generated from ivoa-std/doc-template
-
Notifications
You must be signed in to change notification settings - Fork 15
Open
Description
UCS-2 has been a BMP-only serialisation of unicode codepoints that's been deprecated since forever. We really can't have that in our specs any more. In principle, UTF-16 is a largely-compatible successor. We should use it instead, as suggested by https://wiki.ivoa.net/internal/IVOA/InterOpJune2025Apps/unicode-notes.pdf.
However, for us it's not that simple because for BINARY2's sake we need to know exactly how many bytes a unicodeChar[4] (say) has. If we say the array size is the number of unicode codepoints (which, I give you, is what everyone will expect), then that could be 8, 10, 12, 14, or 16.
I'd like to direct further discussion to PR #68.
Metadata
Metadata
Assignees
Labels
No labels