device | Lead | List row | |---|---|---| | Eyebrow label | yes | no | | Weight | 600 | 500 | | Description | full width, roomy | compact | | Space around it | generous | tight | | Separator | hairline below | none | The cost, stated plainly: the distinction is **subtle**. One weight step is not much, and the eyebrow and whitespace carry most of it. That is the right trade for a developer tool — but it means the copy in the eyebrow is load-bearing, because it is the clearest signal that this item is different. **The eyebrow reads "Currently building"**, chosen over a neutral "Featured". It says something rather than labelling something, which is what a portfolio built in public wants. The known cost: it is a claim about the present, so pinning a finished project makes the page lie. The answer if that becomes a problem is a per-pin label, not a blander default. The general rule this sets: **hierarchy comes from weight, colour and space. Reach for size last, and outside a page title, probably not at all.** ### The front door Steid is meant to be someone's site, reached at their own domain — `jpgill.dev`, not `git.jpgill.dev`, because a `git.` subdomain announces a Gitea clone and this is portfolio-first. That makes the root a signpost rather than a page: unclaimed it sends you to setup, and otherwise it forwards to a profile — your own when signed in, **the owner's when not**. The last case is the one that matters. A stranger arriving at the apex came for the profile and will never sign in, so a sign-in prompt there is the front door answering the wrong question. It only applies when the owner is unambiguous: with more than one user "the owner" has no answer, so the root falls back to a generic landing rather than electing someone by storage order. ### What it needs that does not exist yet - **`updated_at` on a repository.** The design promises recency and nothing stores it. Asking git costs one fork per repository — a twelve-repo profile would be ~150ms of `execve` before rendering. So it is a column, touched when a push is authorized. Slightly wrong if a push then fails; the alternative is wrong more expensively. - **A `/{handle}/repos` index.** The Repositories tab needs a destination; only `/{handle}/repos/{name}` exists. - **A way to choose the lead.** A `pinned` flag on the repository, set from the settings page that now exists. No pin, no lead section — automatic "most recent" would put a dotfiles tweak at the top of a portfolio. ### Mocking is harder than it looks here **Tailwind classes the app does not already use are not in the built stylesheet.** `build.rs` scans the real sources, so a mockup written against the served CSS silently loses any new utility — `gap-7` and `tracking-widest` collapsed a nav into `OverviewRepositories3WritingProjects` before this was understood. Iterate either in the app itself or, for throwaway exploration, in plain CSS against the theme's custom properties (`--background`, `--foreground`, `--muted-foreground`, `--border`, `--surface`). Dark mode is a `.dark` class on an ancestor, not `prefers-color-scheme`; a mockup without it renders light. ## Stack Settled in [0005](decisions/0005-tailwind-and-copied-components.md). | Concern | Choice | |---|---| | Styling | Tailwind via Topcoat's build script — no Node | | Theme | `styles.css` at the package root: design tokens, dark-first | | Components | `topcoat ui add` for primitives, hand-written for Steid's own | | Mono | IBM Plex Mono, for paths, hashes, branches, handles | ### Tokens Defined in `styles.css` on `:root` (light) and `.dark` (dark, the default): ``` background surface foreground muted-foreground primary primary-foreground success success-foreground confirmations, git additions warning warning-foreground caution, git modifications destructive destructive-foreground errors, git deletions border ring shadow-xs shadow-sm ``` `surface`, `success` and `warning` are Steid's additions to Topcoat's `neutral` theme; the rest were retuned to the palette below. **Components reference tokens, never raw colours.** A hardcoded colour will not follow a palette change and will not adapt to the colour scheme. This is the convention to enforce in review. ### Working on it ```bash topcoat dev # rebuilds CSS and re-bundles on change topcoat asset bundle # manual builds only; a stale bundle serves stale CSS topcoat ui list # what the registry offers topcoat ui add # copies source into src/components/ ``` `components.toml` is the record of what is installed; don't duplicate that list here. **A registry component can need more than a copy.** `select` draws its chevron with an Iconify icon, which needs the `icon-iconify` feature on `topcoat` *and* the icon set staged in `build.rs`: ```rust topcoat::icon::iconify::BuildConfig::new().icon_set("feather").stage().unwrap(); ``` The build fails with a message naming the missing set, so it is discoverable — but the build.rs edit is easy not to expect from a command that only advertises copying a file. Icons are embedded at build time; nothing is fetched at runtime. `topcoat ui add` also rewrites the module list in `src/components.rs` by appending, so re-alphabetise it afterwards. ## Prior art Attempt #1 (`steid-backup/AGENTS/UI.md`) has a complete 643-line design system — OKLCH light/dark palettes with concrete token values, a type scale, spacing scale, and component markup for sidebar, file tree, commit bar, breadcrumbs, badges, empty states. It was written for Tailwind + DaisyUI. Its palette has been mined already — the dark-first OKLCH values at hue 260 are what `styles.css` was retuned to. Its component markup has not, and comes with two caveats: it specifies DaisyUI, which is a separate choice from Tailwind and not bundled by Topcoat; and it was written for a GitHub-shaped forge rather than a portfolio-first one. It is also, on its own, longer than every other doc in this directory combined — for an app that had about nine pages. Take the palette and the principles. Don't re-specify components before there are components.