Make a reader, step 1: show an article

A Tersenet article is plain UTF-8 text that already reads well as it is. A reader's job is to show it a little better: headings larger, paragraphs wrapped to the window, links you can follow. You can write one in any language, with any way of putting text on a screen, and you do not need the Terse32 virtual machine at all: articles and the network are the whole of Tersenet's reading side.

This first step needs no network either. You will show an article that is already a file on your computer.

An article to start with

Save this as hello.txt:

# Hello

My first article.
Two lines, one paragraph.

## A short list

- one item
- another, with `code` in it

The rules you need first

An article is a series of blocks. Blank lines separate them, and each block is known by how its first line starts:

Inside a line, text between two backquotes is a code span, and a link is a terse: address, bare or written as [text](address). An article's address is terse: and 64 hexadecimal digits. Only terse: addresses are links: a web address is ordinary text.

That is enough for a first version. Anything that matches no rule is ordinary text, shown as written. A reader that does not lay tables out yet may show a table's lines as a paragraph, and one that does not draw figures may show a figure as its text: the format allows both, so leave them for later.

Check yourself

The Tersenet sources come with a test for exactly this. The folder vectors/articles holds 58 small articles, each .txt beside a .dump file that lists what a reader must find in it, one block or piece per line. For hello.txt it is:

heading
  text "Hello"
paragraph
  text "My first article. Two lines, one paragraph."
subheading
  text "A short list"
item
  text "one item"
item
  text "another, with "
  code "code"
  text " in it"

Make your program print the same, and compare. The reader tersereader (in the sources as native/reader-go) prints it for any file:

tersereader --dump hello.txt --tokens
tersereader --dump hello.txt

The second command shows the article laid out as text, with its links numbered: a good picture of what yours should do.

tersereader is itself a reader written from the specifications alone. With the network left to a separate program, as step 2 shows, it is about 5,300 lines of Go. A reader is a project you can finish.

What you have now

A program that shows Tersenet articles from files. Next, get them from the network: Make a reader, step 2: fetch an article.

The complete rules, with every edge case, are in ARTICLE.md; Writing an article shows each construct as source and as shown.