Re: Feedback on 0.3 Format

[Graham <dtcd@xxxxxxx>]

On 19 Dec 2003, at 5:38 pm, Sam Ruby wrote:

> What bothers me more is calling option 2 html.  Particularly as a 
> significant portion of the titles out in the wild are of form 2: 
> intended to be text/plain (i.e., no markup present), and yet are 
> escaped in this fashion.
>
> Telling such content producers to not escape their data in this 
> fashion is a non-starter.  Telling such content producers to mark 
> their content as if it were html invites the inclusion of markup, 
> which many feel is undesirable.

But it is already HTML and should be labeled as such! It contains an 
< which is part of HTML's layer of escaping, nothing to do with 
mode="escaped". The double-escaping when using text/html mode=escaped 
comes from 1*XML + 1*HTML, not 1*XML + 1*mode="escaped".

Taking sides:
"It is the intent of the author that you should apply a level of 
unescaping to what's in the Infoset (i.e. after one level of XML 
unescaping) and expect to get markup that will be meaningful to the 
application that processes the content."
vs
"It is the intent of the author that what's in the Infoset (i.e. after 
one level of XML unescaping) is a data string of MIME type 'type'"

(Note the second covers two attributes, and in one much less tortuous 
sentence)

The name "escaped" has never been very appropriate, so deriving 
behavior from it is even less so.

Graham