I think the negative responders to this sort of proposal are largely failing to recognize that such a system would be:
a) highly automated
b) smarter than they seem to give it credit (e.g various levels of 'taxation' at popular brick and mortar government-audited points)
c) viral
Alternatively - they do understand this and they're resisting it in perhaps the only way that *might* work: overwhelming popular rejection/distaste + playing down the practicality of it.
I asked this question on stack exchange a while back:
Is there any way the Bitcoin network could resist a viral tainted-coin tagging system implemented by regulators?
http://bitcoin.stackexchange.com/questions/2119/is-there-any-way-the-bitcoin-network-could-resist-a-viral-tainted-coin-tagging-sThe answer seems to be, technically no, politically/socially maybe.
I feel confident that some government somewhere will try it if Bitcoin ever goes mainstream.
If they are 'heavy handed' in applying taint - then the more tainted coins (and the higher taxed they are), the larger the split in fungibility.
If they produce a massive split in fungibility this way - all they succeed in doing is creating a subset of bitcoins which are largely used for blackmarket transactions.
If you end up with a chunk of tainted coin.. you then have an incentive to spend it on 'blackmarket' goods and services (rather than reporting the source or paying the audit-point taxes to spend it cleanly) - increasing the size of the blackmarket economy the taint system is supposedly there to control!
Because of this - I think that although such taint systems are inevitable, they'll be largely self-limiting and only effective if applied judiciously to the worst of the worst.
Most of the complaints here seem to be of the sort: "I don't like it!!!"
Well duh.. but can we move beyond that and look at what might happen whether you like it or not!?
I'm not convinced that it's as big a problem as some think (due to self-defeating effects if too heavily applied).. and I'm also not convinced that just naysaying it will stop it from popping up. Better to model a light-touch vs heavy-touch taint system and try to get a feel for how it might play out.