Re: The elements to which unicode-bidi:isolate applies

> The spec says
>
>  # User agents that support bidirectional text must apply the Unicode
>  # bidirectional algorithm to every sequence of inline boxes
>  # uninterrupted by a forced (bidi class B) line break or block boundary.
>  # This sequence forms the "paragraph" unit in the bidirectional
algorithm.
>
> In what way is this not sufficient to address your concerns?

Taken by itself, it is perfectly explicit and sufficient.

The problem is that the unicode-bidi:isolate spec is also perfectly explicit
and sufficient, and, it seems to me, conflicts with the above:

for the purpose of bidi resolution in its containing
paragraph (if any), the [unicode-bidi:isolate] element itself is treated as
if it were an
Object Replacement Character (U+FFFC).

When both specs apply, I think it should be either implicitly obvious or
explicitly stated which one wins. I do not think it is implicitly obvious,
so I would like to state it explicitly.

Once we are at it, it is nice to mention the other cases
where unicode-bidi:isolate has no effect, like the first two you mention,
although this fact is implicitly obvious from the specs, just for the
convenience of the reader (not as a normative definition), but I would not
have brought those cases up if it were not for (what I think are) the
ambiguous ones.

Aharon

On Sun, Sep 26, 2010 at 4:26 PM, fantasai <fantasai.lists@inkedblade.net>wrote:

> On 09/25/2010 05:25 PM, Aharon (Vladimir) Lanin wrote:
>
>> The current definition
>> <http://dev.w3.org/csswg/css3-text-layout/#unicode-bidi>
>>
>> of unicode-bidi:isolate states:
>>
>> "For the purposes of the Unicode bidirectional algorithm, the contents
>> of the element are considered to be inside a separate, independent
>> paragraph, and for the purpose of bidi resolution in its containing
>> paragraph (if any), the element itself is treated as if it were an
>> Object Replacement Character (U+FFFC). (If the element is broken across
>> multiple lines, then each box of the element is treated as an Object
>> Replacement Character.)"
>>
>> I think that it is worthwhile to stipulate that unicode-bidi:isolate has
>> no effect whatsoever on any element that either creates a separate UBA
>> paragraph, or serves as a UBA paragraph break in its containing UBA
>> paragraph. Such elements include: any element taken "out of flow" e.g.
>> with float or position:absolute, any element with display other than
>> "inline", <br bidibreak=hard>, <textarea>, <input type="text">. This
>>
>
> Elements that are out-of-flow do not affect their surrounding contents.
> That is the definition of out-of-flow, and it needs no further
> clarification here.  As for atomic inlines such as <textarea> and
> <input>, they are handled as U+FFFC under the non-textual entities
> clause (which will be further clarified in the next revision of CSS2.1)
> and the bidi independence of their contents are covered by the spec text
> quoted below.
>
>  stipulation is particularly important for the display:block and <br
>> bidibreak=hard> elements, for whom unicode-bidi:isolate's specification
>> that the element be treated in the containing paragraph as if it were
>> U+FFFC, i.e. bidi class ON, conflicts with the proposed specification
>> <
>> https://docs0.google.com/a/google.com/document/edit?id=1zR06HjhVvt7ySAJeqq7zQZSzpLTRgV8AoQNo_vNznX4&pli=1&authkey=CIXomoYI#
>> >
>>
>> that these element be treated in the containing paragraph as if they
>> were bidi class B.
>>
>
> The spec says
>
>  # User agents that support bidirectional text must apply the Unicode
>  # bidirectional algorithm to every sequence of inline boxes
>  # uninterrupted by a forced (bidi class B) line break or block boundary.
>  # This sequence forms the "paragraph" unit in the bidirectional algorithm.
>
> In what way is this not sufficient to address your concerns?
>
> ~fantasai
>