Protocol Action: 'Right-to-left scripts for IDNA' to Proposed Standard

Slim Amamou slim at alixsys.com
Sat Feb 13 10:49:42 CET 2010


On Sat, Feb 13, 2010 at 5:44 AM, Abdulrahman I. ALGhadir
<aghadir at citc.gov.sa> wrote:
> Hey,
>
> Well as we know the IDNA protocol didn't adapt bidi algorithm (UAX #9) fully. They disallowed all bidi markers (LRM,RLM,...) which are they used to solve problems from this kind.
>
> So, what can be done?
> 1) adapt or create new algorithm in IDNA protocol to solve this issue (doubt it).

That won't be too difficult. Two choices :

  - Enforce LTR context (and thus network order) for the domain name
display. This is my preferred choice, since it preserves a unique
presentation for the domain name in every context and for all locales.
i.e. the domain name will look the same in china, morocco, england...

  - Define the behaviour of labels with respect to the period
separator for both LTR and RTL context.

ex.
The period MUST define a hierarchy relation between it's left label
and it's right label
in LTR  context the the hierarchy is such as right label is
hierarchically superior to left label
in RTL  context the the hierarchy is such as left label is
hierarchically superior to right label

> 2) make all hyper links to be treated in special way rather using UAX#9 (application wise possible).

This won't be too difficult since links, IRIs and domain names are
already parsed for purposes other than display.

> 3) enhance the existent bidi algorithm by adding new rules to support hyperlinks (possible).

I don't think this problem could be solved by unicode. The reason is
it's really tied to IDNs structure. IDN is a notation, like numbers
(I'm choosing this example purposefully because numbers are presented
LTR in a RTL context), and as such it has to have it's own coherency.
So you can't solve the problem in a general way.

This is actually a class of problems, let's call it "New notations",
which deals with the presentation of new notations introduced by new
and socially pervasive technologies (it's also the case of IRI for
example). IETF will have to tackle this problem one way or another.


-- 
Slim Amamou | سليم عمامو
http://alixsys.com


More information about the Idna-update mailing list