Dear CJK GP members, FYI, actually, my implementation of proposed algorithm is using the same method with following. I wanted to explain proposed algorithm as generic as possible, so the explanation and implementation was slightly different. I'm very happy if someone can reword it as generic as possible. Regards, -- Yoshiro YONEYA <yoshiro.yoneya@jprs.co.jp> On Fri, 15 May 2015 23:21:00 +0900 Yoshiro YONEYA <yoshiro.yoneya@jprs.co.jp> wrote:
Dear CJK GP members,
As discussed today's meeting, I did reword of step 3 and 4 as follows.
Step 3
From LGR-M, each GP picks up the variants corresponding to every CP in its LGR-1 – Variant type of variant(s) inherits its original variant type (or subtype if any) value in LGR-1 – Variant type of variant(s) not defined in LGR-1 is defined as: . "blocked" if the variant is not in LGR-1 repertoire (codepoint of the variant is marked as out-of-repertoire for next step) . Otherwise, inherit its original variant type (or subtype if any) value in LGR-1 repertoire.
Step 4
Each GP picks up codepoints marked as out-of-repertoire in previous step and extracts corresponding variants from LGR-M to establish symmetry. - Variant type of reflexive codepoint marked as out-of-repertoire is "out-of-repertoire-var" - Variant type of another corresponding variants are "blocked"
Regards,
-- Yoshiro YONEYA <yoshiro.yoneya@jprs.co.jp>
_______________________________________________ japanesegp mailing list japanesegp@icann.org https://mm.icann.org/mailman/listinfo/japanesegp