W3C home > Mailing lists > Public > public-i18n-bidi@w3.org > April to June 2013

Using bidi control characters in HTML

From: Richard Ishida <ishida@w3.org>
Date: Wed, 24 Apr 2013 16:25:24 +0100
Message-ID: <5177F964.4040508@w3.org>
To: public-i18n-bidi@w3.org
I've begun revising the article we have[1] that tells people how to use 
Unicode control characters when markup isn't available, in light of the 
new characters that will be available with Unicode 6.3.

I think people need to know *how* to use the control characters, rather 
than just what they are. (And the couple of examples that are in the 
article also need some revision.)

One way would be to say "Read the article about inline bidi markup, and 
use character x,y or z wherever you would have used inline markup."

This would result in many users resorting to tightly wrapping all 
opposite-direction phrases in pairs of control characters.  I'll refer 
to this practice as TWODP.

I wondered whether this is the best approach when not dealing with markup.

If you know the direction that needs to be applied to text, you are 
typically dealing with one of the following scenarios (examples given 
for LTR context):

[a] embedded base direction	eg. abc "ghi FED" jkl

[b] neutrals on boundaries	eg. abc !FED ghi

[c] lists			eg. abc FED, IHG jkl

[d] number isolation		eg. abc FED 3 ghi

Although TWODP reduces the brain strain for the author, scenarios b, c 
and d can actually be resolved using LRM/RLM, and this involves a simple 
insertion of a single, non-paired character.

What should we recommend?

[1] http://www.w3.org/International/questions/qa-bidi-unicode-controls

Richard Ishida, W3C
Received on Wednesday, 24 April 2013 15:25:56 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:24:40 UTC