Re: Comments on section 2 of RFC2396bis

Chris Haynes wrote:
>  "Tim Bray" proposed:
> 
>>- ASCII characters which may legally appear in the component MUST
>> appear directly as themselves, i.e. 'a' may not be encoded as %61.
> 
> My understanding is that the decision on whether or not an ASCII
> character 'may legally appear' (at any particular position in the URI)
> is not at all well defined and is scheme-specific.

Scheme-specific I believe.  "Not well defined" seems like a symptom of a 
broken scheme.

> 
> Witness the several recent requests to this list by Israel Viente re:
> file:/e:/xxx.pdf  versus  file:/e%3a/xxx.pdf and the like - I don't
> think anyone was able to give him quick, unambiguous answers.

This would tend to support the general feeling that the "file:" URI 
scheme is basically broken.  Further evidence of that fact is that there 
is very poor interoperability in the field.

> Unless and until the URI RFC _and_ all schemes can be shown to have
> completely unambiguous syntaxes in this respect, developers will have
> to 'play safe' and escape questionable characters.

I see your point, but if a scheme is so broken that a programmer can't 
be sure what needs escaping, then it needs fixing.
-- 
Cheers, Tim Bray
         (ongoing fragmented essay: http://www.tbray.org/ongoing/)

Received on Saturday, 8 March 2003 18:23:42 UTC