About this archive

L.O.S.T — the Library Of Secret Text — is a static website written in plain HTML5. No frameworks, no build step, no server code. Open it straight from disk or host it anywhere that serves files.

The palette

Two colours carry the site: neon purple and neon green . Purple is the accent — headings, the wordmark, download buttons — and green is the signal colour used for Open here and the active viewer. Amber (search) and red (close) appear only where a third meaning is needed. Every colour is declared twice, once per theme.

Where the documents live

8utp5678364bhkdsytr/ is the site root. The three pages and the assets/ folder sit in it, and the documents live in two folder trees inside it — PDFs under pdf/, plain-text transcripts under text/ — each split into one sub-folder per subject (blk, div, enl, ilu, mas, occ, psy, rel, spi, spm, tar, wht, wit). The shelf is stocked in pdf/occ/, pdf/mas/, pdf/wit/, pdf/psy/, pdf/spm/, pdf/enl/, pdf/ilu/, pdf/spi/, pdf/tar/, pdf/div/, pdf/wht/ and pdf/blk/; only pdf/rel/ is empty and waiting for whatever you file next.

Links from the library page are relative to that root, so the pattern is simply:

viewer iframe  →  src="pdf/rel/LOS-REL-001-gospel-of-thomas-translation.pdf"
               →  src="text/rel/LOS-REL-006-shepherd-hermas-fragment.txt"

Open here loads the document into the viewer at the top of the reading room, so you keep your place on the shelf — the viewer has Download and Close buttons of its own.

The shelf scans the folders

A static page cannot read a folder by itself, but it can ask the server for that folder's listing. On load, library.html requests pdf/ and text/, then every sub-folder it finds inside them, reads the links the server sends back, and builds a card for every file it finds. Nothing is hard-coded: copy a file into either tree — loose, or into a subject sub-folder — and it is on the shelf at the next reload (or the next press of Search, which re-reads both trees before it filters). The same scan feeds the counts on the left-hand buttons, the search box and the figures on the home page.

The folder listing has to come from the server, so run the site from a server rather than opening the file directly. jwebserver -p 8000 (Java 18+, run inside 8utp5678364bhkdsytr/), Apache, nginx, Node's http-server and PHP's built-in server all list folders by default. Two rules keep it working: do not put an index.html inside pdf/, text/ or any subject sub-folder (the server would serve that page instead of a listing), and keep the file name free of spaces and question marks.

If a host refuses to list folders — or the page is opened straight off the disk over file:// — the page reads library-index.js instead: the shelf, the subjects and the counts come from that file, and the note above the shelf says so in amber. It is a .js file so a <script> tag can load it — the only way to read a file from file://, where fetch() is blocked.

That file is written by the one tool in the project, tools/ScanLibrary.java, which walks the same two trees offline and records every document it finds with its subject, code and size: java tools/ScanLibrary.java. It is only needed for hosts that will not list folders — when the live scan works, there is nothing to run.

Adding your own documents

The categories

Light and dark

The round button in the top right swaps the whole palette. Your choice is remembered in the browser's local storage, and a tiny script in each page's <head> applies it before the first paint, so the page never flashes the wrong colours. Every colour used for text is defined twice — once for the dark theme and once for the light one — so headings and labels stay readable on both.

Images and buttons

All artwork is hand-written SVG, so it stays sharp at any size and adds no dependencies:

The viewer

Pressing Open here on a card does not navigate away. It points an <iframe> at the file and shows the viewer panel above the shelf, so a PDF renders in the browser's own reader and a text transcript renders as plain text. The viewer carries its own Download and Close buttons, and it writes the file path into the page address as library.html#pdf/rel/your-file.pdf — copy that link and it reopens with the document already loaded. Press Esc to close it.

The buttons down the left are navigation, not just filters. Choosing a subject or a format while a document is open closes that document and scrolls to the shelf you picked, so the click always lands you somewhere new instead of leaving the open file in the way. Search is the exception: it narrows the cards underneath and leaves the open document alone, which is what you want when you are reading one thing and hunting for another.

The three format buttons count the subject that is on screen, not the whole library. Choose Freemasonry and they read Everything 6, PDF 3, Text 3 — the two figures always add up to the subject's total, so you can see the shape of a subject at a glance before opening anything. All subjects puts them back to the library-wide totals.

Notes on sourcing

Enter the Library → Back home