Q2: What mapping function should be used in a revised IDNA2008specification?
alexander.mayrhofer at nic.at
Tue Apr 7 15:53:29 CEST 2009
> I've seen 3 variants of the position on mapping "A" to "a".....
> 1 - casefolding (UTR#21 section 2.3 paragraph following S.3)
> is required
> 2 - lowercasing (UTR#21 section 2.3 bullet S.2) is the one that's
> required, but casefolding is OK if others prefer it
> 3 - lowercasing is required, casefolding is harmful and
> should be avoided.
> Which is your preference?
I knew at some point someone would reveal that i know way too little about Unicode :). I've read the specs and looked at the mappings charts at http://www.unicode.org/charts/case/ , and (from the probably limited perspective of being focused on german script) i think that:
- lowercasing (2) would at *least* be what users would expect to happen - no user would understand if suddenly "http://Österreich.at" would not work anymore.
- full case folding (1) would retain backwards compatibility for the "sharp s" case (upper and lower case map both to "ss", like it used to be, so "http://www.faßbier.at" ("draught beer") still works, and would also remove the potential that ligatures can be used for phishing.
So, after spending 10 mins on that, i *think* at least for characters relevant for german script, full case folding (1) is the best solution. Please educate me about the ratholes. I do think that titlecase is harmful.
And, i think any mapping would need to be REQUIRED, for a consistent user experience. We would probably need to introduce a new terminology for strings before the case folding / lowercasing - but those strings will be what users type in. (and, from a GUI perspective it would make sense that the case folding / lower casing output is visible to the user, much like today when a user enters http://www.IETF.org , the browser switches the location to http://www.ietf.org/ )
Ugh, more ratholes on the radar, i fear..
More information about the Idna-update