A bug, a bug, and a very annoying one!

[Richard Frith-Macdonald]

> So really what it comes down to is that &sect produces an error
> condition and thus the behavior is potentially undefined (OW tries to
> get around the problem by assuming the semicolon and going on after
> making the replacement, Netscape by assuming the author meant the &
> literally; why is the OW response "more correct"? Is there a place
> where this kind of error is defined to be handled as OW does and not
> as Netscape does?).  In either case it is just better to avoid the
> error condition in the first place.
>
Actually, when OmniWeb assumes the semicolon, it is imitating the Netscape  
bug - Netscape (or at least the older versions - I'm not sure about the recent  
ones) assumed a semicolon, when it really shouldn't.  If this emulation is  
turned off, OmniWeb acts (AFAIK) according to the spec.  I think OmniWeb is  
'more correct' because it behaves as the spec says it should.  Whether or not  
this is the most useful behaviour is perhaps debatable.

The relevant part of the HTML2 specification is at -
http://www.w3.org/pub/WWW/MarkUp/html-spec/html-spec_3.html

I don't think that this detail is changed in the 3.2 spec.