> > > would "-" work (this would imply that "-" is not permitted in > > > the label name-space)? There's also %, which has already been used > > > for IPv6. > > > > Amusingly, there's a separate thread on crossbow-discuss about allowing > > '-' as an interface/datalink character -- so that'd need to be stopped > > yes, I noticed, but this is a totally different context- we are not > talking about the namespace for datalinks, but rather > about a *separator* char between the datalink name (where the exclusion > of '-' has been noted, and serendipitiously ends up avoiding namespace > conflict) and the address slot/identifier itself
My point is that if the IP interface name can contain a "-", then "-" becomes ambiguous in your proposed context. -- meem
