Some numbers:
Assume it takes on average 30 seconds to verify 1 MB of typical transactional data (k =0.5 min / MB). Since T = 10 min, this means the maximum average blocksize (network capacity) is limited to:
Seffective = T / (4 k) = (10 min) / (4 x 0.5 min / MB)
= 5 MB.
QED. We've shown that there exists a limit on the maximum value of the average blocksize, due to the time it takes to verify a block, irrespective of any protocol enforced limits.
Great work Peter, but do we have any empirical evidence for the 30 seconds? Seems surprisingly high and I would have guessed just a few seconds.
No, I just made it up. I think I'll change it to 15 second, as I agree it's probably too high. I guess we could estimate it by looking at the ratio of empty blocks to non-empty blocks produced by F2Pool.
If someone wants to tabulate that data, I'll update my post.