Re: URL parsing and IPv6 addresses

>          I think we need to send a resounding "NO!" to the IPNG wg.
>  Surely they can come up with something better.   I don't see why/how
>  dotted decimal is insufficient.   If there are four dotted decimal
>  numbers,  that's clearly an IPv4 address.   if sixteen,  then it's v6.
>  Does anyone in this group see a problem with that?

I think people that have to work with IPv6 addresses on
a daily basis would see a big problem with that (which is
I imagine why the IPng group isn't doing it that way).
They mainly need two things, 1) compactness, and 2) the
ability to easily visualize the various hierarchical/
administrative boundaries that exist in the address.

The address you suggest looks like this:

That can't easily be visually parsed.

By moving to hexadecimal, the address gains 1) considerable
compactness, and 2) makes the bit-boundaries in the address
easier to mentally parse.  By having 7 delimiters rather
than 15, the whole address can be much more easily understood
at a glance.  (I have heard that humans can mentally distinguish
about 7 objects before the brain starts internally grouping
them.  I personally max out at about 3, but that is neither
here nor there.)

The above address in IPng form looks like:


Its also a brain twister, but much much easier to deal with.
Anong other things, one can easily mentally split this address
into half, and each half into half again.   From there, one
can get a handle on its internal structure.  That is impossible
with the above long string.

So, I think you can say no to the IPng on using colons because
of the (computer) parsing problem (at least with the shorthand
notation they have), but I don't think you can say no to their
basic idea of using 8 hex chunks separated by some delimiter.


Received on Tuesday, 6 August 1996 20:27:47 UTC