Re: The HTML q element can sometimes be useful. Discuss.

On 4/27/2016 2:31 PM, Tex Texin wrote:
> Asmus,
>
> Regarding " Here's a clear case of something that markup could (have been) used for to provide scope, and so obviating a need for heuristic scanning (which will fail in the case of non-balanced parens."
>
> I hope you are not suggesting that markup be used to solve problems with the bidi algorithm. The bidi algorithm needs to work at the level of plain text.

Bidirectional layout does not need to (only) work on plain text. It has 
to work.

If (hypothetical) parenthetical expressions where done as elements in 
HTML, then for HTML, there would be no need to run a heuristic to figure 
out the scope of nested parenthetical expressions.

I wanted to state no more and no less.

>   As I think you know, I don't believe markup should have the equivalents for the Unicode bidi controls. It makes sense to have the larger flow controls (page, paragraph), but not for the inline text direction. Using markup for the finer text direction means that the bidi source must look incorrect and authors can't use bidi-knowledgable plain text editors. It also means the text cannot be sourced from databases or elsewhere, where the content can also be shared with non-html platforms. Not to mention the fact that markup can't be used within some of the html attributes so they cannot have their direction set properly by bidi markup.  I have been ranting about this ever since utr #20 said that bidi markup should be favored over Unicode controls.

This is an orthogonal issue in a way.

The bidi controls are inextricably mixing styling and semantics. They 
force directional interpretation of text, but they do not tell you 
"why", and not all of them can be copied into new locations in the text 
with predictable results.

Semantic markup instead always identifies the structure of the text; 
this can then be analyzed and turned into instructions to override the 
bidi behavior (or, in the case of the parens, the mirroring behavior).

At this point, we don't have standardized semantic markup for 
parenthetical expressions, so the issue is moot - merely an interesting 
"what-if' to contrast with the <q> element.

As to being able to use a plain-text editor: try keeping (deeply) nested 
quotes straight using the <q> element. I bet you have to be pretty 
expert to be able to predict what the appearance will be.


>
> Besides, I am not sure that markup would do a better job of figuring out which way to mirror in multiple nestings, any more than it  can figure out which quote to use in the same scenarios.

Actually it can (could). Because, unlike quotes, the mirroring doesn't 
change by nesting level. The bracket pair would be explicitly tied to 
the scope of the element and whatever the mirroring is for one, that's 
the mirroring for the other.

If you had a way to disable the bracket resolution from the bidi 
algorithm for any text that you had semantically marked up, you would be 
able to safely include unpaired brackets in your text, something that 
might be useful for mathematical texts.

Anyway, all hypothetical.

A./


>
> (Sorry, bidi markup touches a raw nerve!)
>
> tex
>
>
>
> -----Original Message-----
> From: Asmus Freytag (c) [mailto:asmusf@ix.netcom.com]
> Sent: Wednesday, April 27, 2016 1:50 PM
> To: Tex Texin; 'John Cowan'; 'Florian Rivoal'; ishida@w3.org; 'W3C Digital Publishing IG'; 'www International'
> Subject: Re: The HTML q element can sometimes be useful. Discuss.
>
> On 4/27/2016 1:44 PM, Tex Texin wrote:
>> Interesting point about ASCII quotes Asmus. Thanks.
>>
>> Also your point about nesting parentheticals is true. However some parentheticals use square brackets. [sic] - or dashes - just to complicate things.
> You would have to have a parameter (I tried to allude to that with the "multiple types").
>
>> The interaction of bidi with lang and quotes and nesting would probably make most heads explode.
> Well, for parenthetical expression, the bidi algorithm was just tweaked to allow it to compute nesting levels on the fly to make sure that the brackets do not get mirrored the wrong way around (for each parenthetical both brackets should match in mirroring state, either both on, or both off).
>
> Here's a clear case of something that markup could (have been) used for to provide scope, and so obviating a need for heuristic scanning (which will fail in the case of non-balanced parens.
>
>> tex
>>
>>
>> -----Original Message-----
>> From: Asmus Freytag (c) [mailto:asmusf@ix.netcom.com]
>> Sent: Wednesday, April 27, 2016 1:24 PM
>> To: Tex Texin; 'John Cowan'; 'Florian Rivoal'; ishida@w3.org; 'W3C Digital Publishing IG'; 'www International'
>> Subject: Re: The HTML q element can sometimes be useful. Discuss.
>>
>> On 4/27/2016 12:17 PM, Tex Texin wrote:
>>> Regarding usefulness, the q element is helpful to screen readers.
>>>
>>> Florian commented on the lack of a sentence element. Since paragraphs are a series of sentences, and are generally clearly separated by periods, perhaps there is less of a need.
>>> However, we might contend that if there is a need to semantically earmark quotes, there is also a justification for doing so to questions and exclamations.
>>> (I am not asking for this, I am just noting the analogy.)
>> I suspect that getting "nice" quotation marks from ASCII-only input was a big driver. That added a convenience not needed for other punctuation.
>>
>> I also wonder whether their use is (or can be made to be) beneficial for bidi layout.
>>> Certainly, they would help screen readers. And to the extent that question marks and exclamation marks change with language, the alleged benefits of the q element providing quote marks would also apply to these statements. Has there been any demand for this?
>>>
>>> (I don't think <em> or <strong> are equivalents for exclamations,
>>> although they might be used along with them.)
>>>
>>> Parenthetical remarks would also benefit from having their own
>>> elements
>> These are interesting because they use paired punctuation in a similar way (although there can be multiple types that are not necessarily "nested" the same way quotations are. (The scope ordinarily nests, but while more than one set of marks can be used, they are not alternating by nesting level).
>>
>> A./
>>
>

Received on Thursday, 28 April 2016 00:36:12 UTC