Re: Right-to-Left Punctuation Problem

From: Philippe Verdy (verdy_p@wanadoo.fr)
Date: Fri Jul 29 2005 - 16:23:35 CDT


I forgot to say that the effect of RLE terminates at end of a paragraph. So
when you send a plain-text email like this one, whose paragraphs will be
splitted into several lines with limited length, the final punctuation may
occur on a distinct line which, according to default Unicode rules, will be
handled as a separate paragraph that does not start with the RLE control.

For this reason, the line inherits from the main diractionality of the
document, and then starts and ends in LTR context, and the final punctuation
appears on the left.

To solve this problem for LTR documents, you can use HTML email because its
parsing will keep paragraphs logically joined even if the message is encoded
on multiple lines.

The other solution is to read the email using a RTL display preference.
(You'll note that this changes the way the English text is displayed, and
its final punctuations may appear on the wrong side as well)

The good question to ask is whever there is a solution to specify within the
document itself the default directionality, so that it will span several
lines or paragraphs.

Using RLO to force the directionality is not a solution because it also does
not span paragraphs (its effect is to to force the directionality not only
characters with weak direction, as done with RLE, but also override the
directionality of characters with strong directionality such as digits or
Latin letters).

Using RLM may be another option to mark a single character (used for example
just before the final punctuation) with another default directionality.
However it will not be always reliable for emails, because there may
sometimes be a line split between the RLM and this punctuation (this should
not occur unless your line has no spaces and consists only in a single long
word terminated by this punctuation). Generally for plain-text emails, using
RLM is the best option.

----- Original Message -----
From: "Philippe Verdy" <verdy_p@wanadoo.fr>
To: "Adam Reisman" <adamreisman@hotmail.com>; <unicode@unicode.org>
Sent: Friday, July 29, 2005 10:46 PM
Subject: Re: Right-to-Left Punctuation Problem

> There's no anomaly here: what you have created is a left-to-right document
> that contains insertions of Hebrew. At end of the sentence, the document
> properties apply to the final punctuation.
> If you want to make sure that a complete Hebrew sequence (including its
> punctuation) will be interpreted right-to-left, you have to surround the
> Hebrew text with right-to-left bidi controls: The next paragraph for
> example starts with a RLE (Right-to-Left Embedding) Bidi control, and it
> is terminated by a PDF (Pop Directional Format), so that the cursor goes
> to the right when I use the "End" key:
>
> ‫המשפט הראשון בסדר! המשפט השני הוא גם־כן בסדר! אבל הפיסוק למשמפט האחרון
> עובר להתחלת הטכט!‬
>
> Your other example:
>
> ‫למה?‬‬
>
> The PDF at end is not absolutely required (only if you intend to resume
> the LTR processing of punctuation for your document), you can use a single
> RLE control at the beginning of your text.
>
> The semantic of the punctuation is not changed (only its presentation).
> You don't need other RTL punctuations...
>
> Also, you don't need RLE/PDF if your document is created using a default
> RTL context, for example if it uses a Hebrew charset, or if there's some
> other meta-data specifying that its language is Hebrew or that its
> presentation should be RTL by default for characters with weak
> (contextual) directionalities such as punctuation.
>
> Shalom!!
> ‫תודה רבה!!‬‬
>
> Philippe.
>
> ----- Original Message -----
> From: Adam Reisman
> To: unicode@unicode.org
> Sent: Friday, July 29, 2005 8:57 PM
> Subject: Right-to-Left Punctuation Problem
>
>
> Has there ever been talk about creating a complete set of right-to-left
> Hebrew punctuation. I'm having problems with punctuation direction...or
> maybe I'm just doing it wrong. I have no problem when Using MS Word, but
> when I type up an email in Internet Explorer, the punctuation always
> jumps to the wrong end of the sentence. Then when I type the next
> sentence, the punctuation in the previous sentence jumps back to where
> it's supposed to be. This isn't a problem until I get to the last
> sentence of a paragraph or if I type a one-sentence line:
>
> המשפט הראשון בסדר! המשפט השני הוא גם־כן בסדר! אבל הפיסוק למשמפט האחרון
> עובר להתחלת הטכט!
>
> In the above example, the exclamation point on the far right, should be on
> the far left because it belongs to the last sentence.
>
> שלום!
> In the above example, the word "shalom" should have the exclamation point
> on the left.
>
> למה?
>
> Why?
>
> Is this a configuration issue I can correct, or is this something I'm
> doing wrong, or just something that people have to deal with? I can
> partially solve this by using the gershayim (U+05F4) as quotation marks,
> and the sof-pasuq (U+05C3) as a colon, but that's not enough. I thought
> about using Arabic punctuation, because it's coded right-to-left (it has a
> character I could use as a Hebrew comma, and the semi-colon and question
> mark are close enough even though they are inverted), but I worry that it
> will come out garbled when the person I write to tries to read it.
>
> (as an aside, another reason for Hebrew punctuation would be that in the
> Times Roman font, the punctuation doesn't match the Frankruhl Hebrew! The
> Hebrew period and all other punctuation dots should be diamond-shaped).
>
> Any thoughts or suggestions?
>
> תודה רבה!!
> Thank you!!
>
> P.S. Thank you for all who advised me on a font maker. I think Fontforge
> is the program I will start with, but I will also check out all the
> shareware suggestions you gave me :-)



This archive was generated by hypermail 2.1.5 : Fri Jul 29 2005 - 16:24:46 CDT