On Wed, Sep 28, 2011 at 11:30 AM, Boris Zbarsky <bzbar...@mit.edu> wrote: > On 9/28/11 2:24 PM, Dimitri Glazkov wrote: >> >> Can you help me understand what the issues with fallback are? > > Sure. If I want to attach a component to a <table> and to do that I have to > write: > > <x-my-table> > <tr><td>Content</td></tr> > <x-my-table> > > and somewhere before that point register that x-my-table should be treated > as a table, then the whole thing falls apart in a UA that doesn't understand > that registration. > > That is, this approach works if components are only used for situations that > absolutely cannot work without the component available, as far as I can > tell.
Ah, I understand! It's the same issue that you and Roland were articulating in the other thread. So, this is really a parsing issue, right? We have some invariants about parsing context validity for a (small, but angry) subset of HTML tags. Hixie, is this the same problem you were mentioned as "doesn't have fallback behavior"? :DG< > > -Boris >