Overview

This document contains edits by the Integration Panel:

1. added "date" to metadata
2. added "language" to metadata
3. added a placeholder for the missing description element, but filled it with review notes
4. added comments to sequences to indicate sequence number from the proposal document
5. added a rule for each sequence used in the definition of other rules, so sequences can be named in rules that are using them
6. edited rules to use named rules for sequences
7. changed type="out-of-repertoire" to "out-of-repertoire-var" to follow conventional use
8. changed automatic comments on out of repertoire "char" elements to "Not part of reperotire" to match conventional use
9. changed automatic comments on variants to "Cross-script homoglyph"
10. added a named class for "v"
11. changed rule 14 to use cless for readability
12. added a named class for "k"
13. added a rule "ck" for "c+k" to make WLE rule 7 and 11 easier to follow
14. changed "[ ]" to "( )" in comments or removed.
15. applied XML source code formatting convention: space before end of element " />"
16. removed unused rule "followed-by-c"

Additional Review Notes

Issues found:

1. Several metadata fields are missing (listed in the next item) for some, placeholders have been provided
2. Language
3. Date
4. Description
5. References
6. References are not applied to "char", "var" elements, etc.
7. rule "between-v" does not exist as numbered rule in Proposal
8. Unicode version is out of date with current MSR (6.3.0) but this may change before the proposal is ready; at that time "unicode-version" element needs to be edited to match the latest MSR.
9. In [Proposal-Myanmar] the key CM42 is used twice in Table 8-B (this is correct in Appendix B).
10. In [Proposal-Myanmar] the key SV5 is used twice in Table 8-B (this is correct in Table 8-A).
11. variant type="out-of-repertoire" should have been "out-of-repertpoire-var" to follow conventional use
12. Commenting does not follow conventional usage, particularly for variants, many comments empty
13. For left side context rules that contain a choice of single character classes as well as some sequences: if the sequences all end with code points in one of the character classes, adding the sequence does not alter the set of labels matching the context rule; the sequence is redundant and can be removed from the rule. Only sequences that end in different character class need to be added. (Examples are Rules 8, 9 and . However, Rule s 12 is fine, it contains only sequences; for Rule 10, the sequences end in -k, which is different from the "shan_vowel" class).
14. Many sequence definitions in the repertoire are redundant. A sequence can override a generic context rule, but if the generic context rules already allow the elements to combine into the sequence, it does not need to be specified. An exception is the case where a sequence is a target of a variant mapping. In the Myanmar LGR, all sequences of type c-m can be removed, for example (but sequences of type c-m-m or dv-dv would need to be retained).
15. the classification "myanmar" seems to be unused
16. the classification "shan" seems to be unused
17. comments mention the tags "c_pao" and "m_pao" but these are not defined.
18. literal sequences in rules should be expressed as named rules and referenced by name for clarity
20. literatl code points for which tags exists (k and v) should be replaced by named character classes of the same name in the rules for clarity

References

[Proposal-Myanmar]

Myanmar Generation Panel, Proposal for a Myanmar Script Root Zone Label Generation Rule-Set (LGR), [DATE and URL TBD]