Zim - A Desktop Wiki
Formats
Formats parse plain text with markup into a parse-tree
All Formats should support the IO object interface. Do not assume that the IO object always maps to a file, this is responsibility of the Repository object. DBI interfaces could be used by the Repository object, in which case IO::Scalar could be used to feed the Format.
TODO: document parse tree
TODO: Document how to write parsers
Comment on format restrictions and how to deal with them, e.g. number of header levels, ital and bold versus emphasis and strong etc.
Main difficulty is how to do linking. Explain that when we do not get an URL we will apply zim's internal lookup. URL syntax to overload zims lookup, e.g. page://foo (zim:// already used for interwiki) - or use property ?
----
Differentiate between import/export and read/write. See e.g. the script for converting Tomboy notes. For import/export we potentially need control over the storage, while read/write only converts one file at a time and let the default storage worry about where the files go.
----
- dokuwiki (subclass wiki ?)
- extra options for images
- nested tags
- bullet lists have whitespace indent
- locking mechanism ?
- honor change log
- check media dir
- add title read-only on top
- renders a link like ":users:pardus" as "pardus"
- need "prefix" or "base" to keep link direct editable
- numbered lists with "-" prefix
- kwiki
- reversed order for headings
- single chars for markup / and *
- uses [=text] for inline verbatim
- uses -text- for strike
- supports html entities with &entity;
- camelcase for links and [text url text] for forced links
- "!" escapes camelcase or url
- lists use multiple * or 0 characters
This website was written in Zim !