The electronic document generator


Novelang aims to support a wide range of languages, at least those with Roman characters. There are three aspects to consider:

  • The character set of the source document(s).
  • The character set of the rendered document.
  • The characters recognized by Novelang grammar.

Taking advantage of the underlying Java platform, Novelang supports numerous charsets. See the list.

Source document charset

Source document charset is set at Novelang startup, using --source-charset option (documented as an HTTP daemon feature).

Reading a source document in a charset which is not the one expected may result into incorrect character display, and even make the source document unreadable.

Charset mismatch can happen when working across different platforms with different default charsets. For Western European versions of Mac OS X, MacRoman is operating system’s default charset, while for Western European versions of Microsoft Windows it is Cp1250.

In order to prevent various headaches, you must be aware of the charset of your source documents, and use a text editor which is explicit about the charset in use.

Recommended source document charset is UTF-8, which supports a wide range of characters.

Rendered document charset

Rendered document charset is set at Novelang startup, using --rendering-charset option (documented as an HTTP daemon feature).

The character set of the rendered document makes only sense for text-based formats like HTML. For HTML, Novelang will do its best to provide named HTML entities (making HTML source more readable). But, unless you need some special transcoding operation, UTF-8 will always be great.

PDF don’t care about rendered document charset as it uses Unicode internally. But rendered document may look wrong with a font that don’t support its characters. Luckily, Novelang supports a preview of available fonts (with /~fonts.pdf pseudo-document).

Characters in the Novelang grammar

The characters recognized by Novelang are hardcoded in its guts. While Novelang reads and writes a lot of charsets, only some of the characters in this charset are supported. In order to keep the Novelang grammar meaningful, it’s not possible to admit any character, especially symbols and punctuation signs. Letters, like Roman ones with diacritics, will be added on a case-by-case basis.

By now, Novelang supports all letters of those languages: English, French, Hungarian.

Additional resources about Unicode