Supporting labels such as windows-65001 for utf-8 seems bad. Any chance you could align closer to Gecko and http://encoding.spec.whatwg.org/ ? Supporting more labels is not necessarily in your benefit. E.g. sites might rely on them not being supported.
Yes, I want to switch from enumerating all ICU aliases to having a built-in table in WebKit. That will have the side effect of removing support for some crazy encoding names.
That said, the ideal resolution would be to have an ICU "standard" for HTML5, so that the right names would be in ICU. This is exactly what those are for, and I find it less than great that HTML encoding names are tracked in the Encoding spec, not in ICU. See <http://demo.icu-project.org/icu-bin/convexp> - there are separate namespaces for MIME, IANA, and others. HTML is just another case.
ap, I don't understand that feedback. Why would it not be good to have a standard for what ICU should implement? ICU is not the only implementation.
True. Maybe that would be another project under Unicode consortium umbrella (CLDR perhaps)? The stellar history of the consortium's interaction with many interested parties makes it a desirable gatekeeper for a sensitive spec like this.
I don't really see what moving it brings us. It's mostly a research project.
See also: bug 28760.
This was done a few years ago.