[Zope3-dev] Re: XML header and TAL interpretor

[Tonico Strasser]

Andreas Reuleaux schrieb:
> On Thu, Sep 01, 2005 at 11:44:48AM +0200, Tonico Strasser wrote:
> 
>>...
>>And, as long pages are served as text/html they are treated as old-style
>>HTML by browsers anyway[1].
>>
>>XHTML pages served as text/html must follow the compatibility
>>guidelines[2]. E.g. in ZPT HTML mode, elements like <br/> will be
>>converted to <br /> automagically (with space before the slash). That is
>>a good thing IMHO.
>>
>>[1] Serving XHTML 1.0
>><http://www.w3.org/International/articles/serving-xhtml/>
>>
>>[2] HTML Compatibility Guidelines
>><http://www.w3.org/TR/xhtml1/#guidelines>
>>
>>Tonico
>>...
> 
> 
> Tonico, I think you misunderstood something: While it is true that
> pages served as text/html are treated as HTML by browsers,
> there are still quite some differences how they treat the HTML,
> i. e. in which mode they operate (doctype switching).

I know.

> Serving pages as text/html just makes the browser more forgiving.
> E. g. a plain <br> tag won't normally be flagged as an error, while in
> application/xhtml+xml it will be flagged as an error by the browser
> (and Mozilla/Gecko does so). IE does not understand
> application/xhtml+xml, so this is only a choice for intranets without
> IE right now. For the Internet there is currently no choice but text/html
> (maybe negotiating the MIME type depending on the browser).

Ok.

> Also there is nothing wrong with serving XHTML as text/html (except
> that you are missing some browser error messages - if you want to make
> sure your XHTML is correct, you can always rely on external
> validation, see below)

Ok.

> Still - and that is my point - even IE (that only understands
> text/html) has different modes of operation (basically standards mode
> and quirks mode), triggerd by doctype switching.

What's your point?

> If you want to use XHTML and make sure that that XHTML is correct,
> this is what you can do:
> 
>   (1) rely on the zope machinery to find possible mistakes: XML-mode
>       is not as forgiving as HTML-Mode for page templates - and I
>       think that is a good thing, and therefore I am also in favor of
>       making XML-Mode the default (+1).
> 
>       Be aware though that even in HTML-Mode you can currently write
>       perfectly valid XHTML pages - you are just loosing Zope's
>       ability to detect XHTML errors.
> 
>   (2) rely on external validation: html://validator.w3.org
>       this should probably be done much more often anyways
> 
>   (3) rely on the browser to flag errors
>       e. g. Mozilla/Gecko in application/xhtml+xml, NOT possible
>       in IE.
> 

I do the same :)
http://mail.zope.org/pipermail/zpt/2004-July/005387.html

> Beware: this is all about what input is accepted as valid XHTML,
> interpreting the input as XHTML is the next step: normally you want
> that XHTML to be interpreted in standards mode (that's why you wrote
> XHTML in the first place) and you tell the browser by doctype
> switching.
> 
> Philipp: If XML-Mode is the default, and therefore no xml-declaration
> is needed (and the aforementioned unfortunate IE bug is therefore
> circumvented) how would one toggle on HTML-Mode? - Am I missing
> something? The situation right now is
>   xml-declaration -> XML-Mode
>   no xml-declaration -> HTML-Mode
> HTML mode might still be needed in some situations (Fred gave some
> examples)
> 

Tonico