Open NorbertLindenberg opened 5 years ago
Isn't Javanese going through USE now?
According to the Introduction, Javanese goes to the USE in Windows 10, which implies that in Windows 8.1 (which was the first release to support Javanese) it's still handled by its own shaping engine.
Anyway, I wouldn't mind if this page disappeared; I just want to let font developers know that its value as a reference document is questionable.
I just want to let font developers know that its value as a reference document is questionable.
Right... That's, I guess, implied. :))
Since Windows 8.1 was EOLed over a year ago, maybe the page attempting to document its Javanese shaping engine can now be garbage-collected?
It appears that this specification has been abandoned by Microsoft, and bugs in it will not be corrected.
I reported the following issues to Microsoft four years ago, on November 24, 2014:
1) The table defining the character classes has some lines that probably shouldn’t be there, from “X+” with the Myanmar code points down to the redundant “X*”.
2) The table defines a group “J” for joiners, but then that class isn’t used in the regular expressions defining clusters. Where do they fit in? The section Recommended Glyphs mentions they may be used as generic bases, but they’re not included in the definition of class “GB”.
3) The section Recommended Glyphs lists several other characters that are not included in the definition of class GB before mentioning that they may be used as generic bases.
4) The table defines classes “MR” and “MY”, but then later refers to classes “MCR” and MCY”. I assume these are supposed to be the same.
5) The class “O” is defined as “SCRIPT_COMMON” – is that meant to be the value common for the Unicode property script? If so, a number of SCRIPT_COMMON characters are listed as belonging to class “GB” (00A0, 00D7, 2012–2015, 2022, 25CC, 25FB–25FE) or “WJ” (2060), and at least “GB” glyphs are treated differently from “O” characters.
6) The table defines a class “VPst”, but that class is not used in any of the regular expressions defining clusters. A comment under the regular expression for complex clusters implies that it should be used in it.
7) The class “WS” is defined as “white space” – is that meant to be the Unicode bidirectional class? If so, the white space character 000C also has the property script common. Or is meant to be the Unicode character category Zs? In that case there’d be a conflict for 00A0, which is also listed in class “GB”. Also, if ZWSP is meant to be 200B, that’s not a white space character under either definition.
Document Details
⚠ Do not edit this section. It is required for docs.microsoft.com ➟ GitHub issue linking.
ID: b136ee1d-ce2e-f8cb-47ef-fe1181b73a50 Version Independent ID: b48c1d37-72b0-7893-d248-2514bf4a7661 Content: Creating and Supporting OpenType Fonts for the Javanese Script - Typography Content Source: typographydocs/script-development/javanese.md Product: typography