Comments on idnabis-rationale-01

[Frank Ellermann]

John C Klensin wrote:

 [U-label]
> Frank would like to specify length limits for these strings.

IIRC I supported "one non-ASCII is good enough", like others,
and you wanted another minimum.  Likely punycode has a clear
maximal lengths in code points, i.e. resulting in 63-4 LDHs
for the A-label, it would be nice to note this.  Maybe it is
already noted in RFC 3492, then let's forget it, it is only
relevant for developers (if at all), not in the "rationale".

 [A-label]
> I believe Frank would like to specify length limits for 
> these strings.

Yes, 63, as stated in RFC 1034, minus 4 for "xn--", is 59.
Likewise 4 for "xn--" plus something is at least 5.  That
is of course not new, but readers of the "rationale" might
wish to get a feeling of what this is all about. 

>   Example: "nonsense"

Not exactly the most interesting LDH example, it has no D
and no H :-)  

>> 2) what are "a--abcdef", xn---ghikl" in your terminology?
 
> The first is an LDH-label even though it makes me cringe.
> The second is IDNA-invalid because it looks like an A-label
> but is not a valid one.

However LDH, and in a certain sense more LDH than "nonsense".

  xn--label = "xn--" ( L / D ) [ *57( L / D / H ) ( L / D ) ]

xn---ghikl is no <xn--label>, because the third Hyphen is no
( Letter / Digit ).  Therefore it can't be a (valid) A-label,
and I don't need a decoder to see this.

>> 4) how do you name "xn--abcdef" with no punycode conversion
>> to Unicode?
 
> IDNA-invalid, since it starts with "xn--" and does not have
> a mapping to a U-label.

That is an <xn--label>.  To check that it is no (valid) A-label
I'd need a decoder.  Applications hopefully treat this "as is"
without crashing, and registries hopefully tell me to get lost.

 Frank