Re: HTML 2.0 LAST CALL: Numeric character refs

Eric W. Sink (
Mon, 5 Jun 95 09:23:52 EDT

>Hmm... I guess I saw it differently. I'm sorry to drag this out, but
>Eric: could you give the chair's position on this, so we can move on?

1. I agree that constructs like &N; where N>255, are errors.

2. I do not believe that the RFC needs a full specification of how a user
agent should handle bad HTML, though having one would certainly have saved
me personally a lot of time. ;-)

3. I agree that Netscape and Mosaic handle this case in an essentially
broken fashion. We shouldn't spec "char ch = atoi(N);".

So, I recommend we simply not spec the behavior of a user agent for this case.

Eric W. Sink
Senior Software Engineer, Spyglass