Without (I hope) straying into the depths of an Abusenet "*.advocacy"
type posting, here's my quick opinion:

How many people there are who are willing to port may not be such a
huge issue.  Remember Fred Brooks (I think) citing an old saying:
You can't get a baby in 1 month by getting 9 women pregnant.

If the concern is for extant ports, then you should start by
asking which systems you really care about, and then see not
only how many of them are supported, but how well they are
supported.

If the concern is about *portability* (presumably because these
hypothetical governments have as-yet unmade hardware and want
a UNIX-like OS on it ASAP), then neither the number of old ports
nor the number of porters is really a direct answer.

Counting ports also seems an exercise a little like counting clouds
that have drifted into and through one another.  Should AMD64 and
i386 be distinctly counted?  How about Amigas and Atari STs?
What's the criteria?  A NetBSD page claims that NetBSD/Alpha supports
more Alpha hardware than any other (free?) OS.  Is that still true?
Could it be stretched to be called multiple ports if you squint?

Nevermind the points already raised about single source trees, and
whether the systems are actively supported or are purely historic.

I think that this exercise is futile without a lot of details
made clear up-front.  And then you run into the debate of whether
the details render the answers meaningless or have been picked in
order to favor an up-front bias.


Basically, "Number of ports" is a benchmark, and without a bunch
of picky rules, it's more like saying "how many primes per second
can you generate" than "how far can you get in the Sieve in one
second".  Whatever value the latter may have, the former is rather
pointless.


-- 
  "I probably don't know what I'm talking about."  http://www.olib.org/~rkr/