Peter Kirk wrote:
This is not a practical use of variation sequences if, by this, you mean use of variation selectors. What are you going to do, add a variation selector after every single base character in the text? ...
...
... Are you expecting fonts to support the tiny stylistic variations between Phoenician, Moabite, Palaeo-Hebrew, etc. -- variations that are not even cleanly defined by language usage -- with such sequences?
No one has suggested this.
Then what is Ernest suggesting? He wrote that the distinction between stylistic variants of unified scripts could be done with variation sequences, i.e. a sequence that 'always consists of a base character followed by the variation selector, may be specified as part of the Unicode Standard'. He then went further and wrote:
My point was that I have seen enough evidence to
absolutely convince me that if both glyph repertoires
are unified in a single script, variation sequences
would be *necessary*. [My emphasis.]So what is he suggesting if not that every single base character in a text would be followed by a variation selector character in order to make a plain-text distinction between stylistic variations?
Why not change the friggin' font? Why not use something other than plain-text?
The solution may be a catch-all, but the problem is a real one. Dr Kaufman's response makes it clear that to professionals in the field Everson's proposal is not just questionable but ridiculous. There is certainly some PR work to be done in this area, not name-calling.
Peter, are we talking about the same thing? Ernest is suggesting bizarre measures to deal with a problem -- in my opinion, a non-existent one -- that he sees in *unification*. You are arguing against Michael's *dis-unification*. The ridiculousness of Ernest's suggestion to use variation selector sequences -- indeed, perhaps he intends it to be ridiculous to make a point -- is an argument in favour of dis-unification, since the alternative for making a plain-text distinction is so daft.
My question, again, is whether there is a need for the plain text distinction in the first place?
John Hudson
