Bisq excited me enough to do some quick research about the current status of Bisq, and I sent some questions to Manfred on Slack. He nicely responses with detailed answers. I think these answers would be helpful for others who have the same questions, so I share them here to get more exposures and perhaps it would help generate new ideas around this topic from the community.
Questions from me
The Bisq client currently looks quite heavy for normal users without technical backgrounds especially for the ones with limited knowledge of how it works for exchange and Bisq’s arbitrator model. Also, it is a client only works on the desktop. Considering the most Internet users are on mobile devices, with the various censorship environments in different countries, I think it is hard for Bisq to penetrate the market quickly to grow a sustainable user base for trading economic cycle. The issue is that the trading volumes on Bisq are extremely low, in my view, compared to the centralized exchanges. First-time users spent some time to figure out how it works, but realized there were few trading offers on there and give up eventually.
I am fascinated by the idea of decentralized exchange so I can transfer my values with minimum cost/time span/risks, and I believed it is highly demanded in the market. But due to the usability of the Bisq software, I think it is hard to make it go into the mainstream and persuade users to switch from centralized exchanges.
Nowadays most of the users are using the mobile phone to access the internet. The Bisq client seems only support to be installed on the desktops. So I think if Bisq can be running on the mobile phones, it would awesome and probably could boost the users base that further grows the community to a trading feasible state.
To resolve these adoption issues to bootstrap the user base, I am thinking if the following solutions would work for Bisq with my limited knowledge:
- Prompts users message about the Internet censorship issues, and provide options for them to configure a network proxy to get around their local firewall restriction.
- Port the client as the mobile app to boost the usability and user adoption.
- Allow the users to get offline after they post the trade offers, and get notifications when there are updates on the offer so that the makers don’t need to wait for a long time in front of their desktop just for takers to view their offers. Also, this would increase the offer volumes. From my understanding, the maker has to have the client on to keep the offers visible to others, and they can’t just exit the client until the trade is completed.
In terms of upcoming technologies for the decentralized exchange, how do you see atomic swap, which seems to be an efficient mechanism for decentralized exchange? Will it be a feature on Bisq so that there is no need for arbitrators for altcoin exchanges? So far, the most interesting and valuable part of Bisq, from my understanding, is its method of decentralized exchange for fiat-cryptocurrency.
I had a look at the DAO whitepaper, the trading fee looks like the engine of the whole economic cycle for the Bisq token. There are few things I still not quite sure:
- When Bisq has a market price, does that mean the trading fee will be based on the market price rather than a fixed number of Bisq?
- If Bisq is used for a discount trading fee, does that mean the arbitrators would receive discounted/fewer returns on the same effort?
- With the upcoming technologies to promote efficient cryptocurrency exchange trading with nearly zero cost/no need arbitrator(correct me if I am wrong), like atomic swap, what would be the advantages for Bisq with a trading fee?
Answers from Manfred
re GFW:
yes due the GFW tor is blocked by default.
we have plans to support plugable transports (meek, obfusc4). a new version of our tor library (supporting plugable transports) is actually ready for integration but would require some good testing and a bit of integration work. with that it should work by default to get over the GFW. but we have atm other priorities though any dev could work on that…
we would also need chinese translations and a UI dev for fixing layout issues due to longer translation strings. and for china we would need the decentralized arbitration system ready because i think english might be a problem for some traders there. atm we cannot open the arbitration system to anyone outside the core team because that would be too insecure. so our prio is to get that decentralized arbitration system done first. that requires the full dao as well so those 2 parts are our main prio atm.
so from our side expanding to china carries too much dependencies and effort for now. but once the full dao and arbitration system is in place we have definitely plans to do that. if anyone else wants to work on parts of that (translations, layout fixes,…) it would be very welcome of course!
re maker keep running app:
yes the maker need to have running their app but he does not have to be active (the take offer happens by the app itself - the app is playing the role of a server for the taker). once a trade starts (a peer takes your offer) you dont need to keep the app online but check at least once a day for trade messages.
re mainstream adoption:
i am aware of the usability drawbacks compared to mobile or web apps. but to have a real decentralized p2p app comes with some limitations. theoretically it could work on a mobile (beside the issue for makers to stay online - that would not be feasible with mobile due battery drain) - but the resource requirements (bitcoinj, p2p network, tor) might be a bit tough, but more problematic is the engineering effort and limitation on android. java8 should be already working on modern fast phones but on ios there is no java support. the effort for getting such a android-only, takers-only, fast-modern-devices-only mobile app is probably about 6-12 months. we dont have the resources for that.
to have a second alternative implementation would make deveopment much more complicate as well (one change on one app must not break the other app).
a better approach IMO with less effort would be to use the apis (we are working on a trade api) and use a light “remote control” mobile client. that could be even html based so we would get a web app as well. drawback with that solution is that it requires a server where the real app is running and most mainstream users are not able/willing to do that. it would require a service provider who provides the hosting/setup for very small costs and do it in a trustless way (it contains the wallet so you dont want the service provider to have access). there are some rough ideas how that might work but thas kind of a research project.
so you see it is not that easy. breaking the decentralisation aspects would break the whole system. bisq only makes sense if it is really censorship resistent. so i think for the near future it will be limited to the hard core bitcoin users who are able and willing to take a bit of extra effort to get privay and censorship resistent.
compare it to running a btc full node (bitcoin core) - it requires some extra effort and is not that suitable for the mainstream, but u dont get the same level of privacy and censorship resistent without it. it will never be a mass market but is the base for a system on which others can build mass market products which have lower standards on privayc/censorship resistent. for bisq that might happen once the apis are in place. then people can build semi-centralized web or mobile apps and use bisq as backbone. we dont endores that as it does not carry our core values (privacy, censorship resistent) but as we cannot prevent it and as it has the advantage to add liquidity to bisq (also for those who run the real p2p app) it is something i want to point out as an alternative path for those who aim to get more into the mainstream market.
if you have other ideas please let me know!
re atomic swap:
we have rough plans to support that in future, however the engineering effort is rather high and it will be probably only feasible of some main currency pairs.
i think once the apis are out and you can do a fully automated altcoin trade the cost/benefit balance is much more on the current model with arbitrators. for altcoins there is very low risk as there is no room for real disputes (the altcoin blockchain is the truth).
yes, i agree the fiat part is the core feature of bisq and that is the part where no other project is competing (decentralized/privacy protecting).
re dao:
-
yes the trade fee in BSQ will be based on the BSQ market price, which will be a bit problematic in the beginning if volatility is very high.
-
the trade fee in BSQ will not be paid to arbitrators but will get “burned” (un-colored) - thus distributed to all stakeholders as we reduce total amount of BSQ by burning, makeing each BSQ a bit more valuable. arbitrators are contributors and get paid via compensation requests.
-
i think to get to that state it will take several more years if it ever really happens. if there are many trustless exchanges which works faster/better/cheaper then bisq will fail if it has not adopted to those new approaches. but after 4 years in the space i have not seen much advances. the idea for atomic swaps is out since many years (was not 100% safe without malleability fix due segwit) and nobody took the opportunity to work on it (mercury was the only project which was in production but server absed). i think it is over-hyped atm and the problems (engineering, resource requirements) are not well recognized. i was diving once deepter into that and was considering to use that instead of the arbitration model and do crypto-crypto only, but i concluded that in a pure decentralized way its not very feasible. maybe for 2,3 currencies but not 10 or 100. u need to implement and run the infrastructure (like bitcoinj) for each coin, that costs a lot of effort and resources. using rpc interfaces and outsource that to the user to run a full node for each altcoin is a usability killer. on a server thats not such a big deal but on a p2p app it is. so for a server model it can work but do you gain much regarding privacy/trustlessness compared to a much more simple centralized exchange model?