Quote from: David Rabahy on March 09, 2016, 04:57:53 PM

Quote from: adamstgBit on March 09, 2016, 04:18:03 PM

Quote from: David Rabahy on March 09, 2016, 04:04:11 PM

To scale we must either;

1) have fast enough hardware
2) improve the algorithm, i.e. verify without pairing
3) verify in the background (although the verification has to happen sometime)
4) constrain the sum of sigops/transaction in a block

Can someone please provide an example of transaction that requires two or more sigops?

this is a none issue
#1 #2 is has already been done

it's not like these crazy computational intensive TX are legit, you only get into trouble if you try to allow this crazy spam like TX with thousands of inputs.

miners are allowed to orphen a block for ANY reason. I think it's perfectly valid to not allow spam TX designed to slow down validation time of the block.

1) Oh? What speed hardware can handle what number of sigops? With quadratic growth eventually some number of sigops will exceed the capacity of any conceivable hardware.
2) Please help me understand; I am not aware of this development yet.

So, each input comes with a sigop; I see, actually that makes perfect sense. Hmm, would it make any sense to create a number of smaller transactions that consolidate the many inputs? Perhaps it would be possible/better to avoid creating addresses with small amounts in the first place. Just trying my best to understand. I appreciate your patience and efforts.

i dont know the details...

I assume hardware is always getting better because HELLO! and libsecp256k1 offers 5x validation speeds.