The application that has taken shape since the beginning of the CHI fellowship executes queries against the British Museum’s (BM) SPARQL endpoint. The BM system returns results serialized as XML- or JSON-LD. The application updates its data store (Redux) and renders the collection of results as a simple list. This is straightforward enough. On the other hand, I’ve had to think a bit more about how to render single query results in response to user selection of any given list item.
The basic data structure in semantic web technologies is the Resource Description Framework (RDF) triple. The triple is itself a graph, expressing a subject and an object’s relationship through a predicate. The Rosetta Stone, for instance, appears as a subject in ~180 statements. Given the centrality of the RDF triple in semantic web systems, it’s no surprise that a common way to represent the resource interrelationships expressed in these triples is by showing networks of nodes and edges; subjects and objects as nodes, predicates as the edges connecting them. This seems the most intuitive way to render single BM collection objects.
RDF and the Resource Description Framework Schema (RDFS) are fundamental to the semantic web technology stack as it’s developed in the past two decades. RDFS extends the constructs defined in the RDF specification to enable kinds of expressions like generalization. While RDF does allow for classification (instance-class relationships) with the instanceOf property, the RDFS extends the framework to allow for the expression of constructs like generalization hierarchies, where a class relates to another class as its sub- or superclass, with the subClassOf and subPropertyOf properties. For an RDFS model (e.g., each version of CIDOC-CRM), implementation in the Web Ontology Language (OWL – another important technology in the semantic web stack) adds formal logic to the model and thus enables automated reasoning with it.
I’m working on using these technologies to enrich what are otherwise pretty boring network visualizations. The lodlive.it project is a good example of visualizing data structured as RDF triples in a compelling way.
Last week I was able to attend the annual conference for the Association of College and Research Libraries (ACRL) in Baltimore (proceedings here). One talk struck me as particularly relevant to cultural heritage informatics. The authors explored humanities scholarship collaboration across institutional and disciplinary boundaries, focusing on the fifteen institutions that constitute the Humanities Without Walls (HWW) consortium (which includes Michigan State). The authors reported the themes that emerged from qualitative analysis of semi-structured interview transcripts, and represented the resultant collaboration network as a network graph with dimensions including team-size per project, number of grant awardees per institution, and number of connections between projects and institutions.
The authors also advised of ways to engage with these innovative networks of humanities scholarship collaboration. For the Adaptive Research Practices theme that emerged from the analysis, the authors advised intervening at the level of student-instructor interaction. This fellowship’s emphasis on learning techniques for creating and maintaining web content represents this sort of engagement. Regarding the Networks of Scholarship theme identified in the qualitative data analysis, the authors recommended that information professionals support places for scholars to interact across conventional boundaries, for instance in research centers, communities of practice, or other hybrid organizational forms.
Finally, the authors encouraged information professionals to foster experimentation in novel forms of Scholarly Communication and Dissemination. Significant, but certainly not the only developments include the use of blogs and personal websites for scholarly communication and dissemination, the use of platforms like GitHub, GitBook, or Authorea, and the use of networks like SSRN, Mendeley, Zotero, or ResearchGate.
Considering electronically-mediated scholarly collaboration, one challenge to effectiveness is the very knowledge in which researchers trade. Codifying knowledge that is part of a rapidly evolving domain is a problem particularly for the knowledge work of distributed research teams, whatever the branch of the sciences (Bos et al., 2007). On this point, researchers in computer-supported cooperative work (CSCW) have advanced our understanding of the need to enrich virtual knowledge-work-spaces, suggesting features like document annotation or authorship tracking to produce affordances like knowledge evolution monitoring (Malhotra and Majchrak, 2012). One particularly exciting development that seems to address this obstacle to collaboration is content that is published on the web not only for dissemination but, even more, to elicit wider contribution. In this respect, these projects (perhaps a subset of what have been called virtual research environments?) constitute something more than scholarly communication/dissemination. Examples of this kind of project include the Modes of Existence project (site / source), and the Platform for Experimental and Collaborative Ethnography, which grew out of the Asthma Files project (site / source).
As I mentioned last month, one of the ideas of the semantic web is to render data from specialized, disparate sources comparable, and this is achieved by developing specifications like CIDOC-CRM. One implementation of CIDOC-CRM is the Erlangen CRM. Heritage institutions like the British Museum use implementations like this to organize their collection. It is implemented in the Web Ontology Language (OWL) and can be browsed in an ontology explorer like Protégé or by just reading the XML.
“immaterial items, including but not limited to stories, plots, procedural prescriptions, algorithms, laws of physics or images that are, or represent in some sense, sets of propositions about real or imaginary things and that are documented as single units or serve as topic of discourse” (CIDOC-CRM, n.d.).
According to the documentation, a set of exemplary instances of this class are the common plot points of Kurosawa’s The Seven Samurai and Sturges’ The Magnificent Seven. A query to a SPARQL endpoint in order to materialize that collection’s Propositional Objects might read as follows:
# declare a prefix
# this allows us to refer to objects in the schema directly rather than by their full URI
# e.g., in the query below, crm:E89_Propositional_Object rather than the full URI http://erlangen-crm.org/current/E89_Propositional_Object
PREFIX crm: <http://erlangen-crm.org/current/>
# a) the variable that the server should return (?instance)
# b) that the server should return unique instances only (with the DISTINCT modifier)
SELECT DISTINCT ?instance
# specify the pattern for the server to try to match
?instance a crm:E89_Propositional_Object
# state how the response should be ordered…
ORDER BY ?instance
# and the quantity of instances to limit the response to
My project involves working with some of the technologies of the semantic web. The main idea of the semantic web (or web 3.0, and in Berners-Lee’s language the “read+write+execute” phase that will supersede the “read-only” phase of web 1.0 and the “read+write” phase of web 2.0) is for web services to reason automatically about resources. Robust descriptions enable the linking of heterogeneous resources. Semantic web services commonly use the Resource Description Framework (RDF) to represent entities in terms of subjects, objects, and predicates.
Take the hyperlink, for instance: where a classic hyperlink connects one document on the web to another, web 3.0 proposes to link data within a document to data within another. This proposition depends on data modeling work: specifying a domain’s things as categories (classification), identifying supersets and subsets of its classes (generalization), and specifying the part-whole relationships in which its classes participate (aggregation).
Where descriptions of resources are robust, we are able to build services to reason about them intelligently and automatically. Europeana, the European Union’s platform for heritage content, provides good documentation and maintains a SPARQL endpoint, which I’ve found helpful for learning about these technologies. See the video I’ve posted above for more information.
This division includes activities of libraries and archives; the operation of museums of all kinds, botanical and zoological gardens; the operation of historical sites and nature reserves activities. It also includes the preservation and exhibition of objects, sites and natural wonders of historical, cultural or educational interest (e.g. world heritage sites, etc).
This division excludes sports, amusement and recreation activities, such as the operation of bathing beaches and recreation parks (see division 93).
Moving forward, I’ll need to do some reading about the extent to which activities in this space can be described in concepts like between-firm value networks and within-firm value chains composed of business processes, not to mention the very concept of the firm, and whether organizations acting in this space can even be described as such.
I’m thinking this project will involve almost as much conceptual plumbing work as it will design and development. I’m not unhappy about that, but it should go without saying that suggestions on literature to check out will be welcome!
This past Friday, we talked about licensing. As I’ve mentioned before, I’m interested in information and communication technologies (ICTs) and traditional knowledges (TK), which are the product of the intricate relationships between indigenous peoples and the specific places in which they have lived or used to live for long periods of time. Questions of licensing that are specific to this class of knowledge – the question of who has the right to access, use, modify, distribute, profit from, or otherwise relate to some piece of TK – are warranted by the historically antagonistic relationship toward TK of modernization projects like residential school systems for indigenous children, or the removal of artifacts for sale, analysis, and display in museums or private collections.
I’d argue that at this point in time, ICTs and traditional knowledges relate in at least three ways:
ICTs can be instruments for the organization of TK. Knowledge objects like place names, demonyms, and flora and fauna names have historically been a) subordinated to western names, and b) relegated to arcane corners of the Library of Congress (LOC) or Dewey Decimal subject headings (“Mythology, Other”). The Xwi7xwa subject headings developed by the Library of the University of British Columbia is a widely cited positive example of the organizational role of ICTs in TK. The modeling of Andean craftsmanship documented in Brownlow et al. (2015) is a good example of ontological organization of TK.
ICTs as aides in the application of TK alongside (if not as integral than at least as supplementary to) professional knowledges, as in the domain of natural resource management. An example of this might be web resources documenting species names and associations between a species and an environment, habitat, or another species (as a reference, for instance, in administering and complying to Canada’s Species at Risk Act). Given the history of extracting TK, provisioning such a web resource would need to involve careful consideration of the problem of TK licensing; the problem of whether and how to provide for more sensitive items (e.g., information about the associations between a given species and uses such as medicinal; food and beverage; technology, arts and crafts; and social or religious rites).
I recently switched to one of these wireless service providers with a sort of pay-as-you-go model, and so now I’m (trying to be?) thriftier when it comes to the data I consume. Earlier today, as we wrapped up our cultural heritage project pitch websites and learned about mapping tools, the option I was recently given on my phone by the Google Maps application to download my area for navigation offline prompted me to take a step back and think about my team’s project website in terms of design choices, especially with regard to the cost of accessing it.
Providing some context: for the cultural heritage project, we thought of a smartphone application repackaging the audio tour mp3s at Stockholm’s Vasa Museum, currently available on their website, in order to streamline visitors’ audio tour experience. Beyond thinking about the means by which end-users would get ahold of the application (i.e., by streaming the audio, downloading all mp3s at once, or downloading them one by one during a tour), we did not get too much into the details of implementation, but the project pitch site can serve as a kind of placeholder for thinking about the implementation of cultural heritage projects.
First, the networking section of the Google Chrome developer tools provided some useful information about our project site. At its root, our site weighed 6.4 MB, and took 5.26 seconds to load on my home wireless network. The images, alone, accounted for 6.3 of the 6.4 MB. The larger one of the two images (4.3 MB) is hosted with GitHub here in the U.S. and took 513 milliseconds for the browser to retrieve; the smaller one (2.0 MB) took tens times as long (5.14 seconds) to retrieve from its server in Stockholm. Then, using Xcode’s Network Link Conditioner to simulate full network throughput and a latency of 500 milliseconds (about the latency of a satellite connection), I found that our site took 40.58 seconds to fully load in a browser in those conditions.
Hi there, CHI community. I’m Michael Nelson, and I’m really excited to be a part of this cohort. I’m in my second year of the master’s program in Media and Information at MSU’s Department of Media and Information. After watching movies and eating food, I’m most interested in exploring scholarly communication and knowledge management. I’m currently exploring those latter two topics in research for my master’s thesis and in other projects at my college. I also enjoy project-based work on these topics, as I found this past summer working at the Inter-university Consortium for Political and Social Research as part of the University of Michigan iSchool’s Research Experience for Master’s Students program.
In short, I’m enthusiastic about these topics both as a researcher and as a practitioner.
Right now, I’m thinking my project might involve following up on some coursework I’ve done in the field of information and communication technologies (ICTs) for development (the field commonly known as ICTD or ICT4D). I’m especially interested in ICTD projects that have been applied in the context of indigenous knowledges.
My interests are informed by the work I did as an undergraduate and the work I’ve done since then. Studying history at North Park University in Chicago, I became especially interested in knowledge production, specifically knowledge in the human sciences. Then, the ups and downs of working at a startup investment fund and subsequently at a Chicago high school during the years between undergraduate and graduate study spurred my interest in returning to school in the domain of information science.
Looking forward to collaborating with and learning from other Fellows this year, and also to discussing all that right here on this blog with the CHI community!