Re: NEL where?

[John Cowan <jcowan@reutershealth.com>]

Tim Bray wrote:

> Really?  So you wouldn't add it to the production for "S"?
> I.e. you wouldn't be able to have a NEL in the middle of
> a tag?  All you're saying is that when it's in the data
> stream it should be mapped to \n?


That is what *I* mean, and what I will push the Core WG
to do.  Whether they do that or not is out of my hands.

> In this case, we have a 
> tempest in a teapot.


Yes.

> But I didn't think that's what the
> mainframers wanted, they wanted to be able to say
> 
> <img src="x.jpg"
>      alt="the mighty X" />


The IBM note (http://www.w3.org/TR/newline) does ask for that
capability as well, modifying the production for S in section 2.3.
It seemed to me (NSFTCWG) to be unnecessary.  IBM has not yet
weighed in officially or unofficially, though.

> I thought David Carlisle pointed out a real problem:
> if I can have NELs in the middle of my encoding declaration,


If, yes.  But that would involve changing S.


> but it
> was a nice characteristic that the XML encoding declaration 
> was all characters with values < 128... i.e. you could read
> and interpret bytes before you decided to turn on the
> UTF-* machinery.  -Tim


Of course, that only works if you confine yourself to ASCII-
compatible encodings.

-- 
There is / one art             || John Cowan <jcowan@reutershealth.com>
no more / no less              || http://www.reutershealth.com
to do / all things             || http://www.ccil.org/~cowan
with art- / lessness           \\ -- Piet Hein