I assume the reason behind this is to block people from encoding data into the block chain by doing thousands of worthless transactions. If I recall correctly someone recently encoded the URL to a child porn site into the block chain, this is the only reasonable response to prevent it from happening again IMO.
Encoding data into the blockchain isn't a bug--it's a feature (perhaps THE feature). When you want distributed consensus (e.g. for free-as-in-speech domain names), it's the only game in town. People should be able to put whatever data they want in the system, assuming they cover the cost.
Well, miners can ignore transactions that have small fees or no fees. If there were not enough space in a block for all outstanding transactions, the transactions with the least fees will just be ignored.
So, there is an economic mechanism already in place to make sure dust transactions are paid for.
So that's not the problem that's being addressed.
EDIT: Well, actually, it's a "tragedy of the commons" situation. Having tons of dust transactions is seen as bad for Bitcoin (for reasons that are not yet 100% clear to me). So the devs think there is a need to dis-incentivize that. But the economic mechanism I explained above has not yet kicked in, so right now, you can make tons of dust transactions without having to directly pay for it. The burden of storing those small outputs for all time then falls on everybody who uses bitcoin.
"If you have a better suggestion for fixing the problem of new users wasting lots of time gathering tiny drips and drabs of bitcoins, and then getting upset when they can't spend them (because it costs more in fees that they are worth), I'm open to suggestions."
This type of spam is solved through minimum fees for transacting very small amounts of bitcoin.
If a user is willing to spend $10 to insert a URL in the block chain, there is no way to stop them without making bitcoin useless.
The move to treat very small amounts of bitcoin as simply invalid is intended to reduce the bloat of the unspent address cache (called the UTXO). This is explained at the top of the linked page.
It's more to do with certain services (SatoshiDice) sending very small, useless transactions through the network (SatoshiDice is a gambling service, they send a small transaction to tell you you lost). All those small transactions put a burden on the network.
Regarding the CP URL, it was actually several hundreds of links in a large FAQ, and I don't think this changes anything regarding encoding data in the block chain.
From what I read on bitcointalk I believe this applies to miners too (if they use bitcoind). However it would be possible to run a mining pool which would accept small txs with no fee, and clients would still accept them as valid blocks.
I don't get this problem. It's not like people literally actively read the BlockChain.
(Also, IIRC, wasn't their a proposal to make the data a hash or something mandatory so that it couldn't be random|misc-manual data? Or was that unfeasible for some reason?)
