User:Sebastians

Father Christmas

Wiki War will be back

---

Performing a search on Google is a bit like looking at a Wunderkammer (cabinet of curiosities). Of course, this comparison is not really apt as the search results mostly aren't curiosities (in fact, they should be way more often!). However, the search results page is an assemblage of things found in a space, that has expanded to such a degree, that we can't make any sense of it anymore without the help of search engines (which, however, aren't and shouldn't be the only solution to this problem). This space is in fact our externalized collective memory.

By crawling the web, and by analyzing, categorizing, and storing what they find along the way, search engines try to create a paradigm (in the linguistic meaning: a collection of signs) out of which answers to our questions can be constructed. The construction of meaning depends a) on the existence and availability of elements (what we put on the web, and what then gets indexed by the search engine) and b) on the rules applied in order to construct the answer.

It is obvious how the construction of meaning, the formation of knowledge, is shaped by algorithms (rules). However, these algorithms are far from being transparent. In fact they are well kept secrets, often being regarded as a myth. In this sense, the comparison to the Wunderkammer starts to regain some meaning, as Wunder in German not only means curiosity, but also miracle.

It is not only knowledge that equals power, but also the way how knowledge is being shaped. Therefore, algorithms that are not transparent establish and maintain a hierarchy of power.

Google's Search by Image

The decoding of visual signs has long remained a domain of the human being. However, as technology learns to “see”, to decipher visual signs, images become searchable information, too. What really fascinates me at the moment is Google's Search by Image, a feature that allows the user to submit images as “questions”.

How does Google make sense of this immense pile of images, by looking “at” the image, and not at the textual context only? How does the algorithm work, if at all? Is this a useful tool to gain insights into our collective visual memory? Or is it simply a tool empowering media companies and their lawyers, as the search feature makes it in fact very easy to find “illegal” copies of a an image?