Total received that excludes change sent from the same address

[lirazsiri]

This was reported on reddit when I asked for feedback on blockexplorer.com, which is based on the latest version of Insight.

Total received on Insight is 59.8951 BTC:

https://insight.bitpay.com/address/19SokJG7fgk8iTjemJ2obfMj14FM16nqzj

On all the other non-Insight block explorers it's 0.4999 BTC:

https://blockchain.info/address/19SokJG7fgk8iTjemJ2obfMj14FM16nqzj
https://www.biteasy.com/blockchain/addresses/19SokJG7fgk8iTjemJ2obfMj14FM16nqzj
https://blocktrail.com/BTC/address/19SokJG7fgk8iTjemJ2obfMj14FM16nqzj

[bendelo]

A fix was submitted for this bug but the pull request was closed without discussion:
bitpay/bitcore-node#376

[dabura667]

imo, the issue of which output is change should not be guesswork.

If we only calculate change for transactions that are done through Copay or something where we know what the change is:

1. We would be lowering the user's privacy
2. Any transaction not done by any wallet using insight in the backend would just be an educated guess.

I am more surprised that other block explorers have followed suit with blockchain.info (not exactly a shining beacon of light in the bitcoin community when it comes to sound development decisions) and decided to calculate sent and receive based off of guessing.

I agree that a good compromise would be to ADD new attributes called "receivedChange" and "spentChange" and let the user decide.

however, I can not see any real world use for this... as any meaningful use of "change" data would be easily cached on the application side.

If you'd like to start reporting to insight with signed reports of "HEY THIS IS MY CHANGE KTHXBYE" then it would be a different story.

[bendelo]

If a transaction sends funds from A to B and C, then I agree, we should not try and guess which is change.

If a transaction sends funds from A to B and A, then we should not count the change sent to A in the received total for A. This is what every other block explorer does, Insight is standing alone.