|
 |
|
|
Information and People
I've noticed a few more hints that our concept of data, realized in computers through file stores, databases, and other structures, is not at all the same as how people truly interact with information. Does It (the information) Matter? First, a person looks at a bit of information, say an article about some new product, and decides whether that information is something that matters. Mattering is, of course, an opinion. It is influenced by one's present-state beliefs and expectations of what the future will be. I daresay that a patient on his or her deathbed would not think that a two-page explanation of the new features in Microsoft Power Point would be compelling to read. (Though, having said that, I can think of at least one exception: the original inventor of Power Point might actually wish to read that on his deathbed. Maybe.) Whether information matters or not could also depend on factors like these:
Although some of these issues have analogs in computing (such as "what format is it"), machines do not approach data with anything like a question "Does it matter?" Computers either take what they are given, or follow algorithms to filter and evaluate data as it comes through an input port of some kind. Is It Credible? Next, a person who consumes some information makes a judgment about the information's credibility. Often, this judgment isn't completely informed by any meaningful criteria: "There are lots of statistics, therefore it must be credible". Credibility derives from several factors, some in the control of the author, some in the control of the publication (context), and some factors are in the hands of the person who receives the information. Authors of information can add credibility by doing this:
Context can affect credibility:
Readers or recipients of information can influence the credibility of the information in this way:
Computers don't have any evaluation of credibility, per se. With computer worms and viruses becoming a threat, operating systems and anti-virus software are beginning to evaluate whether software has been "signed" or whether it may be running without the end user's permission, but the computer itself is not making the judgment: the anti-virus software contains rules that capture the judgment criteria of professional anti-virus experts at the software company itself. The computer just does what it always does. Same goes for computing concepts like authorization and authentication of users, which might be argued as a form of credibility. The process of authentication validates that the user is able to reproduce steps that should not be easy to reproduce (for example, the user knows a password, which should be difficult to guess). Even if this is an algorithm (which it is), I'd also add that authentication is a boolean "You pass / you do not pass" situation. In the real world, people may still read something that they do not assign full credibility to, and may hold the information for some time, perhaps looking for counter-arguments because they want to test, after the fact, whether the content really should be believed or not. Imagine, for example, if Time magazine, a fairly well respected journal, ran an article claiming that the President of the United States is actually a being from another planet. You might doubt the story, but you might then watch other stories to see whether they actually support or refute the claim. Machines don't do this, and most algorithms would not approach a data problem with anything resembling this approach. Is It Appealing? Appeal is another one of those squishy criteria that humans demand of information, but computers do not. An article with wall-to-wall text and no white space or headings might be the best content in the world, but many people would avoid it just because it appears too unappealing in its density. The same article could become more appealing, with the same content, with some carefully chosen white space. Appeal could apply to the form or the content of the information. Humans want their information presented in a way that appeals to their sense of self. Technicians want information presented in a no-nonsense manner. Visual artists want information presented in a visually pleasing manner. Executives want information presented in an easy-to-scan form. Whether the previous assertions about how people want their information are always true or not, the point is that people make subjective judgments about information. These judgments may or may not be relevant to the overall fit of content to information inquiry. Computers do not make subjective evaluations of incoming information, and algorithms that attempt to do so are only imposing a scaffolding of instructions between input and data storage. Is It New? "Is this information new to me?" asks the recipient. A fascinating article about spiders might be completely unimportant to someone who chairs the American Arachnological Society. People will judge information based on whether it is something that they know already or not. There are some folks who only read what they already know. Others build on their preexisting knowledge. And, in fact, everyone does both, really. But, imagine if a newspaper always had front page headlines like "Sidewalk On Main Street Is Made Of Concrete" and "Wood Is Flammable". Totally redundant information would not be very interesting, and probably would not be valuable to you. Do I Hate It? People are social creatures, and one of the side effects is that we tend to relate to the rest of the world (at whatever scale you like) in terms of "us" and "them". "Us" could be my family, my professional industry, those who share my hobbies or interests, those who are of my religious faith, etc., and "them" is everyone who is not "us" in a given context. This social division affects how we interpret information. If I were a Rush Limbaugh fan (which I am not), I would regard an article by Michael Moore with a lens of discredit and hate. Sports fans tend to appreciate any information they get from commentators who root for their team. If my wife informs me of a unicorn in my back yard, I would give her a little more credibility than I'd give a passing stranger telling me about his unicorn. Part of whether I hate the information or not is whether it's tailored to who I am, too. My larger self, my opinions, and my goals are all lenses that cause me to be sensitive about certain things. If information is presented that dismisses or challenges these, I will tend to disregard the rest of the information, even if it has merit of its own. Obviously, machines don't hate information. The concept of data completely ignores this subjective judgment that people in the real world apply when they interact with reality. Where Does It Lead? Information in the real world is not usually considered an end in itself. I think most information that I deal with feels more like a stepping stone than a static "fact". I am constantly evaluating information in terms of what the larger implications are, where it will lead, what it means. In some cases, I'm synthesizing the information with everything I know, in other cases, I'm seeing where it leads in terms of research. Maybe I need to learn about X before I can really understand Y (or even know that Y is relevant in the first place). Computers deal with data as if each data point were a static fact. Trend analysis and so forth apply algorithms to information, but computers usually are only able to evaluate these trends in terms of prescribed criteria; computers cannot synthesize new information as meaningful in any sense that mimics human information processing.
|
