2022-07-12
[15:32:23.0216] <sffc>
FYI Richard Gibson ryzokuken Romulo Cintra and others: Mark Davis made a PR to clarify the UAX 29 conformance rules in response to our conversation last week. See here: https://github.com/unicode-org/unicode-reports/pull/40

[15:33:28.0730] <Richard Gibson>
I get a 404 Not Found

[15:33:43.0437] <Richard Gibson>
must be a private repository

[15:34:24.0526] <sffc>
oops, it must be a private repo. The gist of the change is to adopt conformance text more similar to that found in UAX 31: https://www.unicode.org/reports/tr31/#R1

[15:35:42.0944] <sffc>
It means that implementations that say that they conform to UAX 29 must either use the default UAX 29 rules _or_ declare that it is tailoring those rules

[15:37:34.0946] <sffc>
On our end, this doesn't mean anything yet, since ECMA-402's conformance language to UAX 29 is still fairly weak on its own, but at least it means that we could strengthen it if desired.

[15:38:06.0321] <sffc>
I chatted with Mark this morning about it, and he suggested that CLDR could also release off-the-shelf "segmentation profiles" tailored to certain use cases

[15:39:37.0023] <sffc>
In the longer term, we could then expose a menu of profiles to Intl.Segmenter users, with one profile that we choose as the ECMA-402 default.