Записная книжка сисадмина Seven Myths of Google

    Seven Myths of Google
    I love Google. As a librarian, I love the way it can thrust me into the middle of the web and give me the answers that I need. But, like most love affairs, there are irritations, particularly for information professionals. In the spirit of friendly criticism, here are seven myths Google encourages that make librarians’ jobs more difficult.
    Google frustrates the life out of me because it’s so sure of itself, so convinced that it is the god of the information age. In reality, it’s the best we have for an internet plagued by a lack of metadata.
    What is more, Google haunts me.
    Everywhere I go, it’s urging me to log in. If I don’t log in, it still knows where I live. Will it show up one day at my doorstep demanding that I search more?
    But enough of my own paranoid vision. I want to lay a measure of blame on Google because it has created a mythology that challenges the development of information literacy in its users and complicates the professional lives of librarians. Thus, on several fronts, I accuse Google of presenting us with a false, or at least misleading, vision of the world of information.
    Lest you think I am singling out Google, my comments apply equally to Bing. Yahoo, and a lot of other search engines. Google, however, is pervasive worldwide. Its influence affects the performance and perception of all web search engines.
    While Google would never state many of these openly, here are seven Google myths we can blame it for promulgating:
    Of course search is simple. Google was the first search engine to work, really work. And Google taught us that there is no reason why anyone can’t seek and find. Are you able to type a word or two? Then you’ve aced search. There is no need to fear or doubt, because Google's algorithms will do the work for you.
    It’s a central message of any major company — promote its product as the answer to everyone’s needs and dreams. If that company is focused on search, then it is going to tell you that it has simplified search. To be sure, Google's algorithms are fairly amazing, and people actually do find what they’re looking for a lot of the time.
    But let’s consider what it means to say that search is simple. In the future, when search engines are powered by injected human brain tissue or who knows what, search will become intelligent to the point that it will be simple. But the mantra “Search is simple’’ is a semantic nightmare.
    Throwing words into a box is simple. Getting a bunch of results is simple. But search is not simple. To conceptualize what search is really about, we need to recognize first that it is a highly complex process. In the mind of a human being, a concept is comprised of subject matter, specialized focus, and a particular motive. This concept can be expressed in words, but probably not well in search terminology.
    A searcher, for example, wants to learn the best way to move apps to the start page of a particular tablet so that the apps are available when the tablet is turned on. That same searcher does not want the tablet to display only the most recently used app icons on the start page. So this becomes the search terminology: apps move to start page [name of tablet]. And it might work. Or not.
    What is more important is the fact that a complicated idea got turned into stilted words that look like baby talk. In search, the adult has to talk like an infant while expecting adult results. Even a natural language search boils down to the engine using only the most significant words, while the rest of the words in the question are ignored. Search is not simple. It has been given the illusion of simplicity by a search engine that has a vested interest in convincing you that it is simple.
    We can succeed without metadata because Google wants us to believe AI will triumph (someday). In September 2013, Google announced a new algorithm, which it called Hummingbird. Fueled by the need to serve users who use audio via phones to do searches, Hummingbird enlists natural language and semantics to make searches that are more like real questions. Natural language and semantic search, of course, is not new. The original AskJeeves was built on it years ago.
    What is really interesting about Hummingbird, however, is the fact that Google is finding itself caught between the two major ways to do verbal search — keywords and metadata. The net, for the most part, lacks metadata, so Google has been developing its “Knowledge Graph," a set of subject pages that build relationships around concepts. Using the data from millions of search histories, the Knowledge Graph is trying to map the various ways searchers conceptualize ideas with keywords. Hummingbird feeds, to a large extent, on Knowledge Graph information and thus, ultimately, has created keyword searching on steroids.
    At the same time, Google wants webmasters to insert details that look more like traditional metadata into their web-pages. Google and other search engines are active supporters of schema.org, which develops more or less standardized markup language for websites. Yet the folks at Google are aware that it will still be crawling a lot of websites that won’t add markup tags.
    Google thus remains focused on enhancing keyword language. Essentially, it has to succeed without metadata. This puts the ongoing development of metadata, not to mention the continuing use of metadata in academic databases, in doubt. If Google can actually succeed without it, do we need to spend all that money providing it in databases that currently use it?
    Google wants us to believe that no one wants advanced search. It has, therefore, hidden this capability. The last time I looked, it was under the settings link at the bottom of Google’s home search page. While Google only buries its advanced search, Bing no longer has it at all. Why this de-emphasis on advanced searching? Two answers are the most obvious. First, hardly anyone uses it. Second, Google and other search engines have a vested interest in making basic search successful.
    The “No one uses it, so why emphasize it?" argument makes sense in a way. Any product that doesn’t sell should be canceled, right? However, information professionals know a tool that can enhance search but isn’t being used provides an opportunity for education, not cancellation. I am always perplexed by the line of argument that says, «They won’t use our catalogue and databases, so let’s give them a discovery tool.» Why do we educate in every other sphere but let users do what they want without guidance when it comes to search?
    The second reason to de-emphasize advanced searching comes from the fact that search engines have a vested interest in doing magic with basic searching. That's where the bread and butter of Google's reputation come from. Why stress advanced tools that put search more in the hands of the user, when it makes the best business sense to have basic search running rings around competing search engines?
    Users like their results to be personalized, thinks Google. Most of us are well aware of the personalization done by search engines, using our search histories as data to shape unique results for each of us. We want this, according to Google, so that we can get results more in tune with our interests. But Google benefits when it can tailor its ads more personally to us and make more money.
    I’ve pretty much given up trying to stop personalization, because Google prevents so many things from happening if I’m not logged in. But personalization is all good anyway, isn’t it? For the researcher, it’s not. Researchers alter their identities on a regular basis by seeking out new topics that may have nothing to do with what they searched for yesterday or last week. Personalizing results limits options. It’s not all good.
    Algorithms do the evaluation for you. Let’s face it. The content in the web is a mixed bag. Alongside high-level academic and scientific work is Uncle Fred’s home remedy for lumbago and Ditsy Eddie’s latest conspiracy theory. Not to worry, because the algorithms rank results by relevancy. The first page of them is the good stuff, and the shoddy and lunatic material is far down the list, right?
    Google doesn’t actually say this, but it prides itself on its algorithm structure, and the implication is clear: Trust Google and it’ll get you the good stuff. For users, this is terrific. Why do the hard work of evaluation when the search engine can do it for you? Except that reality is far from this kind of Utopia. Try a search on 9 / 11. The day I did this (early 2014), the first five results looked fairly solid, but the sixth was a conspiracy theory site. High quality is not guaranteed.
    Google is only a search engine, but it carries the aura of being some sort of quasi-human information vendor who actually presides over a library of websites. It wants to be a universal library and, for many users, it is. Users regularly say, “I found it on Google," as if it were a department store for information or a library. The belief is that the algorithm curates the collection so that Google can deliver the best resources for any topic. Google is thus more than a search tool. It is, for many users, a library collection.
    Librarians, who actually know what constitutes a genuine library, beg to differ. We understand that careful collection development and extensive metadata are the elements that lead to high quality. When quality is not controlled as resources are added, a search engine such as Google has to find some way to vet the data, doing it at the point of search. Yet it’s not at all certain that its algorithms, now or in the near future, can do that task well. Thus, the web can’t be considered to be a library or a viable alternative to traditional libraries.
    If you want users to decide what they want to access, without the «censorship” of traditional library selection limiting their options, then Google manages something that some people might call a library. But there is a lot to be said for real people determining what will be included in a catalogue or database.
    The real problem with Google as a library, however, is that searchers can only access easily what they can get for free. In our world of academic publishing, most of the stuff people pay for never even shows up in a Google search. Whether or not we like the academic commercial machine, most of the really important current books, articles, and so on are not available to searchers via Google. My biggest concern lies here — that avid Google users, thinking that the web provides the whole world of information, are missing the most significant resources.
    One click is the only way to go. People want easy.
    They want quick. Google has committed itself to providing easy and quick. It’s in Google’s best interests to give us a lot of hype about how easy it is, with one click, to find things with the superior Google algorithms.
    Question: Do you really think we'd have library' discovery tools today if Google hadn't created the mythology' that no one should have to perform more than one click between search words and results? The real problem with our rapture across keyword searching, and hence our disdain of anything that requires multiple clicks, is that it has dumbed down search processes that actually use metadata. If more than one click is inherently bad, then we are left in utter dependence on keywords and algorithms, neither of which can come near to delivering the kinds of resources a multi-click faceted search, using metadata in a sophisticated database, can give us.
    Google has been an amazing boon to the general public interested in finding information and to librarians as well. I'm not the only librarian to love Google. So why knock it? Google, after all, has had to cope with nearly zero metadata and uneven quality of resources. It’s doing what it can.
    I blame Google because it hides its inadequacies, replacing them with hype about its power. Its mythology' is powerful, and the general public, student populations, and even some information professionals accept the myths as truth.
    Pulling people back from total dependence on a tool that promises much but is so ultimately limiting is a big task. That’s why I believe information literacy is so crucial.
    William Badke.

    0 комментариев

    Только зарегистрированные и авторизованные пользователи могут оставлять комментарии.