At 05:37 05/03/08, Julian Reschke wrote:
>
>Tim Bray wrote:
>> On Mar 7, 2005, at 10:31 AM, Martin Duerst wrote:
>>
>>> >for some implementors, HTML is actually easier, if they are handing a chunk of bytes to an HTML rendering control, they'd rather not reconstruct the syntax from the infoset, they'd rather just take an opaque chunk of bytes.
>>>
>>> I really hope they don't take an opaque chunk of bytes, because if
>>> they do, they'll have no clue about the character encoding.
>>
>> Not true; if you can parse the XML to find the <atom:content> then you know the character encoding and can tell your HTML renderer. -Tim
>
>I think Martin was talking about the producer: if all he has is a chunk of *bytes*, there's no reliable way top put that into an HTML-typed text construct (you'll need characters, not bytes).
It works both ways. If all you get is a chunk of bytes, things won't work. Of course the Atom processor knows more, but Tim didn't say that it would pass that info on to the renderer.
Regards, Martin.
