RE: lineBreakType

Hi Shaun,

I've never seen a 1.1 XML document in the wild yet.
But it's a good point.

Indicating the number of bytes would work, but would make create room for possible mistakes like changing the encoding but forgetting to change the default line break size.

Another way would be to store the actual string representing the line break. But that may not go well with the white space handling of attributes according XML.

As for the name, I kind of agree: since we have 'storageEncoding' 'storageLineBreak' would be consistent.

Cheers,
-yves



-----Original Message-----
From: Shaun McCance [mailto:shaunm@gnome.org] 
Sent: Tuesday, October 23, 2012 3:57 PM
To: public-multilingualweb-lt@w3.org
Subject: lineBreakType

Two thoughts on lineBreakType:

1) The attribute name, taken on its own, implies it's something about the content itself, rather than some storage mechanism.
I think it should be prefixed with "storage" like the other Storage Size properties.

2) I think I was the one who brought up allowing NEL. I realize now that XML 1.1 allows any of the following as a newline:

  U+000D U+000A
  U+000D U+0085
  U+0085
  U+2028
  U+000D
  U+000A
  U+0085

If you use U+2028 and UTF-8, that's actually three bytes. Does anybody use XML 1.1 or U+2028? Probably somebody somewhere.

Perhaps instead of specifying the newline type of the storage mechanism, we should instead just specify how many bytes a newline takes up, e.g.

  storageLineBreakBytes="2"

--
Shaun

Received on Tuesday, 23 October 2012 22:12:34 UTC