L2/10-044 Title: Suggestions for Restructuring UTS #10 for Unicode 6.0 Author: Ken Whistler Date: January 27, 2010 Action: For consideration by the UTC I have an old action item, 108-A029, to provide recommendations on restructuring UTS #10 for clarity. This document is provided to fulfil that action item. This is another of our documents which could use a substantial editorial pass on detailed content, but for now I am again suggesting just section level rearrangement and coalescing, to help the overall structure of the document. What I'm asking for is general consensus that this outline of section-level restructuring would be a valuable rework of the existing document, so that the authors can go ahead and prepare a new proposed update for Unicode 6.0. The UTC would then get another chance at the subsequent meeting to verify the results before committing to approval for Unicode 6.0. Here are the suggested changes, at a high level. 1. Remove the 3.1 Linguistic Features subsection head. The subsubsections under this all arguably have something to do with linguistic issues for collation, but the subsection doesn't hang together as is and is out of place in Section 3, Collation Element Table. 2. Hoist the levels of 3.1.1 Multiple Mappings, 3.1.2, French Accents, and 3.1.3 Rearrangement. 3. Move 3.1.4 Default Values into Section 3.2 Default Unicode Collation Element Table. 4. Move 3.1.5 Collation Graphemes and 3.1.6 Combining Grapheme Joiner into Section 6, Implementation Notes, and possibly retitle those subsubsections. 5. Move 7.3.1 Tertiary Weight Table into Section 3.2 Default Unicode Collation Element Table. 6. Move 6.9 Tailoring Example: Java into Section 5, Tailoring. 7. Coalesce 7.2 Canonical Decompositions and 7.3 Compatibility Decompositions as subsubsections under Section 6.5 Avoiding Normalization. 8. Hoist the levels of the remaining subsections of 7 Weight Derivation. 9. Split the existing 7.1.4 Trailing Weights section. The part that stays in Section 7 Weight Derivation should be a short, generic discussion of the use of high values to derive trailing weights. The rest should be split off into a new section: Hangul Collation Issues. (This anticipates possible refactoring of the Hangul content, based on what else happens in 6.0 and after regarding Hangul.)