Bidirectionality and Domain Names
Intended Audience: |
Software Engineers, Systems Analysts |
Session Level: |
Intermediate |
Unicode's ability to represent multilingual text makes it good candidate
for establishing a domain name structure. Unicode brings not only an
encoding framework, but also support for things like bi-directional
scripts. Additionally, the collection of Unicode's character equivalences
is both desirable and at times necessary given Unicode's goal of encoding
natural language text. These equivalences, however may present problems in
the context of domain names.
Unicode's Bidi algorithm as currently specified may also be unsuitable for
determining an appropriate display ordering for domain names. Specifically,
the Bidi algorithm itself possesses a set of implicit assumptions about the
usage of common characters. This set of assumptions may not be applicable
to domain names. Domain names use the same repertoire of characters that
appear in text. This requires a different algorithm for handling domain
names.
In this paper we propose how domain names can accommodate different reading
orders. In particular, this paper offers an algorithm for determining the
display order "reading" of domain names. Additionally, we relate this
notion to Unicode's Bidi algorithm.
|