Data on confusables

[Gervase Markham]

On 27/07/09 07:54, Mark Davis ⌛ wrote:
> Now, as with any statistics, the data is only an approximation.

It seems to me that the appropriate question to ask when judging impact is:

What percentage of domain names contain at least one character which is 
confusable with another character permitted by IDNA2003, but no 
characters which are confusable with characters permitted by IDNA2008?

In other words, how many domain names move from the "possibly spoofable" 
category into the "not spoofable category"?

You say that in IDNA2008, 4.17% of PVALID characters have different 
IDNA2008 PVALID character they are confusable with. What is the 
percentage of IDNA2008 PVALID characters which are confusable with a 
PVALID character in IDNA2003? (Yes, I have asked that question exactly 
as I meant it.)

Gerv