In my ignorance, I'm going to back out of this debate.
My position is the XSL-FO spec was written by people fully appreciative
of the delicacy of writing mode (lr-tb etc).
I'd prefer ODF to follow that model rather than invent new ones.
A couple of comments Thomas.
On 14/06/07, Thomas Zander <> wrote:
> > Inferring the wanted text mode from the character seems somewhat
> > authoritarian to me.
>
> In KOffice we use the FIRST typed character on a new paragraph for the
> whole paragraphs writing mode. Its been shown to be correct in almost all
> cases.
> It surely is something that users have shown appreciation for.
Which seems to me like setting a precedence?
I can't say if your inference is appropriate.
>
> Do note that you need the text-direction for a paragraph and that the user
> just right-aligning text by pressing that toolbar button is not
> sufficient.
I think right-alignment and text-mode are orthoganal?
> > How would you define the directionality of auto?
>
> In practice I think you can call it 'undefined'.
IMHO it could / should be defined.
Options may be by the application, or inferred from the language
and country code.
>
> I'll explain it in a different way;
> All the paragraph properties coming from a style have a defined property.
> Think about the strike-through-color to the fontfamily properties. They
> are all predefined when you start a new paragraph. The text-direction is
> different. Before you start to type a paragraph it doesn't always need to
> have a direction.
I believe it should. I'd be surprised if my Writer application started laying
out text rl-tb.
> The point at the end of the day, naturally, is how we want to make things
> easiest for the users of the ODF implementations.
For all users, yes.
> And I have to say I didn't even find any way to alter the directionality
> of a paragraph in OOWriter, so don't try things on that app unless you
> already know how to do such things ;)
I must admit in my biassed ignorance, I haven't tried.
>
> As an end user I can expect certain workflows. If the text-direction is
> set on a page, then I think we can expect the new paragraph to get that
> text direction. Even if the first character typed is different. This
> makes sense for 1 English sentence in a further Hebrew text.
> On the other hand, there are cases where I want the application to be free
> to deduct the text direction from the text typed. And this usecase is
> currently impossible.
IMHO, XSL doesn't do that.
Quote. Conforming implementations must support at least one of the
"writing-mode" values defined in this Recommendation. Although
writing-mode is defined as a Basic property with an initial value of
"lr-tb", it is not the intention of this specification to impose this
particular, or any other, writing mode value on conformant
applications. If an implementation does not support a writing-mode
used in a stylesheet, either explicitly or by relying on the initial
value, it should display either a 'missing character' glyph message or
display some indication that the content cannot be correctly rendered.
end quote.
I.e. it should be set.
> > Direction may not be implemented.Doesn't mean to say it's wrong.
>
> I'm not sure I follow this line.
Open Office may not have implemented writing-mode.
Not doing something doesn't make it wrong. Simply says not implemented.
I'm hoping we have an I18N person in the house!
regards
--
Dave Pawson
XSLT XSL-FO FAQ.
http://www.dpawson.co.uk
