But isn't that the exact point? That it would start to be a hindrance to merchants so they'd stop reusing addresses?
Just thought of something... how does the BIP32 protocol addendum limit the usable address space based on whatever seed is generating the string of future keys? Reading the wiki pages now, will edit this post if i find an answer.
What do you mean limit the usable address space? There is no change in the number of possible addresses.
For some reason I was thinking that using a seed->prng, and sequential nonce would end up with a subset of the 2^160 (#?) key pairs.
If the entire world economy used a new key pair for every digital transaction (the rate of which is increasing exponentially) for the next 50 years... where would we end up in terms of collision probability? I dunno, i think i'm way off topic here. We can let this sidebar fall away now...