Or one can simply allow Moore's law to take care of the bloat problem.
If computation was the barrier, that is not certain because Moore's law may be ending...
The size issues are bandwidth and storage, not computation. The law is not broadly applicable to bandwidth, and scaling will not end with the end of process shrink. Moore's law will apply to storage in the near term, however, as it will be essentially determined by flash densities. It appears to have enough headroom left to fully accommodate XMR.
I use the neutral term "size" instead of "bloat" because bloat implies the content does not add value, which I consider misleading.
Bandwidth scales slower than Moore's law but is not in danger of ending:
http://www.nngroup.com/articles/law-of-bandwidth/Neither can scale fast enough so XMR to reach 10 million users in 3 years, because XMR would need to scale 10^3 in 3 years.
And there is a computation issue on verifying transactions if you want to scale up micro-transactions volume (e.g. 100,000 txs per sec):
https://bitcointalk.org/index.php?topic=103172.0You appear to lack domain-specific knowledge that I have because I am in the trenches of development and you are not. Practice and application trumps theory.