Consensus Call on Latin Sharp S and Greek Final Sigma

Alexander Mayrhofer alexander.mayrhofer at
Mon Nov 30 16:22:08 CET 2009

> > I would be perfectly happy with the introduction of the "ß" 
> if it would
> > happen on greenfield. Unfortunately, "ß" is being built on a heavily
> > contaminated brownfield.
> Thanks for stating that clearly, Alex. The question then is, 
> how can we get
> back to a "greenfield" with the entire German alphabet 
> including sharp s.


Good point - to get back to the greenfield, we'd need to achieve the following:

(i'm ignorant now, and only considering the web)

1) Get browser vendors to agree on a *single* mapping scheme.
2) Get them to implement IDNA2008 and that single mapping scheme.
3) Deploy the updated browsers to the vast majority of users, preferrably in a short timespan, so that the period of uncertainty and incompatibility is as short as possible.
The worst thing to happen would be vendors implementing different mapping schemes, or some ignoring IDNA2008 completely. The character is "tainted" already anyway, but either of those two things would render it unusable.

> If we make sharp s DISALLOWED now the decision will be 
> definite (and in my
> eyes wrong or at least at a wrong level (layer)). If we make 
> it PVALID this
> will leave the choice open to the registries (!) for the 
> future. And there
> are many ways to get this fixed considering reasonable planning and
> timeframe. Bundling or grandfathering would be one way, 
> waiting until almost
> all applications switched from 2003 to 2008 would be another 
> one - even if
> it takes time. At the end can also refuse to have 
> sharp s registered
> at all, not even thinking about a possible rollout. So my 
> recommendation
> would be to keep the possibilities open for the future 
> because circumstances
> may change in the near and far future. There should be no 
> exigent reason for
> you to bury the sharp s today for all time. And finally 
> mappings like in
> 2003 will no longer be part of 2008 anyway.

As "definite" as the timespan between IDNA2003 and IDNA2008? (sorry, couldn't resist.. - and i know how much of a hassle it is to update a protocol ...)

I haven't seen many other protocols where identifiers are immediately re-assigned or re-defined. This seems to be a rare exception.


More information about the Idna-update mailing list