The Limitations of Search Engines Lie in Specificity
The article referenced in the post:
https://www.wired.com/insights/2014/05/searches-created-equal/
This article, which appears in a science and technology news website called “Wired”, analyzes the nature of most search engine queries and their intended goals. Paulo Guadiano, who wrote the article, asserts that, oftentimes, search engines like Google can be overwhelming and frustrating due to their inability to produce the exact results that we want. Guadiano goes on to offer a multitude of explanations as to why that might be the case, including limitations like “filter bubbles” that have been added to modern search engines. Filter bubbles, which are utilized by Google, filter results based on the user’s information (e.g. location, search history, etc.). The following is a link to a TED talk about filter bubbles and their effects on search results (https://www.ted.com/talks/eli_pariser_beware_online_filter_bubbles?language=en). Other limitations include the tendency of users to only choose links on the first page of the search results, something Guadiano says is a well-documented phenomenon. These challenges are to be expected; there are 15 trillion web pages to narrow down.
Guadiano, however, suggests a bigger problem with search queries; he challenges readers to search for things like “I want to hang a nice picture on my wall” and see for themselves how search queries can produce frustrating results. He points out that many of the search queries that result in poor results are “open-ended”. In other words, searching for a nice picture doesn’t pose a specific goal for the search engine. What one individual determines to be an acceptable picture is horrendous to another individual. Search engines are built to retrieve specific information and the results for a query about a nice picture to hang are bound to disappoint some (or most) users. Another example would be searching for a good present for a significant other or for suggestions for restaurants in a new city. Specific searches are not subject to these problems; anyone who inquires for the address of a museum or its hours of operation will find the answers, provided they exist on a web page.
Similar to what Guadiano does in this article, in INFO 2040, we highlighted the limitations of search engines mechanisms like ranking. The problem of synonymy occurs when words mean the same thing and useful results are left out because of this difference. For example, “green onion” and “scallion” mean the same thing and a search for one would leave out results that would appear in a search for the other. Another problem is polysemy, which occurs when one word means many different things. An example of this phenomenon would be “jaguar”, which could produce results for both the animal and the car company. “Apple” could produce results for the fruit, or for the company. These problems, just like Guadiano’s, are also the result of a lack of specificity. The search engine has no way of knowing exactly what the user is searching for, culminating in frustration. These problems are as old as these search engines themselves, because, as we learned, the initial goal was to create automated information retrieval systems for repositories. However, our coursework did not stop here in defining the problems facing search engines. PageRank, which Google and other large search engines base their algorithms on, face the challenge of pooling pf PageRank values in pages that have no links leading out since they are tailed onto strongly connected components (SCC).
Link analysis reveals some critical challenges facing modern search engines and their ability to deliver users the results that they want. However, both Guadiano and our coursework highlight the flaws in our search habits. The open-ended and ambiguous nature of search queries leads to flawed results. As a society, we rely heavily on search engines to provide answers conveniently and quickly. But because of our dependence, we treat search engines as living breathing entities that hold all the answers to our questions. It would be wise to address our searching habits and analyze whether or not our searches are bound to end in frustration. Guadiano recommends that, the next time we search, we decide if we are searching, or simply exploring.