Re: HTML APIs

At 03:32 PM 12/29/98 -0500, keshlam@us.ibm.com wrote:
>Revisiting the HTML DOM, which I'd mostly ignored in the past...
>
>If someone hands a parser a badly-formed HTML file without the <HTML>
>wrapper, does the DOM assume that an HTML element will be synthesized to be
>the single content root (DocumentElement)? Or should an HTML DOM accurately
>record the multiple-root structure of the broken file?

We say NOTHING about what an implementation does with badly formed HTML.
Obviously Netscape and Microsoft are free to continue their practice of
"fixing up" broken HTML and somehow representing it in the DOM.  The DOM WG
has argued this one to death, and all we can say is that if you want to
*ensure* interoperability of your HTML page across all compliant DOM
implementations, you *have* to make it valid HTML 4.0.  If you're
comfortable with placidly assuming that NS and MS will "do the right thing"
with invalid HTML, fine, but don't whine at *us* when your pages don't work
with some new version of a browser ;-)

Mike

Received on Wednesday, 30 December 1998 10:12:32 UTC