The real issue: interopability, and a proposal (Was: Consensus Call on Latin Sharp S and Greek Final Sigma)

["Martin J. Dürst"]

Hello Shawn,

On 2009/12/01 19:14, Shawn Steele wrote:
>> There seems to be little existing deployment of ß-labels out there, at least on the web
>
> I would guess that be a case against making eszett PVALID.  Existing users needing Eszett are instead making due with ss,

Yes. Existing users made due with "ue" instead of "ü" before IDNA 2003, 
and so on. Existing users might have registered a name with an sz, but 
when they try to input it on a browser, it's converted to ss, so they 
don't see the point of actually using sz.


> so adding Eszett explicitly would put those users at risk.

How/why? How did adding "ü" put the users of "ue" at risk?


> It'd be better to enable them to have a display form so they could indicate that Eszett was their preferred form.

Please note that TR46 doesn't allow that. The only thing it allows is to 
preserve an sz, but not to go back from an ss to an sz if the sz was the 
preferred form.

BTW, I'm rather sceptical about the differences between lookup mapping 
and display mapping in TR46. It looks good on paper. I'm confident that 
browser vendors will get it right, at least after a few iterations. But 
it will be very difficult to get this right for all kinds of other 
infrastructure. Having a special, new, provision for only four out of 
100'000 characters isn't a good precondition for wide, consistent 
implementation. (It's not the fact that there are two mappings, or the 
specific provisions of these mappings, that I'm worried about; that can 
easily be handled by a good library. It's that each script dealing with 
IDNs will have to figure out where to use one or the other, and keep 
both in sync,...)

Regards,   Martin.

-- 
#-# Martin J. Dürst, Professor, Aoyama Gakuin University
#-# http://www.sw.it.aoyama.ac.jp   mailto:duerst at it.aoyama.ac.jp