IPv6 address/port format

[Rene Mayrhofer]

MARAY Tamas wrote:
> 
> > >Wrong. Conflicts with the UNIX shell.
> >
> > Right, if you care about proposed standards.  RFC 2732 says to use
> >
> > http://[FEDC:BA98:7654:3210:FEDC:BA98:7654:3210]:80/index.html
> >
> > If you disagree with RFC 2732, take it up with the Internet standards
> > process.
> >
> > In any case, there are a lot of characters that conflict with the Unix
> > shell, all of which are valid in the pathname section of a URL.  '?'
> > means "any character" to some shells, does that mean we forbid its use
> > for CGI scripts?
> 
> Of course not. That is an unlucky case and we must live with it
> since we have no impact on it anymore. But why shouldn't we create a
> solution for IPv6 URL syntax which satisfies all the requirements
> if we are still in time and position? If we choose the syntax watchfully,
> such kind of conflicts can be avoided.

But I think that there is another point that has to be taken into
account: Users must live with the solution. When we want a smooth
upgrade from IPv4 to IPv6 (OK, as smooth as it can be anyway), then the
change on the user interface should be minimized. 
Remember, the majority of people using the Internet does not care about
IP, they only use their tools like Web-browsers or email-clients.

I think it is very important for the acceptance of IPv6 that users
should not notice the change at all. With URLs, there would only be a
very small change if the above format would be used.

But you are completely right: There is a problem with UNIX shells that
can be solved easier when we think about it now. Is there an easy way to
cope with it without changing the format for URLs ? Can the "[" and "]"
simply be escaped in the shells ? This should be possible when the tools
are aware of it. 

cheers,
Rene Mayrhofer