I looked at the trunk code for gawk on Savannah when I wrote the above. It passes the output of time(0) straight into srand().
The size of the dictionary doesn't matter (given that it's more than about 70 words); the limiting factor is the entropy in the RNG seeding.
I have another version that I use that stuffs srand but in the end I figured srand from a 250k dictionary is still better than picking words out of your head from a ~1k dictionary