Before we wander too far down this path, there was a very interesting paper at the last sigcom in which the authors managed to shrink the size of the then current routing table (~40k routes) into less than 200K of memory. In short, I differ from Mike in that my values for "believed to be viable" differ, apparently wildly, from his and brians. I'm unconvinced that this will remain a true, long term technological argument. I'd like to see something besides, "too hard with 1990's technologies". "Long time" ought to have a better spec than say, Internet Dog Years?
brian's reason is exactly the goal which was in mind:
to bound the maximum complexity of the default-free region at values believed to be viable with some margin.
Like everything else, 13 is an engineering compromise - chosen to balance one set of considerations against a bunch of others, and after ruminating over it a long time, the consensus was 13 was the best choice.
-- --bill