On 9/22/2018 6:01 PM, Jonathan M Davis wrote:
For better or worse, English is the international language of science and
engineering, and that includes programming.
In the earlier days of D, I put on the web pages a google widget what would
automatically translate the page into any language google supported. This was
eventually removed (not by me) because nobody wanted it.
Nobody (besides me) even noticed it was removed. And the D community is a very
international one.
Supporting Unicode in identifiers gives users a false sense that it's a good
idea to use them. Lots of programming tools don't work well with Unicode. Even
Windows doesn't by default - you've got to run "chcp 65001" each time you open a
console window. Filesystems don't work reliably with Unicode. Heck, the reason
module names should be lower case in D is because mixed case doesn't work
reliably across filesystems.
D supports Unicode in identifiers because C and C++ do, and we want to be able
to interoperate with them. Extending Unicode identifier support off into other
directions, especially ones that break such interoperability, is just doing a
disservice to users.
