Conformance according to rules HL4 and HL5

iancu · **Joined:** Thu Feb 11, 2010 12:47 pm **Posts:** 11

Rule HL4 states that the bidi algorithm can be applied to text segments. Rule HL5 states that text may be flanked by additional characters to provide artificial context. By combining the two rules, one can infer that an implementation may insert characters in the text to process (at places it identifies to be segment boundaries) and apply the algorithm to the augmented text, while remaining conformant.

Is that a correct reading of the rules?

unicode · **Joined:** Thu Dec 23, 2010 7:46 pm **Posts:** 1

Mark Davis replied in another context:

Quote:

Yes, if you mean logical insertions. That is, you could have

<p title="XXXX">

and display that as if it were logically

<p title="<LRM>XXXX<LRM>"> // or embedding

but the inserted characters wouldn't show up in the display, of course.

Mark

matial · **Joined:** Sun Aug 22, 2010 5:14 am **Posts:** 5

In the coming version of ICU (4.8), a draft function has been added to implement HL5. Look for ubidi_setContext (ICU4C) or setContext (ICU4J).

The Unicode Consortium Discussion Forum

Conformance according to rules HL4 and HL5

Who is online