i wonder how much damage fake inputs would do to heuristics

gingeropolous: you mean generating extra zero-value outputs in transactions?

This idea has come up from time to time, for sure

Looks like the PoPETs page limits are much better for CLSAG

and I could submit there without having to remove anything

Next deadline isn't until August though :/

I should be able to post a revision on IACR later today

Should there be some kind of posting about that?

@gingeropolous HUSH does this IIRC

There's an important ratio to consider if implementing this in a decoy-based system

You could define it a few ways (per txn, per unit time, etc) but the gist is ratio of # outputs created to # outputs showing up in ring signatures

Also depends on what heuristics you care about

Updated CLSAG preprint: eprint.iacr.org/2019/654

(same link as before)

TeX source also updated: SarangNoether/skunkworks 372ff1b

sarang, not even generating zero value outputs, but making a "ring signature" for an input where 0/11 are the real input

Not much of a signature anymore =p

Since you no longer need to know any private keys

And you need to private key to generate the key image, or the check will fail, and your ring will be known to be fake.

How about sending 0 amount outputs back to yourself instead of to a random address and use those (if such an amount can be "spent") ?

Or, it that's feasible, use such an output in sweep_single ?

(As a second ring)

if there's a fee paid somehow, seems legit

sending more than one output to yourself at a time can be risky

since loops might be created when you spend them

yeah

Hmm that's fascinating. So the input amount from an entire ring could be 0

And zero fee txns are allowed (just not propagated)

So could we theoretically have a 1-in/1-out transaction where the input is a dummy output with 0 XMR, the fee is 0, and the output is empty? (a miner would have to put it in the block themselves)

That would balance

pointless but interesting

^ the motto of a lot of mathematics Isthmus =p