Jump to: navigation, search


Revision as of 03:28, 25 January 2015 by RyanC (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)

NOTE: this proposal, as written does not make sense. See discussion page.


A "Reputation Registry" is proposed which would be the data store for a web of trust or review system, guaranteeing open access and persistence for user-contributed reviews.

Possible names

  1. Reputation Registry, RR
  2. Reputation Data Base, RDB
  3. Feedback Registry, FR
  4. Feedback Data Base, FDB
  5. Web-of-Trust, WOT


Objective: The purpose of the Reputation Registry is to provide a permanent, open-content, and open-source registry in which to store information typically associated with an entity's reputation or quality. The namecoin blockchain would provide the storage medium for this reputation-related information.

The problems to be addressed are as follows:

  • Reputation systems (vendor ratings, business reviews, etc.) are often proprietary. Ratings, reviews, and other reputation-related assessments may be deleted or edited by the system owner, thus contaminating the authenticity or veracity of the assessments.
  • Reputation systems may be impermanent. For example, when the Silk Road closed, all of the buyer and seller feedback disappeared.
  • The information contained in proprietary reputation systems, e.g., ebay's seller feedback, is not freely or easily accessible.

The benefits of storing this information in the namecoin blockchain:

  • The contributor maintains the ability to modify his or her entries in the registry; participation is voluntary and revokable.
  • There is a cost associated entering a record, which should discourage Sybil attacks.
  • The records are open so that developers may build value-added products which compute ratings, risk scores, etc., for entities which are rated.

Related Reading

Reputation System

Past and Future of Reputation and Karma

OpenBazaar's reading list


Objective: Need to balance specificity and generality, and provide an unambiguous and easy-to-parse record.

Name: rr/<name>. <name> indicates the entity which is being evaluated

(Question: Can there be duplicate names in the blockchain? If not, then it is best to make the name a nonce.)

(Need to check that rr/ is not already in use.)

Value: JSON content.

The following are required:

lname = <long name or identifier for the entity being evaluated>

id = id of the entity providing the review; maybe call this, "au," for author?

domain = <the type of entity being evaluated>


context = <comma-separated list of attributes being evaluated>

(should probably make a specific list of recommended contexts for each domain)
global: recommend, satisfied, scam/scammer
personal: honest, skilled, reliable, efficient, intelligent, communicative, creative, independent, thorough
ecommerce: timely, packaging, bargain, asadvertised
product or trade: cost, delay

rating = <comma-separated list of ratings in the form <ns><rating> as described below>

binary: b+ (affirmative) | b- (negative)
trinary: t+ (affirmative, more likely) | t0 (neutral) | t- (negative, less likely)
scale: s1 (strongly disagree, bad), s2 (disagree, poor), s3 (neutral, adequate), s4 (agree, good), s5 (strongly agree, excellent)
data: d<integer numeric value><units>, units: $ (USD), e (EU), y (Yen), n (Yuan), b (btc), m (milli-btc), s (satoshi), d (days)

meta = <additional information about the entity being evaluated, context, etc.>


Seller Feedback


name : "rr/drpool"
value : 
    "fname" : "drpool"
   "domain" : "seller"
  "context" : "recommend, cost, packaging"
   "rating" : "b+, d100$, s2"
     "meta" : "ebay.com"
       "id" : "id/khal" 


Web Site


name : "rr/facebook"
value : 
    "lname" : "facebook.com"
   "domain" : "website"
  "context" : "layout, usefulness"
  "rating" : "b-, b-"
     "meta" : "The advertisements and auto-playing videos suck!"
       "id" : "3f65e8cede8f45ac8f2e710616@gmail.com" 



  • Need to scrape some WOT or reputation data and put it on the blockchain as a proof-of-concept.
  • The domain of entities to be evaluated may be too large, As written, this protocol seems to seek to replace WOT, Yelp, Angie's List, LinkdIn, etc.
  • Need to write simple web interface to rr/ on the nmc blockchain.