Early on, the Semantic Web was described as "not a separate Web but an extension of the current one, better enabling computers and people to work in cooperation." [1]. Thus, the focus on human-computer interaction has always been part of the vision. Toward the end of 2006, "User Interaction and Applications" was formally introduced into the Semantic Web framework [2], creating an even greater sense of importance for these conversations to take place within the CHI community.
Unlike the current Web, the Semantic Web focuses on structured data rather than on semi-structured or unstructured data. The Semantic Web offers opportunities to HCI design, enabling more dynamic/complex data interactions which, on the Web, have been very difficult to achieve. At the same time, the Semantic Web poses new design challenges for current interactions like general page navigation, search handling, or presentation for different devices like cell phones. Semantic data can be mode-less: it is not already deliberately constructed as a page, or as site-specific data to be poured into a page container. There are instead a sea of potentially inter-connectable links associated with structured data, possibly from many sources. Consequently, there are new interaction challenges for this sea of structured data:
figure 1. An mSpace[3] visualization of the Classical music domain integrates RDF encoded data from multiple Web sources to enable exploration.
We have examples of how structured data can be presented to be explored when the schema of that data is known in advance. Faceted browsers, like mSpace [3] and Exhibit [4], are a few approaches. None of these approaches is fluid at shifting from one domain to another potentially-related domain. For instance, an mSpace on classical music may reveal something related to the architecture of that period; there is no easy way to switch from Baroque music to Baroque architecture in order to begin to build up a picture of their similarities, differences or influences. So how can users effectively explore possible associations from one domain to another? How much context support is needed to maintain a sense of where one has come from relative to the path through the space? How can interfaces make a large range of (possibly tangential) paths available without cluttering the interface or confusing the user?
Another challenge in the Semantic Web space is how to represent information when the schema of the data is not known in advance, when the data is simply being drawn in via its hyperlink: instead of an address to a page, we are linked to an address for a populated schema, or the data is retrieved automatically and incorporated in the background.
figure 2: An outline view from Tabulator [5], viewing data from various RDF sources. Fields are directly selected to form a query, presented in a table, map, or calendar as appropriate.
So far, there are just a handful of attempts to address this raw generic question: Tabulator [5], Disco [6], and the Openlink RDF Browser [7]. Other projects like Haystack [8] have tried the approach of populating areas of a grid with the appropriate data types. These explorers also highlight the usability challenges of data being loaded on demand, rather than all at once. Again, we no longer have the paradigm of the page to work with as a container; what is the approach we use?
figure 3: A view of US Presidential data from Exhibit [4] . The right shows available facets on the data, which when selected create a new result set.
While much of the current focus is on viewing, navigating and managing Semantic Web data, the implications for users creating data with more structure and inter-connectedness is significant. How can we leverage research projects like the ?Information Scraps? work [9] to facilitate structuring information that is easy to enter? How can we support subject experts in various domains to provide structured information without significant additional burden? What is the role of social creation of data, and collaboratively-managed vocabularies that move us beyond the world of unstructured tags?
These Semantic Web challenges bring a new set of problems for CHI researchers to explore. Directly applying existing techniques to this information is often insufficient, so new approaches are required. The CHI community has the expertise and insight to address the needs for Semantic Web user interaction, which offers new challenges for their research. In particular, we are interested in tools, techniques, and evaluation methods for browsing, exploring, and authoring data.
The Semantic Web User Interaction workshop series and online community was established by senior researchers and practitioners as a bridge between disciplines that contributes to the development of more usable interaction with the Semantic Web. A range of reference information is now available [10], along with a growing cross-disciplinary community [11]. The research challenges for interaction design in this semantic space are also part of the new field of Web Science [12]; the Web Science Research Initiative [13] has co-sponsored the past two SWUI workshops.
[1] Berners-Lee, T., Hendler, J., Lassila, O. The Semantic Web. Scientific American, May 2001. Online: http://www.scientificamerican.com/2001/0501issue/05 01berners-lee.html
[2] Berners-Lee et al, September 2006. Fig. 3.2 The Semantic Web Stack c.2006. Page 22. In ?A Framework for Web Science.? Foundations and Trends in Web Science. Vol. 1, No 1 (2006). http://www.nowpublishers.com/product.aspx?product= WEB&doi=1800000001
[3] mSpace project: http://mspace.fm.
[4] Exhibit: http://simile.mit.edu/exhibit.
[5] Tabulator: http://www.w3.org/2005/ajar/tab.
[6] Disco Hyperdata Browser: http://sites.wiwiss.fuberlin. de/suhl/bizer/ng4j/disco/.
[7] OpenLink RDF Browser: http://demo.openlinksw.com/DAV/JS/rdfbrowser/index. html.
[8] HayStack project: http://haystack.csail.mit.edu/.
[9] Van Kleek, M., Bernstein, M., Karger, D. and schraefel, m. c. (2007). GUI- Phooey! : The Case for Text Input. In press, UIST 2007, Rhode Island. Preprint: http://eprints.ecs.soton.ac.uk/13819/.
[10] SWUI web site: http://swui.semanticweb.org
[11] SWUI community wiki: http://www.websciences.org/swuiwiki
[12] Berners-Lee, T., Hall, W., Hendler, J., Shadbolt, N., Weitzner, D.J. Creating a Science of the Web. SCIENCE VOL 313 11 AUGUST 2006. http://www.sciencemag.org/cgi/content/full/313/5788/769?ijkey=o66bodkFqpcCs&keytype=ref&siteid=sci
[13] Web Science Research Initiative (WSRI): http://webscience.org.