Re: [css-text] I18N-ISSUE-337: line terminator handling

[Forwarding on behalf of John Klensin]

 >  Note: John K, could you review this one in particular to see
if you are satisfied?

Two (contradictory) answers:

(1) I'm pretty sure that I understood the CSS text when I read
it in January.   Given the amount of trouble being agnostic on
the line-end convention has caused, I found the text profoundly
unsatisfying then and find it profoundly unsatisfying now.

(2) I am now clear that the CSS group has no intention of
changing the present situation and apparently likes it that way.
While I will continue to disagree, I am satisfied that the text
is adequately clear about that position.

Two observations (not further complaints or justification for
leaving this open unless others agree):

(i) Unless there is general consensus that Unicode's attempt to
introduce an unambiguous Line Separator in form of U+2028 has
been a complete failure, I suppose the CSS document would be
better off either including it as an additional alternative (to
"... document language–defined segment break, CRLF sequence
(U+000D U+000A), carriage return (U+000D), and line feed
(U+000A)...") or mentioning why it is not so included.

(ii) I believe that the Unicode Standard discussion of "NLF"
represents a better approach than the indifference ("does not
define...") expressed in the CSS spec.  I.e., one should be
permissive in what is accepted but should canonical all of them
to a single preferred form.


On 23/04/2014 09:45, Koji Ishii wrote:
>>     The discussion of line termination is very hard to follow.
>>     It isn't clear from reading it whether CSS prefers CRLF, LF, or
>>     other forms in output or whether it intends to provide a
>>     suggestion. The document itself appears to prefer CRLF in
>>     some places and LF only in others (e.g., 4.1.2). If nothing
>>     else, I hope we can agree that switching conventions in a
>>     single document or rendering is probably a bad idea.
>>     This issue may be handled by CSS Syntax, which could be referenced here.
> I believe the definition of a segment break in White Space Processing Details[1] makes this point very clear; CSS does not make any preferences nor suggestions on CR, LF, or CRLF.
> 4.1.2 mentions LF only because it’s talking about transforming segment breaks defined above to LF, which looks fine to me.
> [1]
> /koji

Received on Thursday, 7 August 2014 13:45:09 UTC