L2/04-153

UAXes vs. Unicode book

(Source: Markus Scherer, May 2004)

The Unicode Standard is published in book form but with standard annexes (UAXes) which are published online, with snapshots on the book CD.

For an implementer, single-topic documents like the UAXes are much easier to deal with. Their scope closely relates to an implementation work item. They are easily referred to via online links, especially to particular versions of the documents. Changes are tracked more easily compared to book contents.

Sometimes, contents is moved from online documents to the book, which is unfortunate for technical topics.

For example, for Unicode 4, the contents of UAX #21: Case Mappings was moved from the online document to the book. Simultaneous changes to the contents were not pointed out, and thus were obscured, which most likely led to differences in implementations. It was easier to refer to UAX #21 and its versions for the definition of case mappings than to refer to parts of the Unicode 4 book.

Suggestion for the next major version of the Unicode Standard:

• As much implementation-relevant information as possible in diffable data files.
• Move Case Mappings back into UAX #21.
• Separate out more sections of the book into separate UAXes.
• The book is a good format for overview information, but not optimal for algorithms.