L2/10-307

Date/Time:    Sun Aug  8 12:23:46 CDT 2010
Contact:      timpart@perdix.demon.co.uk
Name:         Tim Partridge
Report Type:  Feedback on an Encoding Proposal
Opt Subject:  Feedback on Karl Pentzlin's Draft Proposal to add Variation Sequences for Latin and Cyrillic letters

(Previously sent to Unicode mailing list)
There are 256 selectors but the proposal only suggests numbering up to 16 effectively deprecating the others. Surely we want all 256?

The Mongolian selectors alter the appearance of the glyph displayed after the character has been evaluated for position in the word and a series of complex rules applied. The user will normally only have to use the selectors in exceptional cases. The selectors are only valid in certain positional cases and have been somewhat arbitarily assigned. It is not the case that selector 1 selects the same alternative form in all positions.

A typical user will see most of the variations in use from the built in rules being applied. There is not a user entity which would be considered variant 1 which is used by a separate community. I regard to proposal to give a name like VARIant-M1 as confusing as they have no basis in reality

I am also have some concerns from a security point of view as the proposal makes variation selectors valid for Latin characters for the first time. The selectors which produce a default behaviour or make one character look like another already encoded seem unneeded and introduce yet more clones of common characters. 

I also have concerns about the proposal to give the non ideographic variants names like VARIANT-1. Surely it is possible to give them descriptive names which would make it easier to understand what is meant? It is not as if we will have thousands of these.

Some parts of the proposal have merit, but I would urge the UTC to hold a public consultation on the matter to allow more time for feedback to be gathered.