L2/06-374

Source: Patrik FŠltstršm
Date: Nov 1, 2006 10:42 PM
Subject: IDNA Proposal Background
To: Mark Davis

[snip]

The reason for not updating the draft (there is a -01, but that was
only due to editorial things (like wrapping) is that we are now in
the "silent period" just before the IETF. Very unfortunate.

Let me explain how the tables are created, so you get data on what we
would like to get input on.

We first looked at the various classes defined in Unicode 5. For each
class, we decided whether we thought that class would be included on
a scale from 1-4 where 4 is absolutely included, 1 absolutely not
included and 2 and 3 in between. The goal is of course to have a
document with only values 1 and 4 in it.

Question is, can we move 2 to 1, and 3 to 4?

You can see the result of that review in section 2.

(1) Please comment on the findings here.

Second step was to look at the blocks of codepoints, and try to say
whether a block is needed or not. This decision is only 50% based on
the findings from the above exercise. Here we also have a scale 1-4,
but also "notes" that try to say what the issues are.

You can see the result of that review in section 3.

(2) Please comment on the findings here.

Third step was to take all codepoints between U+0000 and U+0FFF (as a
"random sample") and apply the classification of classes in the first
step above, and "see what happens". The codepoints get values 1-4
based on the value of the class the codepoint belong to.

(3) Please comment on the findings here.


As you can see above, the process is very much an algorithm, and our
goal is to find a combination of blocks and classes that can be used
as an inclusion based selection algorithm of codepoints for what
codepoints can be used in IDNA.

     Patrik

[snip]