Sender | Message | Time |
---|---|---|
25 Nov 2023 | ||
🍠kbrav | so for example, have a dpath, you know the dpath will get you to the correct zone, and let's say you really trust the zone owner for whatever reason, or you trusted them when they set LOCK == 1. that's how you know the data is good | 01:50:49 |
ddapp | In reply to @kbrav:matrix.orgah i see, i think i have the same understanding as you do ie locked dpath is just an immutable alias for the file, dpath doesn't try to provide any more assurance if someone wants any assurance about the file, someone has to verify the file is not harmful | 01:55:13 |
ddapp | In reply to @kbrav:matrix.org* ah i see, i think i have the same understanding as you do ie locked dpath is just an immutable alias for the file, dpath doesn't try to provide any more assurance if someone wants any assurance about the file, someone has to verify the file is not harmful (by trusting zone owner, manually verify the file etc) | 01:55:56 |
ddapp | In reply to @kbrav:matrix.org
wdyt? : )
| 02:03:52 |
ddapp | In reply to @kbrav:matrix.org*
wdyt? : )
| 02:05:15 |
ddapp | *
wdyt? : )
| 02:07:04 |
ddapp | *
wdyt? : )
| 02:08:21 |
ddapp | * am thinking it's nice zone owner can change their pricing strategy by taking in the plaintext name and block name registrations (eg revert based on length), and avoid adding proxy (to me it's a little complex); code below totally agree with you rootzone doesn't need to change, the rootzone governor can manually grant tld and that is all the object does the reason is, as per ens, thinking there would be one tld eg dmp so there is no collision with tlds in DNS thinking it'd be nice dpaths can be used in browser as URI (need a flag somewhere to state if the URI name is mutable or not); so eg v1.rico.dmp.limo/im. * But DNS should be preferred last, and cross checking multiple ethereum RPC for dpath resolution should be first preferred. Thinking out loud here wdyt? : ) function assume(bytes32 salt, bytes32 name, string calldata plain) external { require(owners[name] == address(0), "ERR_TAKEN"); require(keccak256(abi.encode(plain)) == name, "ERR_NAME"); bytes32 comm = keccak256(abi.encode(salt, name, msg.sender)); require(commitments[comm] >= Appraiser(appraiser).appraise(plain), "ERR_PAYMENT"); // appraiser is upgradeable by zone owner commitments[comm] = 0; owners[name] = msg.sender; emit Assume(name, plain); emit Transfer(address(0), msg.sender, name); } | 02:49:25 |
ddapp | * In reply to 🍠kbrav Wrt deferring pricing (there doesn’t necessarily have to be any at all) think you could just have an upgradeable proxy as the zone, idk that there’d need to be a rootzone update for that. But curious to hear thots am thinking it's nice zone owner can change their pricing strategy by taking in the plaintext name and block name registrations (eg revert based on length), and avoid adding proxy (to me it's a little complex); code below totally agree with you rootzone doesn't need to change, the rootzone governor can manually grant tld and that is all the object does the reason is, as per ens, thinking there would be one tld eg dmp so there is no collision with tlds in DNS thinking it'd be nice dpaths can be used in browser as URI (need a flag somewhere to state if the URI name is mutable or not); so eg v1.rico.dmp.limo/im. But DNS should be preferred last, and cross checking multiple ethereum RPC for dpath resolution should be first preferred. Thinking out loud here wdyt? : ) ``` function assume(bytes32 salt, bytes32 name, string calldata plain) external { require(owners[name] == address(0), "ERR_TAKEN"); require(keccak256(abi.encode(plain)) == name, "ERR_NAME"); bytes32 comm = keccak256(abi.encode(salt, name, msg.sender)); require(commitments[comm] >= Appraiser(appraiser).appraise(plain), "ERR_PAYMENT"); // appraiser is upgradeable by zone owner commitments[comm] = 0; owners[name] = msg.sender; emit Assume(name, plain); emit Transfer(address(0), msg.sender, name); } ``` | 02:49:55 |
🍠kbrav | Ah ok so it’s rootzone except with custom appraiser that can be set by owner? | 03:07:57 |
🍠kbrav | The idea is that this is a different protocol so i don’t think we’d have to worry about dns collision generally | 03:14:03 |
🍠kbrav | Like there’s no TLD that points to a network of eth rpcs | 03:15:36 |
sheikhkebab | why | 03:20:10 |
ddapp | In reply to @kbrav:matrix.orgYep, every zone can have an appraiser if the owner chooses, including rootzone because ya doesn't make sense to use DNS when dmp tries to replace DNS | 03:36:30 |
ddapp | In reply to @sheikhkebab:matrix.orgHi, which part? | 03:37:06 |
ddapp | In reply to @kbrav:matrix.org* Yep, every zone can have an appraiser if the owner chooses, including rootzone because ya doesn't make sense to use DNS when dmap tries to replace DNS | 05:28:42 |
🍠kbrav | i think he's asking me | 07:09:46 |
🍠kbrav | because ICANN doesnt like our stupid coins | 07:10:24 |
sheikhkebab | fuck icann | 08:08:25 |
🍠kbrav | worldstar | 08:09:03 |
sheikhkebab | I always thought dmap was a delegation protocol | 08:14:33 |
🍠kbrav | for identity? yeah that's one application | 08:15:49 |
🍠kbrav | anything where it's useful to map stuff you know to stuff you don't know | 08:19:34 |
26 Nov 2023 | ||
ddapp | Redacted or Malformed Event | 01:50:22 |
ddapp | * Isn't using ipfs gateway to get a cid's content a false sense of security? If the DNS or the gateway are hacked.... Literally just bookmarking a hash, hash the code, then compare before using the code, you'd close all the network related attack surface in theory [1]? Feel free to correct [1]: Im figuring out why uniswap interface would fetch js files ... What's the point of them uploading to ipfs if their frontend loads more code over network? | 01:51:16 |
ddapp | * Isn't using ipfs gateway to get a cid's content a false sense of security? If the DNS or the gateway are hacked...., I assume nothing on the browser side checks the CID Literally just bookmarking a hash, hash the code, then compare before using the code, you'd close all the network related attack surface in theory [1]? Feel free to correct [1]: Im figuring out why uniswap interface would fetch js files ... What's the point of them uploading to ipfs if their frontend loads more code over network? | 01:54:42 |
1 Dec 2023 | ||
lasthayek joined the room. | 12:46:38 | |
2 Dec 2023 | ||
🍠kbrav | would be really cool as a starting thing if we could just type in a couple commands with a dpath to download all the type info as well as UI and docs for a protocol, forget if i mentioned. that was kind of long term vision type stuffs when we put dmap/dpack on hold | 01:32:25 |
🍠kbrav | I say "starting thing" just because most people will want one command or even no commands and a LOC. the more ergo it gets the better | 01:35:05 |
@splessnosi:matrix.org left the room. | 18:16:15 |