gistfile1.txt

ICES:SMALL: An Exploration of Trust-based Currency in a Peer-to-Peer Network
1 Introduction
The promise of online peer-to-peer trading (e.g. eBay) is that it allows even casual users to exchange items
or services in a large, diverse, and efficient marketplace. These transactions are typically between two
strangers and involve asynchronous restricted communications; the buyer often assumes a substantial up-
front risk by paying and then relying on the seller to ship the item. Trust is essential to these transactions.
Online trading systems like eBay often encourage cooperation between strangers by providing a reputation
mechanism, which disincentivizes the seller’s misbehavior and provides the buyer with judgment-supporting
information. However, it has been shown that reputation systems are hampered by the cheap pseudonyms
problem, which penalizes newcomers who haven’t yet paid their dues by accepting poor treatment (or paying
an explicit registration fee) [RZ02, Gro03]. The cost of “dues paying” is unavoidable in systems, which limits
the effectiveness of the market and may exclude many potential users [RO01].
However, it may be possible to circumvent this limitation by considering economic transactions not be-
tween strangers, but between friends-of-friends - that is, between parties who are indirectly connected by
paths of mutual friends in a social network. These weak ties can provide important social and economic
functions, allowing for the spread of diverse information across cliques and communities [Gra73]. In fact,
computer security and internet researchers have leveraged the strength of weak ties to improve the qual-
ity, relevance, and trustworthiness of search results, recommendations, and spam filtering [GH06, GH06,
AMT05, ARH98]. Such systems include an interface that allows users to input data (ratings or reviews)
about their immediate friends, i.e. strong ties, as well as a computational metric that aggregates this data to
infer the larger network of weak ties and display the most promising results. It has even been suggested to
explicitly include individuals in a social network as active participants in an economic transaction, allowing
a mutual friend of two parties to act as an insurer or intermediary in a trade between them [DB05, Fug04].
Can a web-based social-network tool be an effective way of promoting cooperative exchange between
friends-of-friends, just as eBay promotes exchange between strangers?
We will address this question through two phases of user-study evaluations with a prototype called
‘Stuffnet’, a peer-to-peer exchange network based on smart-contracts [MMF01], virtual representations of
IOUs that are only as good as trust in the issuer’s word, but that can be easily transferred throughout the
network. Smart-contracts allow multiple parties in a transaction to follow and agreed-upon set of rules, min-
imizing each party’s reliance on the honesty of each other (and on third parties, e.g. the service provider).
We plan to explore the use of these contracts as a currency that facilitates transactions between indirectly-
trusting parties. Our work will contribute to an open field of research called “human currency interaction,”
which involves the application of human-computer interaction (HCI) techniques (user-centric design, usabil-
ity analysis, ethnographic methods) to systems of exchange [WM08]. Our prototype development process
and experimental design will be focused on two exchange scenarios in which we believe we will be able
to observe not just user satisfaction, but also an improvement in economic efficiency compared to existing
online trade systems.
In the first phase, we will explore the use of web-based social networks to support one of the most
clearly socially-embedded forms of economic exchange: the gift. Specifically, we will build a competitive
online game (inspired by the ‘Potlatch’ [MH00]) that encourages regifting as a way of extending the reach
and effectiveness (and therefore reciprocity) of gifts. Since gifts-in-kind are currently associated with a 10%
destruction of value due to unwanted gifts[Wal93], we aim to improve the efficiency of gift-giving as the first
benchmark for evaluating the potential of our prototype.
In the second phase, we will use Stuffnet as a tool for redistributing secondhand items. We will expand
the functionality of the prototype to include the features expected from an online marketplace: search,
reputation, and dispute resolution. Our hypothesis is that users will be socially motivated to exert effort in
1
assisting their immediate friends in online transactions, by patiently mediating disputes and by risking their
own reputations to vouch for their friends. By defining smart-contracts that can include multiple stakeholders
in an online transaction, an individual can limit his vulnerability by relying on his strong ties while benefitting
from the wider reach and greater diversity of his weak ties.
We believe this approach can channel social influences to facilitate economic exchange within an ex-
tended network of indirectly-trusting peers, especially benefitting low-volume sellers and casual users who
are under-served by existing online trading systems.