Distributed To - Do Lists Using Online Social Networks ( Progress report # 2 ) Clifton

Shopping is a common activity in daily life. Not only do people typically visit multiple locations to purchase their goods, but the realization of a need to purchase a particular item may come at inopportune times. Most people carry scraps of paper for these moments to jot down reminders for themselves to purchase particular items. Furthermore, the typical household common area may have a public todo list, where occupants may scribble down items that need to be purchased on the next trip to the grocery store. Leveraging the opportunistic reminder that an electronic post-it note affords, we observe the potential utility of a system that not only allows multiple people to input to-do items, but reminds participating users at opportunistic times of items to purchase. In this paper, we describe a system that leverages online social network relationships to allow users entering a coffee shop to purchase drinks requested by their friends. By avoiding a two-phase commit protocol, and by providing users with additional contextual information, we hope to minimize the total travel time and the total number of redundant trips necessary, while allowing for a time-optimal solution. We discuss technical challenges of the proposed system, and the remaining issues that still need to be studied. INTRODUCTION The average American spends over 6 hours a week shopping, with much of this time spent traveling from one location to another. In addition to requiring a long amount of time (our most precious resource), this traveling consumes great amounts of fuel at a large economic and environmental cost. In this paper, we present a prototype system that allows friends in an online social network to share “to do” lists so that they can coordinate their shopping efforts in the real world. Our system ties a social network‟s central “to do” list to location based services so that group members learn when they are in a position to help out a neighbor in the network. By simply entering a location to which a friend is planning on visiting at a later date, members are notified via their cell phones that they are in a position to help a friend out by purchasing them an item. Our architecture avoids a distributed locking algorithm, which is likely to be unreliable and slow in our vastly distributed network. Instead, our users participate in a novel distributed, two-phase commit system to ensure that to-do items are purchased in an efficient manner. The awareness of the status of other co-located users facilitates the coordination among them to complete tasks on the shared to-do list without duplication of effort. It is our hope that the widespread use of such a system can save its users an untold number of hours of traveling between shopping locations; traveling that is redundant when one‟s social network is viewed as a whole. MOTIVATING SCENARIO The following scenario provides a high-level description of our system and motivates the need for a distributed to-do list by showing the system‟s positive effects on the total amount of effort spent by a small group of friends. The Status Quo: An Inefficient Shopping Behavior During her morning coffee break, Alice remembers that she needs to buy bread and dog food on her way home from work that day. Unknown to Alice, her co-worker Bob makes a quick trip during his lunch break to the same supermarket that Alice would drive to at 5pm. After picking up bread, Alice hops in her car and continues to the pet store. While in the pet store, Alice runs into her neighbor Carol, who is there buying cat food. They then drive home in separate cars. Although their purchases of food can be completed by other people, they cannot coordinate their shopping well since they have no clue of the friends‟ context. We thus believe that users can leverage their opportunities to ask their friend to purchase what they want by knowing the friends‟ context. In the next section, we will see how the shopping behavior will change with our proposed system. A Distributed To-do List During her morning coffee break, Alice remembers that she needs to buy bread and dog food on her way home from work that day. She adds these two items to her online social network enabled “to do” list. After lunch, Bob knocks on Alice’s door and hands her the loaf of bread that he bought for her. He had gone shopping over his lunch break. On her way home, Alice stops at the pet store to buy dog food, and when she steps through the store door, her cell phone informs her that her neighbor Carol is in need of some food for her cat. Alice checks this item off of Carol’s list, purchases both the dog food for herself and the cat food for Carol, and heads back home for the evening. Figure 1. Many items on people’s to-do lists can be easily purchased by friends in one’s social network, if only friends were made aware that they were in a position to help. In this scenario, the total number of store visits has been reduced from 4 to 2, and the total number of car trips from 3 to 2. Although we just see the case with a small number of friends, we expect that the users could coordinate their shopping more effectively with a large social network.