Florian Cramer on “Why Semantic Search is Flawed”
Posted: November 15, 2009 at 1:01 pm | By: Liliana Bounegru | Tags: florian cramer, semantic search, semantic web, Web 2.0, web archive | 1 Comment
Florian Cramer, head of the Networked Media Master at the Piet Zwart Institute in Rotterdam, ended the last session of The Society of the Query conference. The Alternative Search 2 session presented a few of the latest web technologies as potential directions for the web and search engine design in the near future: RFDa, which would make the shift to what Steven Pemberton named the web 3.0, and semantic search, as implemented in the Europeana project.
Florian Cramer concluded this series of presentations with a critical and somewhat pessimistic evaluation of the current state of the web and the idea of a semantic web and semantic search, as one of its potential futures. His three main arguments revolved around: “why search is not just web search (and not just Google),” “why semantic search is flawed,” and “why the world wide web is broken.”
The first point expressed his frustration with the narrow understanding of the notions of query and search engine on which the conference focused. As he explains, wikis and social networking sites also include the search engine functionalities.
As far as semantic search is concerned, Cramer usefully pointed out to the difference between folksonomies, the currently used form of semantic tagging, and the universal semantic tagging which a semantic web would require. While folksonomies are “unsystematic, ad-hoc, user-generated and site-specific tagging systems,” (Cramer, 2007), like the tagging systems of Flickr for example, the semantic web would require a structured, universal tagging and classification system which would apply to the entire web. Cramer is skeptical of the possibility to create this unified, ‘objective’, meta-tagging system because classifications, or taxonomies, are not arbitrary but expressions of ideologies, which would call for the discussion of the politics of meta-tagging. While meta-tagging may have its advantages, such as arguably empowering the web users and weakening the position of large web services corporations, although still maintaining the necessity of search engines to aggregate data, it also has several potential weaknesses. The semantic web model must be based on trust in order to prevent some predictable problems, such as massive spamming.
In the concluding section, Cramer expressed his concern that the Internet as a medium for publication and information storage is not sustainable and argued for redundancy in web archiving. However this desire for permanence raises questions about the nature of the medium itself.
November 16th, 2009 at 2:01 am (#)
Thanks for this summary. – A few elaborations and minor factual corrections: Search engines do not just exist on the web, but in all digital media devices (including mp3 players, E-Book readers, PC operating systems), and have become one of their crucial selling points; contemporary debates on search engines need to reflect this.
Historically, unified classification schemes/taxonomies have been the center of Aristotelian (scholastic) science and were abandoned by modern (empirical) science for good reasons.
Please allow me to correct the summary “that the Internet as a medium for publication and information storage is not sustainable and argued for redundancy in web archiving”. I did not speak of the Internet as a whole, but only of the World Wide Web, which is massively transforming from an online publishing medium into a real-time software application platform. Commercial search engines and the “walled gardens” of so-called “Web 2.0″ platforms and social networking sites have become so powerful because they make up for deficiencies in the original, still unchanged design of the World Wide Web: the insufficiency of hypertext for information retrieval, the instability of URLs, the absence of distributed/redundant storage, revision control (i.e. document version rollback) and the absence of a public domain full text search index (which everyone could use for building their own search engines with its own filtering, ranking and user interface) as part of the standard technological infrastructure of the web.
Since I had to rush these points in a twenty minutes presentation, my lack of technical explanation and use of terse technological terms probably made many people in the audience misunderstand what I was trying to get across. I am not arguing in favor of setting the web in stone, and “clean it up”, in order to make it a long-lasting archive. Rather, I think that the weaknesses of its current design are the real cause of the current ‘cognitive capitalist’ monopolies over web search and online publishing platforms. The issues can, in my opinion, not be fixed by some alternative to Google and other web services, but only on the deeper level of how the web works.