suppress-script values for fil, mi, pes, prs, qu members
petercon at microsoft.com
Thu Oct 21 16:24:35 CEST 2010
From: Michael Everson [mailto:everson at evertype.com]
> You just said that MS localizes into quz. Not the other forty.
So, you think it would make sense to add s-s to quz but not the rest? (I'll keep that in mind the next time you want to encode characters to complete a paradigm.)
>And unless there is some danger that quz will be tagged with Cyrl or Hans, what is the "actual need"?
> So you want to add subtags here to plug a leak based on the
> way you have "really" implemented. Surely your implementation
> can compare strings in a table without having to parse them.
What MS has implemented used the tags quz-BO, quz-EC and quz-PE, which are the tags that follow the recommendation in 5646 and that you yourself gave: don't include script subtags (quz-Latn-BO, etc) because that would be redundant and unnecessary. But the registry does not capture the fact that that would be redundant and unnecessary because the record for quz doesn't include the s-s field. Now when implementing software that would compare tags, whether the tags are parsed or not, we need additional data indicating that quz and quz-Latn are to be treated as equivalent, and likewise for quz-BO and quz-Latn-BO, etc.
I don't want to add subtags: I'm not asking to add subtags in the registry, and I don't want users to include the Latn subtag in tags based on quz. The most reliable way to effect the latter is to add a s-s field to the record for quz so that the data in the registry reflects what we here know: that a Latn subtag would be redundant and unnecessary when used with quz. (And likewise for the other cases I've mentioned.)
More information about the Ietf-languages