Word division

Words are separated by spaces. The compiler wraps each word in <w> (when the template enables word wrapping) and punctuation in <pc>, and it handles words broken across a line or page. Three markers override the normal division.

| — a word boundary where there is no space

a|b

produces two words with no space between them. Use it where the manuscript has no space but the edition treats the text as two words, or where a parser would otherwise see one word.

~ — one word written as two

upp~haf

is one <w> element containing both parts. The difference appears on the MENOTA levels:

Level	Output
Facsimile	upp haf
Diplomatic	upp haf
Normalized	upphaf

On a single-level document the two parts are joined with a space (as the manuscript has it). On a multi-level document the normalized level joins them, because normalization treats them as one word.

~ before a break — a word continued across lines or pages

hár~//2fagra

is a word that continues across a line break:

<w>hár<lb n="2"/>fagra</w>

hel~///1vlo

continues across a page break. The break is included inside the <w>, so the word is not split between two word elements and its lemma applies to the whole word.

Note

A ~ directly before // or /// makes the continuation explicit. Without it, a line or page break that immediately follows a letter is also treated as continuing the current word, following the common manuscript convention; a break after punctuation ends the word. Writing the ~ is clearer and does not depend on that heuristic.

Punctuation and entities

Punctuation written as an entity is a separate <pc> element when it stands alone between spaces:

þa :middot: þo

Attached to letters, the same entity is part of the word. :middot:Q:middot: is the numeral ·Q·, one word.

A single-letter entity such as :inodot: (the word í) is a word of its own.

Word numbers

Lemmas and annotations are stored by word number: the n-th <w> of the body, as the previews count it. Two things follow:

• Words inside .head{}, .supplied{} and .norm{} blocks are counted like any other word.
• When text is inserted or deleted, the application keeps track of which word is which, so lemmas and annotations stay attached to their words rather than to positions. See Lemmatization.