BTW, you know there *is* a standard ICMP message to let you get the
local netmask by broadcasting...

Phil Nelson writes:
> 
> >  1.  Is there anyone in their right mind on a class A or B that
> >      _really_ wants broadcasts to go to _all_ machines, i.e., that
> >      doesn't subnet internally?  If not, maybe the default
> >      should be 0xffffff00, to avoid boneheads like me.  :)
> >      Or maybe a note should pop up for class A and B saying
> >      `most class A and B folks subnet internally, so the proper
> >      netmask is 0xffffff00.  Do you want to use a netmask of
> >      0xffffff00, or 0xff..0000?'
> 
> This is really hard to know what people do.  Here at WWU, on our
> class B network, I know of netmasks of 0xffffff00, 0xfffffc00,
> and 0xffffff80.  There may be more.  For this reason, sysinst
> gives the standard netmask for the class (or should).
> 
> >  2.  In sysinst's net.c, it appears that the netmask for class C
> >      nets (192.* -- 223.*) is still set to 0xffff0000 --
> >      shouldn't that be 0xffffff00?
> 
> This is a bug and should be fixed.  It may be a case of too much 
> cut and paste.  Sorry.
> 
> -- 
> Phil Nelson
> phil@steelhead.cs.wwu.edu (NetBSD/pc532 machine)
> phil@cs.wwu.edu (work)
> http://www.cs.wwu.edu/~phil