Social Computing Social datorisering? Sociala system? Social informatik? Social data? Peter Lönnqvist, DSV future ubiquitous service environments
Social Computing Vad är det Vad ska det vara bra för Massor med exempel Frågor och diskussion
Social Computing Vad är det Människor är sociala djur
Social Computing Vad är det Kan vi få (en del) sociala behov tillfredställda i en värld där vi ständigt lever med mindre fri tid, på längre avstånd, etc. med hjälp av informationsteknologi? Kanske…lite…
Social Computing Vad ska det vara bra för Intelligenta gränssnitt behöver finna nya sätt att adaptera på Vad händer om man inkluderar andra människor i loopen och låter dem hjälpa varandra? Människor är bra på att ge personliga råd Människor är bra på informationsfiltrering Låt oss ta en närmare titt på hur man kan sammankoppla människor för att de ska kunna hjälpa varandra
Social Computing Vad ska det vara bra för Social navigation Direkt Direktkommunikation mellan användare Synkron/asynkron kommunikation Att koppla samman rätt personer Awareness (medvetenhet?) om andra Att man är medveten om andra gör att man tar kontakt (Ackerman 1995)
Social Computing Vad ska det vara bra för Social navigation Indirekt Kollaborativ filtrering explicit/implicit rating Historieberikade miljöer Försvinnande över tid
Social Navigation The term Social Navigation (SN) captures everyday behaviour used to find information, people, and places primarily through watching, observing, imitating, following and communicating with other people. As a design approach social navigation tries to raise awareness that social activities should be part of our information processing environments. We can distinguish between systems where actors are in direct contact with one another (direct social navigation), and systems where contact is anonymous and indirect (indirect social navigation). Very likely, all these formshave their justification, but each of them brings up a number of research issues discussed here.
Context Some researchers hold the view that SN could be a general process running in the background of all applications, so that all aspects of interaction are suffused with information of other people's actions. A contrasting view bases SN's utility on the task domain and advocates tailoring the design to the tasks, user group and context. Another unresolved design question is whether SN is an inherently spatial concept. Will SN only be applicable in information spaces expressed through spatial metaphors or will spatial metaphors at least be beneficial for social navigation processes?
Extracting information Feedback Many store-based recommender extract recommendations for future transactions from previous purchases. Voting with a credit card is a strong statement indeed. However, these systems do not capture what happened before a purchase: did customers compare many items or did they know exactly what they wanted? Did they come back over and over again till reaching a purchasing decision? Similarly, there is little post-purchase feedback whether an item met a customer's needs.
Extracting information cont’d Bootstrapping How can a SN system provide useful social cues before a significant number of people have used the system? And how can a new previously unrated item be entered in the system? Or, in the case of a SN system where users are present and visible in real-time, what do we do when a user is alone in the system? (after all we need to keep the user in the system long enough to give other users a chance to join him/her)
Extracting information cont’d Ratings Social cues can be captured explicitly or implicitly. Explicitly voting or writing recom-mendations requires a deliberate effort. Therefore, systems based on voting are based on possibly higher quality, but sparser information. On the other hand, systems using implicit information on page views or purchases collect this information as effortless by-products of users' activities. When do we use which of these approaches?
Extracting information cont’d Snowball effects The snowball effect concerns the situation where more and more users follow each other down the wrong path, thus reinforcing the wrong recommendation.
Aggregation Once the input from users is gained, it can be aggregated and refined in various ways depending upon numerous considerations. One consideration is timeliness of recommendations. A forest path is an example of a social navigation structure in the real world, which exists only as long as it is used on a regular basis. Related metaphors could be used to indicate popular event Web sites such as for the Olympic games. Event sites tend to drop in popularity very quickly after the event but a Social Navigation system might continue to rate the site as popular, causing people to visit the site and keep the popularity higher than is justified. Other obvious considerations concern users' tasks and whether the task is to merely locate recently updated, reliable, and recommended information or to find long- trusted generally accepted information. In this second case the information might not be rated as popular, as fewer people access it, but it is well- trusted information.
Information Quality Sometimes SN should be based on popular consent / activities (mass response), where in other cases it is more prudent to base recommendations on a single expert's opinion (expert response). In the first case the population of users providing input may be entirely unrestricted or restricted to a community of interest or community of practice. How can a social navigation system determine who is an expert? One possible solution is to have the expert claim his/her expertise and to use social navigation processes to rate the expert status based on responses by already trusted experts, by the entire population or by a trusted subgroup of the community.
Ensuring trust Trust has become a hot issue lately; indeed trust might be the key issue deciding between success and failure of eCommerce. As long as the quality of recommendations and the competence of reviewers cannot be guaranteed, it will be difficult to convince people to fully trust SN systems. Trust is one of the areas where indirect social navigation provides a clear advantage, because information is aggregated over many people.
Social affordance When a user sees how another user uses a system he/she can learn from watching; a form of social affordance. The watching of other users creates a place for vicarious learning, where new users can not only ask questions but also watch more experienced users in the system. But how do we create designs that make other users' actions visible so that others can learn from them, ideally without entirely compromising the privacy of the first person.
Reshaping experience Social navigation design can involve altering the structure of an information space. An example is the use of recommendation links in systems like Amazon.com. Links to "related items" and "people who bought this also bought X" change the structure of a navigational space. Such systems are a first step towards empowering actors to make the functionality and structure of a system drift and to make our information spaces more fluid, ultimately empowering a user community to customize systems to their needs. Customization of this kind, though, might jeopardise consistency in structure.
Privacy SN will require a certain, limited amount of visibility of our actions, which infringes on our privacy. Privacy concerns are a major stumbling block towards the acceptance of SN systems. Erickson and Kellogg use the concept of social translucence to describe that it is not only necessary to see other actors, but to clearly communicate that what information is disclosing and how it is used. Social translucence entails a balance of visibility, awareness of others, and accountability.
Evaluation There is a difference between concluding that SN happens in the world no matter what we do, and deciding that it is a good idea to design system based on this phenomenon. How will our perceptions of our systems change? Incorporating SN might not, for example, change the paths we might take through the Web, but we think it can profoundly influence how we experience our systems and will significantly change interactions we have with other people mediated through our systems. We still lack the empirical grounds needed to know which kinds of SN forms are most relevant in different circumstances and domains, thus more user studies are badly needed
Social Computing Mer exempel Nätverksspel telefoni/sms andra artefakter
Social Computing Mer exempel telefoni/sms andra artefakter
Social Computing Mer exempel andra artefakter
Social Computing Att tänka på Passande? När och hur ska man koppla samman människor Personalisering Vi människor skräddarsyr ofta information Avatarer Har fysiska begränsningar och kan bidra till onödiga förväntningar Enkelhet Att göra det enkelt att använda är bra HCI Titta på existerande system innan ni bygger egna
Social Computing Peter Lönnqvist, DSV future ubiquitous service environments