OK, I see now that down the road using
a 1 and 100 net address on the lab would
create unmanageable problems if those nets
were ever put into use on the internet...
something NAT couldnt fix. And the
responses saying use 1918 space point out
the potential problems were this lab ever
to leak out an advertisement on to the
internet, etc.... all advice I appreciate
people have taken the time to offer.
But not to be a pest but what are the odds
the IANA would ever allocate the 1 and 100
nets to someone? Is this an unpredictable
matter or is there a schedule of what's
next somewhere? Or which is more likely, the
world adopts IP v6 or the 1 and 100 nets
are deployed on the internet? 
It is
apparent that I really want to use these
address ranges but I do need to grapple
with the possibility that this lab will
need internet connectivity at some point.
Given that unallocated class A address space represents one of the biggest
chunks of remaining address space fairly likely...
you'll notice that 60/8 was assigned in april 03 to apnic, lacnic was
assigned 2 /8s in the last year and so forth...
> But not to be a pest but what are the odds
> the IANA would ever allocate the 1 and 100
> nets to someone?
99%
I can't imagine 100.0.0.0/8 remaining reserved - there's nothing
particularly special about it (100=0x64... a number which represented
in hex has digits which form a power of two in decimal, looks
nice but isn't a special bit boundary or anything).
1.0.0.0/8, well, IMO some chance it may remain reserved for quite a
while. But there's always a chance it could be allocated any day.
As another example, I'd be sure 200.200.200.200 will end up with someone
someday, and I've seen horrible attrocities committed with that IP
by people who don't own it (eg. used as a content destination IP via
satellite that a number of providers then had machines with that IP
receiving UDP), just because they think it looks nice.
1.0.0.0/8 was used by One.Net (an Australian ISP/Telco, who later
collapsed rather dramatically) for their router/link IP addresses. It
was disgusting enough that many wouldn't peer with them.
I don't know what isn't clear about using the allocated network for
internal addresses:
10.0.0.0/16
10.10.0.0/16
10.20.0.0/16
etc
Nice, clear, obviously differentiated blocks. 10.0.0.0/8 is BIG.
You're unlikely to need more than 254 devices on any subnet in a
lab anyway, so you can split down to /24's (or smaller in these
enlightened times of CIDR, but I'm guessing that doesn't look nice
to you).
If you can't find enough nice IP addresses in it to build your lab,
well, that's a really big lab.
David.