[{"data":1,"prerenderedAt":179},["ShallowReactive",2],{"\u002Fblog\u002F2020-10-15-what-a-certificate-on-a-public-chain-proves":3},{"id":4,"title":5,"archiveType":6,"body":7,"category":163,"contextDate":164,"cover":165,"coverAlt":166,"date":167,"description":168,"extension":169,"featured":170,"historicalDate":167,"historicalLabel":164,"meta":171,"navigation":172,"path":173,"publishedDate":164,"seo":174,"stem":175,"thumbnail":164,"workPeriod":176,"writtenDate":177,"__hash__":178},"blogs\u002Fblog\u002F2020-10-15-what-a-certificate-on-a-public-chain-proves.md","What a certificate on a public chain proves, and what a verifier still has to check","retrospective",{"type":8,"value":9,"toc":155},"minimark",[10,14,25,30,33,68,71,75,86,92,108,114,118,121,124,127,131,134,152],[11,12,13],"p",{},"On 26 October Harvest Finance lost about $24 million. Its vaults valued deposits using the state of a liquidity pool, and that state could be pushed out of line by a large trade and pushed back within the same transaction. The chain recorded every step accurately. The mistake was in what the contracts concluded from that reading.",[11,15,16,17,24],{},"I think about the same gap with certificates. ",[18,19,23],"a",{"href":20,"rel":21},"https:\u002F\u002Fmedium.com\u002Ftrivechain\u002Fintroducing-trivecredential-trvc-app-updates-log-13-12-2019-c4847296fb79",[22],"nofollow","TriveCredential",", which we introduced through TRVC.App in December 2019, records certificates on Trivechain using TriveAsset, our token layer. By the end of that year we had issued certificates on the chain for real events, including the hands-on workshop TriveAcademy ran with the Malaysia Blockchain Association. As lead developer of the chain those certificates sit on, I want to be exact about which questions the chain answers and which it leaves to the person checking.",[26,27,29],"h2",{"id":28},"what-the-chain-can-prove","What the chain can prove",[11,31,32],{},"A certificate issued through TriveAsset gives a verifier four things they can check without asking us.",[34,35,36,44,56,62],"ul",{},[37,38,39,43],"li",{},[40,41,42],"strong",{},"It existed by a certain block."," The issuing transaction is in a block, so it existed by the time that block was mined.",[37,45,46,49,50,55],{},[40,47,48],{},"It came from a particular key."," The issuance was signed by whoever controlled the input it spent, and the ",[18,51,54],{"href":52,"rel":53},"https:\u002F\u002Ftrivechain.github.io\u002Ftriveasset-protocol",[22],"TriveAsset protocol"," derives the asset ID from the first input of the issuing transaction, so the ID itself points back to where the issuance came from.",[37,57,58,61],{},[40,59,60],{},"Its details have not changed."," TriveAsset keeps metadata off the chain, on IPFS, and records the file's 34-byte IPFS hash in the transaction. Any copy of the metadata can be checked against that hash.",[37,63,64,67],{},[40,65,66],{},"Which address holds it now."," Ownership follows the output that carries the asset.",[11,69,70],{},"None of these says who issued the certificate.",[26,72,74],{"id":73},"what-it-cannot-prove","What it cannot prove",[11,76,77,80,81,85],{},[40,78,79],{},"Who the issuer is."," The metadata format has an ",[82,83,84],"code",{},"issuer"," field, but it is text the issuer typed. Anyone can issue an asset whose metadata names a university. The link between an issuing address and an institution has to come from outside the chain: the institution publishing its issuing addresses on its own website, under its own domain, where a verifier can find them.",[11,87,88,91],{},[40,89,90],{},"That the key was still theirs."," If an issuing key leaks, certificates made with it are as valid on the chain as the real ones. A verifier needs the issuer's record of key changes, and that record also lives outside the chain.",[11,93,94,97,98,103,104,107],{},[40,95,96],{},"That the certificate still stands."," On a UTXO chain, the output carrying a certificate can only be spent by the key that holds it. An issuer cannot pull a certificate back. Revocation therefore has to be a separate statement by the issuer, and a verifier has to look for it as well as for the issuance. The W3C's ",[18,99,102],{"href":100,"rel":101},"https:\u002F\u002Fwww.w3.org\u002FTR\u002F2019\u002FREC-vc-data-model-20191119\u002F",[22],"Verifiable Credentials Data Model",", a Recommendation since November 2019, gives this its own property, ",[82,105,106],{},"credentialStatus",", and leaves the scheme to each implementation. TriveAsset's Expiration rule can give an asset a lifespan in blocks. It can later be shortened, but only in a transfer made by whoever holds the asset, so it does not give the issuer a way to withdraw a certificate.",[11,109,110,113],{},[40,111,112],{},"That the holder is the person named."," An address shows who controls a key. If a certificate can be transferred, holding it proves possession, not achievement. The protocol's Holders rule can limit which addresses may receive an asset, but rules like this are applied by the software that reads the chain, not by the chain's consensus. A verifier has to use a reader that applies them.",[26,115,117],{"id":116},"keeping-people-off-the-chain","Keeping people off the chain",[11,119,120],{},"A public chain cannot forget, and data-protection law, including Malaysia's Personal Data Protection Act 2010, expects personal data to be correctable. So names and identity-card numbers do not belong in a transaction or in public metadata.",[11,122,123],{},"What can go on the chain is a fingerprint. A plain hash of a name and an identity-card number is weaker than it looks, because the set of plausible values is small enough to try one by one. Adding a random value that only the holder keeps makes the fingerprint useless to anyone who has not been given both.",[11,125,126],{},"The same caution applies to IPFS. The TriveAsset metadata format lets a field be encrypted to a public key, but an encrypted field is published for as long as anyone hosts the file. Encryption that is adequate today has to stay adequate for the whole life of the record.",[26,128,130],{"id":129},"the-verifiers-list","The verifier's list",[11,132,133],{},"This is the list I would give anyone checking one of these certificates.",[135,136,137,140,143,146,149],"ol",{},[37,138,139],{},"Is the issuing address or asset ID one that the institution publishes itself?",[37,141,142],{},"Was that key still the institution's on the date of issue?",[37,144,145],{},"Has the issuer published a revocation for this certificate?",[37,147,148],{},"Does the metadata file match the hash on the chain, and can it still be retrieved? A hash proves a file is the right one; it does not keep the file available.",[37,150,151],{},"Does the person presenting it control the holding address, or hold the random value behind the fingerprint?",[11,153,154],{},"The chain settles one question well: whether the record was changed after it was written. The rest are questions about people and keys, and their answers have to come from somewhere a verifier can trust for its own reasons.",{"title":156,"searchDepth":157,"depth":157,"links":158},"",2,[159,160,161,162],{"id":28,"depth":157,"text":29},{"id":73,"depth":157,"text":74},{"id":116,"depth":157,"text":117},{"id":129,"depth":157,"text":130},"Security",null,"\u002Fimg\u002Fblog-covers\u002F2020-10-15-what-a-certificate-on-a-public-chain-proves.webp","A certificate card with a blue seal resting on a glass block, with a magnifying glass beside it.","2020-10-28","Using TriveCredential and the TriveAsset metadata format to separate what Trivechain can prove about a certificate from what a verifier has to check elsewhere.","md",false,{},true,"\u002Fblog\u002F2020-10-15-what-a-certificate-on-a-public-chain-proves",{"title":5,"description":168},"blog\u002F2020-10-15-what-a-certificate-on-a-public-chain-proves","2020","2026-11-07","1Ap1E99uVKoVl82T8zQ0FzO5xjiKsCkMrMLdl0jCHj4",1794838285843]