At 09:30 p.m. 26/06/2000 +0200, Andreas J. Koenig wrote:
> >>>>> On Mon, 26 Jun 2000 09:48:04 -0400, "Luis E. Muņoz" <[EMAIL PROTECTED]>
> said:
>
> >> I'm not sure, but I could imagine somewhere in the NetPacket or
> >> Network hierarchie or below the Net::DNS there might be a place to
> >> consider settling.
>
> > The modules I referred to deal specifically with representations of
> > IP addresses. They're protocol independent.
>
> > Under 'Net' there is essentially stuff related to networking and
> > actual implementations (protocols, clients, servers, tools, etc).
> > While logically the modules I wrote about could fit in something
> > like 'Net::IP', I'm not sure if this approach would make sense in
> > the long run.
>
> > On the other hand, something like 'Net::Address::IP' might make
> > more sense. The Tie interface could then be Tie::Net::Address::IP
> > so it could all boil down to
>
> > IP::Address --> Net::Address::IP
> > Tie::IP::Address --> Tie::Net::Address::IP
> > Net::Subnet::Count --> Net::Address::IP::Count
>
> > Please let me know what you think about this one, as I feel this
> > makes more sense.
>
>It indeed looks much more intuitive than before. Just one thing,
>there's a long standing tradition to discourage
>extremely::deep::nesting::namespaces. E.g. one could shorten
>Net::Address to NetAddr. No matter how ugly it looks, it seems
>understandable and less to type. Leading to
>
> IP::Address --> NetAddr::IP
> Tie::IP::Address --> Tie::NetAddr::IP
> Net::Subnet::Count --> NetAddr::IP::Count
This is ok with me, but it will create another item into the 'root' of
the name space :) I guess it's your call in this one ;) You might want to
consider Net::Addr::... to avoid the new root name (unless NetAddr already
exists)
Regards.
