BGP table prefix growth

route-views data does not confirm conjecture that more specifcs are currently driving up the BGP table sizes
indeed, for 1.3 years since aug 2001, standalone prefixes grow faster than more specifics (standalone 38%, more spec 32%)
currently at parity
more specifics do churn (appear/disappear) faster
also convert to standalones
but overall change in more specifics not statistically significant

increase in # of top prefixes is caused by:
coverage of new address space 		55%
deaggregation/contraction of existing prefixes 37.5%
expansion of existing prefixes	                5%
aggregation				2.5% 

but growth was in top prefixes, not more specifics