Big ideas about information

Data is a silhouette of reality

Although I’ve been at the University of Washington Information School for ten years now, I still remember my first reaction to the iSchool’s faculty position job ad to which I applied: “What in the world is an iSchool?”

After a decade of hearing that same question, I now know how to answer it reasonably well:

“You know all the data in the world? iSchools study it. Where it comes from, how it’s biased, how it informs people, how it shapes decisions, and how to create systems, both software and otherwise to help people create, store, analyze, and share it. From cradle to grave, information changes lives, and we both study and invent new ways for it to do this.”

I get a lot of different reactions to versions of this answer. Some say, “Oh, that’s really interesting.” or “Is that really a thing?” And sometimes I get, “Oh, is that like computers?” The occasional software engineer reacts with “Information isn’t just bits?” Much like people back in the 1960’s launching the first computer science departments, I regularly find myself having to explain an entire discipline of information science in a few sentences.

I’ve always been a bit uncomfortable in this position. I’m not from the discipline. I don’t know the foundations of information science, how it differs from computer science, or how it connects to the many other information disciplines in academia. I’ve always felt like a bit of an imposter, defining and defending my school when I don’t really feel ownership or mastery over its foundations.

This Autumn quarter, I finally got the chance to fix this: I taught INFO 200 Intellectual Foundations in Informatics, one of the foundational survey courses we teach to mostly freshman and sophomores interested in pursuing our highly popular Informatics major. This was my first chance in ten years to finally learn the foundations of information science and get that mastery I’ve been longing for.

Here was my approach:

• I began last year by crowdsourcing big ideas about information from my fifty wonderful faculty colleagues.
• After organizing all of the fascinating ideas they shared, I did a deep dive into many seminal papers from which those ideas emerged.
• After finishing all of that reading, I spent this past summer synthesizing everything I learned into the design of a digestable survey course for freshman
• I then spent much of this Autumn creating content for the course (lectures, readings, activities, and assignments).

In the end, the result was a course anchored by several dozen big ideas about information. In the rest of this post, I attempt to distill most of these ideas into blog post form, including little examples to help illustrate the big ideas.

What is information?

• Data is something people perceive in the world (e.g., the words on this screen).
• Information is data that has meaning to an individual (e.g., you might interpret these bolded words as terminology I’m trying to define)
• To interpret data as information, data requires context (e.g., the writing above told you that these bullets are big ideas)
• Metadata approximates the context of data (e.g., my name, my bio, and when I published this, all approximate who, what, when, where, and why I wrote this)
• Data itself is encoded, reflecting a designer’s values (e.g., I’ve encoded this in English and Unicode, revealing my western-centric views)
• The process by which data and metadata are encoded is central to interpreting the meaning of information (e.g., you have to know the process I used to devise these big ideas to know if they’re sound)
• The emotions surrounding data mediate how we interpret information (e.g., if you’re a social conservative distrustful of progressive liberal scholars like me, you might have already decided to not believe anything I’ve written)
• Because of all of the above, data and its metadata are not necessarily accurate, ethical, helpful, or fair (e.g., you have to judge who I am and how I wrote this to know the veracity of what you’re reading)

How do people access information?

• The only way to access information is through an interface to an information system (e.g., you’re probably reading this on a computer connected to the internet)
• An information system is a process that organizes people, technology, and data to allow people to create, store, manipulate, distribute, and access information (e.g., Medium’s network of authors and it’s publishing platform and policies is one information system embedded in the much larger information system of the web)
• Information systems span all of human history and all media (e.g, smoke signals, the Library of Alexandria, radio, telegraphs, and newspapers are all non-digital examples)
• Information technology can help automate parts of an information system’s process (e.g., written language, stone tablets, the printing press, and the web all help automate storage)
• No information system is ideal for all tasks (e.g., the internet is really lousy at retrieving memories from my youth because it doesn’t have an archive of them, but it’s really good at quickly finding cat videos)

How do people find information?

• Exposing metadata in interfaces streamlines searching and browsing (e.g., the title of this post and its image helped convey its relevance to you and also helped it appear in search results).
• Interfaces are not natural; they must be invented and learned (e.g., you didn’t always know how to use the device you’re likely reading this on, nor did the device always exist)
• Using an information to find information requires involves sensemaking, which is a subjective, social, evolving, reflective, and narrative process (e.g., you’re not just consuming these big ideas, you’re hopefully reflecting on them, integrating them into your larger understanding of current events such as misinformation, disinformation, and social media)
• We satisfice when we sensemake, limits what we find and what we know (e.g., you probably aren’t going to verify every one of these points I’m making, but instead trust me because of my credibility as a professor)
• Web search engines — which mine web pages for links and text, creating an index that is used to answer queries — only accelerate the retrieval part of sensemaking (e.g., judging the quality, meaning, relevance, and veracity of information is entirely your job)

What does computing have to do with information?

• Computers are just one form of information technology (e.g., before computers, people relied a lot more on books, newspapers, and friends)
• Computers mindlessly and quickly execute instructions carefully written by programmers (e.g., they have no concept of what this sentence means, a programmer just told it to render it on your screen quickly)
• Computers only work with data, stripping information and knowledge down to bits (e.g, Medium doesn’t store the meaning of this post in it’s database, nor the impact of that meaning on your knowledge, just the bits that represent its text).
• It is important to decide what parts of information systems should be automated with computers, and which part should remain in human control (e.g., should Medium use algorithms or people to decide which stories to censor?)

What are privacy and security?

• Privacy is having control over information about yourself and your activities (e.g., I can control whether this Medium story is posted, but I can’t control who sees it)
• Physical privacy is mostly under your control (e.g., if this story was just on paper, I would have considerable control over who read it).
• Digital privacy is mostly under the control of private companies and government (e.g., Medium, not me, controls who sees my writing)
• Getting control over digital privacy requires regulation (e.g., I can’t force Medium to give me control over who sees my writing without passing laws that compel it to)
• Security is the extent to which an information system can make guarantees about who has access to private information (e.g., Medium makes certain promises around protecting sensitive data, such as my password, my unpublished stories, and unpublished comments on my stories).
• Security requires software, hardware, human, organizational, and regulatory vigilance (e.g., even if Medium’s software was perfectly secure, one of it’s engineers could probably access and disclose my private information)

How do organizations shape information systems?

• Organizations shape people’s use of information by designing products and policies (e.g., I can’t really control what information systems exist for sharing writing in the world, only whether I choose to use Medium)
• Organizational culture shapes the information systems that organizations create (e.g., Medium’s belief in an internet funded by great content, not by great advertising makes its writing platform quite different from Facebook’s)
• Organizations make design tradeoffs that introduce bias in their information systems (e.g., by focusing on writing rather than virality, Medium has imposed strict formatting limitations on posts, limiting my ability to express my identity through typographic form)

How do societies shape information systems?

• Values also shape the information systems society creates (e.g., if I’m unhappy with Medium’s design choices, I can use a different platform to blog, shifting Medium’s priorities)
• Laws regulate what we can and can’t do with information (e.g., I can’t post any of my students’ grades on Medium because of the U.S. FERPA laws)
• Information technology may only amplify social change, not cause it (e.g., Medium’s publishing platform can’t make the web any less advertising driven, but it can amplify the values of writers like me who believe in this vision).

Did you think those ideas were interesting? So do the nearly 1,000+ students a year at UW who after taking INFO 200 apply to be an Informatics major. Much of the rest of the degree is about deepening students’ understanding of these ideas, while also teaching them to design, build, evaluate, and redesign information systems that account for these big ideas. The power in their skills is therefore taking the vast quantities of information systems in the world—some digital, some not—and making the tough choices about how to evolve them to better serve the goals, values, and needs of the world.

Because I’m a computer scientist (and students are often choosing between computing and informatics), I think it’s important to explicitly contrast an undergraduate degree in information and an undergrad degree in computing:

• Whereas computing is focused on one particular information technology, informatics is focused on all of them (though our students are primarily interested in computing).
• Whereas computing is focused on algorithms, informatics is focused on data (though our students certainly learn about algorithms, as they’re handy for automating parts of information systems).
• Whereas computing is focused on processing data with a computer, informatics is concerned with the meaning of data in the world.

Because our students are primarily concerned with computing as an information technology, they usually end up working at the same technology companies as computing majors, but overseeing the front end, the databases, and the broader product management concerns. Moreover, Informatics majors find themselves outside of the software industry just as often, working in domains where understanding the particular challenges of a domain is key to finding successful solutions (e.g., health, education, government, science).

After a year of learning about information science and teaching INFO 200 once, I definitely feel like less of an imposter in my school. But I don’t yet feel like an expert—just expert enough to teach hundreds of freshman a new way of seeing the digital world they take for granted.

A few disclaimers:

• Some of the ideas above might be wrong. Science is always provisional!
• Some of the ideas above are over-simplified. In some cases, that was intentional, as this was an intro course.
• I might have misunderstood some of the big ideas in what I read. If I did, let me know! I’d love to learn the nuance I missed.
• Some of the academics in my audience might think “That’s a horrible misrepresentation of information science!” I’m okay with that; I’m less interested in the artificial disciplinary boundaries we create and more in the ideas they produce.
• I’m sure I missed something. Do you have a big idea about information you think is important? Share it with me!