Re: Modes and types on titles etc.

[Graham <dtcd@xxxxxxx>]

On 19 Dec 2003, at 8:04 pm, Henri Sivonen wrote:

> A while back I noticed that Blogger was generating Atom 0.3 feeds with 
> mode and type attributes on elements other than <content>. Then, Mark 
> Pilgrim announced the Atom 0.3 snapshot with "expanded content model" 
> for some elements. I didn't even notice, when I read 
> draft-nottingham-02 that there was a drastic change like this.
>
> Where did this come from? Did I miss something on this list? On the 
> wiki?

I have no idea where it came from either, and I was shocked. But the 
thing is that if you have code to process <content>, the added 
complexity of processing title is pretty negligible. So I think it's 
fine.

> 1a) XML-only titles

What you propose isn't much easier than stripping everything between a 
< and a > in text/html (you have to do something with entities too of 
course). I don't personally expect any serious tag-soup fishing to go 
on.

> 1b) Only straight XML or entity-encoded inline HTML in title
> If we can't agree to XML-only titles, can we at least not pretend that 
> any MIME type is OK in the title?

This is probably something for the inter-op RFC.

> 2) text/plain is inappropriate for denoting whitespace-collapsed XML 
> character data

The default is XHTML, is it not?

> 3)  Need to define the processing of HTML fragments
> Also, it should be defined that HTML and XHTML

Yep, that needs to be covered in the spec.

Graham