Transactions can be rejected because they take too much time to process (misbehaving transactions) or because there are too many of them for amout of EOS held.
Each transaction has to complete within allowed window, or it will be rejected by block producers. This will keep bad code out of the blockchain.
OK, but the bandwidth limit has the same purpose (while distributing resources proportional to stake) . Why do we need both?
Transactions can be rejected because they take too much time to process (misbehaving transactions) or because there are too many of them for amout of EOS held.
Each transaction has to complete within allowed window, or it will be rejected by block producers. This will keep bad code out of the blockchain.
ok, but why not just take a global time limit for processing according to stake? Maybe it is harder to implement
Bad transactions can cause problems for blockchain operation, so have to be rejected regardless if there is unused bandwidth left.