These are chat archives for ethersphere/orange-lounge

4th
Sep 2017
lash
@nolash
Sep 04 2017 12:08
@nagydani did you take a look at the readme yet?
Daniel A. Nagy
@nagydani
Sep 04 2017 13:55
Hi @nolash ,
Not yet. Let me take a look now.
Viktor Trón
@zelig
Sep 04 2017 15:27
@nagydani can we have a call sometime?
holisticode
@holisticode
Sep 04 2017 15:57
@zelig can I do anything else for the errorpages PR? Not merged yet, travis test failing look like related to go 1.9.0 and not related to the PR itself
also have a couple of questions regarding "Generalised swap swear and swindle games"
  1. "Disconnect threshold"
creating an account is cheap, so one could just always only leach and, if reaching the disconnect threshold, simply drop account and create a new one
the issue I see is less about the leacher himself, but rather for the peer in plus , which would provide service for free and loose the credit
holisticode
@holisticode
Sep 04 2017 16:02
  1. Waivers: "Swap index"
What exactly is the "swap index" and when does it get updated?
If peer B issues a waiver, with a specified swap index, but "normal" swap accounting continues if peers continue to operate in the meanwhile as specified in 2.2., doesn't the swap index get updated during this "normal" swap accounting? Doesn't look like service exchange gets "frozen" until all waivers and cheques are cleared?
holisticode
@holisticode
Sep 04 2017 16:11
  1. chequebook solvency
we know from conventional banking there exist special accounts with withdrawal restrictions
what about the option of creating a chequebook where funds can't be "freely moved", also signaling to peers a higher level of solvency trust (if this makes sense at all)
Viktor Trón
@zelig
Sep 04 2017 16:13
swap index is incremented when the peer doing waiving consumes more than it could waive and needs to pay. Then it issues a cheque with a new swap index to close the cheque-waiver cycle
holisticode
@holisticode
Sep 04 2017 16:14
aaah so swap index isn't a measure of "normal" swap accounting, it only kicks in during cheque-waiver cycle
Viktor Trón
@zelig
Sep 04 2017 16:15

what about the option of creating a chequebook where funds can't be "freely moved", also signaling to peers a higher level of solvency trust (if this makes sense at all)

yes, that is easy to do in the contract, when you do assign part of your balance to a specific peer (channel deposit), i claim you implemented a payment channel. Thats clear. Now the exciting bit is when you can do it without touching the blockhain

ie soft channel deposit allocation
holisticode
@holisticode
Sep 04 2017 16:16
right
Viktor Trón
@zelig
Sep 04 2017 16:17
yes indeed, in fact we should call it 'swap cycle index' maybe
holisticode
@holisticode
Sep 04 2017 16:19
:+1:
Viktor Trón
@zelig
Sep 04 2017 16:19
or I am open to suggestions
holisticode
@holisticode
Sep 04 2017 16:20
ability to provide services and earn ETH that way is bad-ass
Viktor Trón
@zelig
Sep 04 2017 16:28
creating an account is cheap, so one could just always only leach and, if reaching the disconnect threshold, simply drop account and create a new one
the issue I see is less about the leacher himself, but rather for the peer in plus , which would provide service for free and loose the credit
Nick Savers
@nicksavers
Sep 04 2017 16:32
Did you have to put B left and A to the right in that diagram? :)
holisticode
@holisticode
Sep 04 2017 16:34
I think a word about where promissory notes are stored would be great (or am I missing something).
Intuitively an issuer issues them and are then stored on the beneficiary's side
are they denominated in wei? because they might well become tradeable as well...
Viktor Trón
@zelig
Sep 04 2017 16:36
to leech, you need to have a chequbook. although you can leech more than your balance, if your strategy is to set one up, leech over your disconnect threshold and let the bank run kill your contract you gained d*n-c where d is the disconnect threshold amount, n is the number of peers you managed to fool and c is the cost of contract creation for chequebook.
holisticode
@holisticode
Sep 04 2017 16:38
ok so amout of c is crucial
Viktor Trón
@zelig
Sep 04 2017 16:57
@holisticode active notes are stored by both parties
Viktor Trón
@zelig
Sep 04 2017 17:03
@nicksavers thanks i will swap A and B there :)
holisticode
@holisticode
Sep 04 2017 17:07
@zelig are these notes tradeable?
Viktor Trón
@zelig
Sep 04 2017 17:10
what would that mean? and which one in particular?
holisticode
@holisticode
Sep 04 2017 17:15
well bonds are potentially tradeable - although the paper specifically mentions a bond has a beneficiary
Viktor Trón
@zelig
Sep 04 2017 17:16
they are meaningful only for the parties involved, unlikely that any c would covet a cheque A gave to B
holisticode
@holisticode
Sep 04 2017 17:17
cheque is backed by future promise of payment - so it could be interesting for B to give it to C if B has any business (obligation) with C
Viktor Trón
@zelig
Sep 04 2017 17:17
the bounty is clossest to something of interest outside the peeer context
holisticode
@holisticode
Sep 04 2017 17:18
in any case, this wouldn't be anything which looks like an immediate need
final question: have you had any thought on swarm tokens as discussed a while ago?
Viktor Trón
@zelig
Sep 04 2017 17:19
not sure its needed once write upsection 5 :)
holisticode
@holisticode
Sep 04 2017 17:20
ah right :+1:
Viktor Trón
@zelig
Sep 04 2017 17:21
atm i dont see it solve anything but my view is limited i can easily be convinced
did you get to section 4 yet?
holisticode
@holisticode
Sep 04 2017 17:22
yes
Viktor Trón
@zelig
Sep 04 2017 17:23
thats whwere explanation became a little sloppier and got no good diagrams
or even ideas for good diagrams
holisticode
@holisticode
Sep 04 2017 17:24
isn't figure 6 related to section 4?
looks good to me
Viktor Trón
@zelig
Sep 04 2017 17:25
yes
oh i meant section 5
holisticode
@holisticode
Sep 04 2017 17:25
aaah
yes that's the sec I need to re-read better
Viktor Trón
@zelig
Sep 04 2017 17:25
and above re swarm token i meant section 6
holisticode
@holisticode
Sep 04 2017 17:26
:) makes sense
what's the plan for implementation?
Viktor Trón
@zelig
Sep 04 2017 17:27
feel free to edit / fix /comment via PR
well,first i need to convince people (inculding me) that this all makes sense :)
holisticode
@holisticode
Sep 04 2017 17:28
To me it does
what we need to pay attention to is that, as with any economically incentivised system, it attracts bad behavior
Viktor Trón
@zelig
Sep 04 2017 17:29
the relatively uncontroversial section 4 bit has a preliminary implemetation here: https://github.com/ethersphere/swap-swear-and-swindle by @orenyodfat
holisticode
@holisticode
Sep 04 2017 17:29
i.e. gaming the system
Viktor Trón
@zelig
Sep 04 2017 17:29
game it, attack it, pull it apart
if you find attack vectors i need to address them
holisticode
@holisticode
Sep 04 2017 17:31
would be good though if swarm's "main" service -decentralized storage - isn't affected though
so I think we need to proceed in stages
  1. "normal" swap accounting
  2. "uncontroversial" insurance and litigation
  3. everything else (in possibly "guarded" environments until proven)
without 1 and 2, swarm isn't complete anyway
also, I wonder how much support we need from ethereum itself
Viktor Trón
@zelig
Sep 04 2017 17:34
well yes, but for smooth insurance and litigation I think you need all of section 5 and for proper pricing probably 6 too....
what do you mean?
from ethereum itself?
holisticode
@holisticode
Sep 04 2017 17:36
well if we start offering parallel payment channels, can it somehow have consequences to the main blockchain? to ETH value?
what are the risks?
Viktor Trón
@zelig
Sep 04 2017 17:39
well, my naive thinking is that any tech that brings secure scalability improvements will be attractive driving eth up.
the risk is if the tech (or the model) is flawed and we figure only after major damage...
lets be good engineers and try our best to mitigate that
holisticode
@holisticode
Sep 04 2017 17:40
:+1:
Viktor Trón
@zelig
Sep 04 2017 17:40
ie. network testing and simulation, testnet, monitoring and your idea of staged delivery
and before all. clear writeup, peer review, .... spec
holisticode
@holisticode
Sep 04 2017 17:42
great :+1: :+1:
holisticode
@holisticode
Sep 04 2017 17:58
@lmars I guess a callback into the mocker would be great when stopping a network?
Currently the mocker (essentially a loop) just keeps running after stopping the network
Viktor Trón
@zelig
Sep 04 2017 18:19
mocker is retired in the PR-ed code btw
holisticode
@holisticode
Sep 04 2017 18:19
you mean eliminated?
has the overlay sim been adapted?
Viktor Trón
@zelig
Sep 04 2017 18:20
@lmars i got a promise from @fjl for reviewing the simulation PR
in fact the network-testing-framework is not in sync with the simulation code in the PR
I wonder if we should rebase network-testing-framework on it...
holisticode
@holisticode
Sep 04 2017 18:21
so latest up-to-date code is in PR?
Lewis Marshall
@lmars
Sep 04 2017 18:21
yep my plan was to rebase once p2p-simulations is merged
Viktor Trón
@zelig
Sep 04 2017 18:22
y... sigh..
holisticode
@holisticode
Sep 04 2017 18:22
ok has anyone tried the overlay sim with the new PR?
or is that one obsolete now
Lewis Marshall
@lmars
Sep 04 2017 18:23
there is no need, the PR just contains a subset of the changes which are on network-testing-framework to avoid trying to PR everything at once
Viktor Trón
@zelig
Sep 04 2017 18:23
i guess it wont work cos it needs the mocker
holisticode
@holisticode
Sep 04 2017 18:24
ok question is: should overlay sim be refactored?
Lewis Marshall
@lmars
Sep 04 2017 18:24
we should just hold off making changes to simulations code until the PR is merged
holisticode
@holisticode
Sep 04 2017 18:24
and is overlay still a good backend to use for frontend?
Lewis Marshall
@lmars
Sep 04 2017 18:24
I'm not sure what you are asking?
holisticode
@holisticode
Sep 04 2017 18:25
i guess it wont work cos it needs the mocker
Lewis Marshall
@lmars
Sep 04 2017 18:25
you should just continue using the network-testing-framework branch because that has the new swarm stuff too
p2p-simulations is just the p2p/simulations package
holisticode
@holisticode
Sep 04 2017 18:25
right, but
Viktor Trón
@zelig
Sep 04 2017 18:25
sure but even within that package there are hanges
holisticode
@holisticode
Sep 04 2017 18:25
mocker is retired in the PR-ed code btw
Lewis Marshall
@lmars
Sep 04 2017 18:26
well, mocker just isn't included in the PR, it is still on the network-testing-framework branch
holisticode
@holisticode
Sep 04 2017 18:27
ok so it's kind of deprecated, coz after PR merge we rebase network-testing-framework
Lewis Marshall
@lmars
Sep 04 2017 18:27
well we decided a long time ago that mocker functionality should be client side using the API
we haven't done that yet though
holisticode
@holisticode
Sep 04 2017 18:27
that's why I ask:
Lewis Marshall
@lmars
Sep 04 2017 18:28
so either we'll just merge the mocker work or move it to the client
holisticode
@holisticode
Sep 04 2017 18:28
ok question is: should overlay sim be refactored?
and
Viktor Trón
@zelig
Sep 04 2017 18:28
yes it better be rewritten in the client
holisticode
@holisticode
Sep 04 2017 18:28
and is overlay still a good backend to use for frontend?
Viktor Trón
@zelig
Sep 04 2017 18:28
yes
holisticode
@holisticode
Sep 04 2017 18:28
:+1:
I'll try refactoring overlay to be rewritten in the client
I guess the discovery test is a good example?
Viktor Trón
@zelig
Sep 04 2017 18:29
btw i am still struggling with the discovery test as we speak ... :(
Lewis Marshall
@lmars
Sep 04 2017 18:29
it is only the mocking that we decided to move
the discovery_test doesn't use a mocker
well, I don't think it does :)