I don't see how this is entirely feasible unless we are talking like changing the units to mBTC.
Now, we have just <21,000,000,000 mBTC that will potentially be available.
Not sure which prior post you're referring to, but all of these proposed changes are technically feasible (if critical mass of term-usage is achieved). No matter what units are chosen for human use, interfaces, and integration with other systems – the core protocol/system always works on integral satoshis. All of the following tallies are identical, and represented the same in the binary protocol and canonical data structures:
21 million BTC
= 21 billion mBTC/"millibitcoin"
= 21 trillion µBTC/"microbitcoin"
= 210 trillion satoshis
Working up from a zib=µBTC unit instead, again all these can be used interchangably without really changing the atomic units of the system:
210 trillion satoshis/cZBC/"centizib"/"cents"
= 21 trillion ZBC/"zib" (aka µBTC)
= 21 billion kZBC/kƵ/"kilozib" (aka mBTC)
= 21 million MZBC/MƵ/"megazib" (aka BTC)
For completeness, you can also think about units that exist but are so large they'd rarely be useful...
210 trillion satoshis
= 21 thousand GZBC/GƵ/"gigazib" (aka kBTC/"kilobitcoin")
= 21 TZBC/TƵ/"terazib" (aka MBTC/"megabitcoin")
...as in, "Some estimate that Satoshi himself may own around 1 megabitcoin (1 MBTC)", or "Mt. Gox misplaced 850 kilobitcoin (850 kBTC)".
Since people easily understand large numbers, especially when rarely used, it's far more likely to emphasize the unique size of these numbers by using the smaller units:
"Some estimate that Satoshi himself may own around 1 million bitcoin" or "...1 trillion zib".
"MtGox misplaced 850 thousand bitcoin" or "... 850 billion zib".