6.1850: Computer Systems and Society

Learning Objectives

The design decisions we make in computer systems imply power structures, affect individuals, society, and the world, even when it seems like they don't. Our primary goal in this class is to make the connection between these technical design decisions — e.g., whether to use a hierarchy or a mesh — and their impact on humans explicit by using an ethical framework as part of our design process.

We'll study an existing ethical framework — the "ethics protocol", designed by Abby Jacques and Milo Phillips-Brown — and apply it to the process of designing computer systems, with the goal of understanding the impact of various design decisions, and the goal of improving the framework itself within the context of computer systems.

Students will:

• Evaluate the consequences of technical design decisions beyond the impact on typical metrics (performance, scalability, etc.); specifically, assess the impact of decision on users/customers of the system, on people who are not users/customers of a system, on society at large, and on the environment. To study this impact, we will often look at what power structures a design decision implies (I.e., who has power over whom).
• Apply an ethical framework to computer systems design, and analyze the usefulness of this framework; evolve the framework as necessary
• Apply this framework to their own designs, to make ethically-informed decisions
• Learn to communicate in forms that are common in the field of computer systems—for example oral presentations and written reports similar to published conference papers in computer systems—and to audiences of both experts and peers.

6.1850 is divided into roughly three modules. After we spend a week introducing the ethics protocol (we'll call that Module 0), we'll look at each of the following through the lens of the ethics protocol:

• Module 1: Examples of large-scale power structures that stem from technical decisions in systems. Here, we'll see big, global systems like DNS, BGP, Bitcoin, etc., and study the power structures that are implicit in them, and the impact of those structures on society. For example, the DNS hierarchy implies that a particular entity owns the root zone; who should that be? What are the consequences? Bitcoin's proof-of-work mechanism gives power to those with the most compute power; what is the impact of that?

• Module 2: Examples of small-scale power structures that stem from technical decisions in systems. We'll ask similar questions here as we do in Module 1, but the systems we study in this module are typically thought of as being at a smaller scale than systems such as DNS. For example, weighted-fair-queueing is fair, for a particular definition of fairness; does that definition also make it fair for the people impacted by the queueing scheme? Are people impacted by queueing schemes?

• Module 3: Examples of technical decisions that limit access and/or impact the environment. In this Module, we'll look at design decisions that result in limited access to a certain population as well as the environmental impact of physical infrastructure for systems. For example, what incentives are there for ISPs to serve rural areas? 5G networks require many cell towers to be placed closely together in order to achieve high speeds; what is the environmental impact of those physical towers? We'll see some of these questions come up in the first two modules too.

In each module, we'll apply the ethics protocol to many of the systems we study. It's this ethical framework, and not the specific systems themselves, that is the overall theme of this class. As such, the list of systems we study might change from the initial plan, especially if there is a system that students are particularly excited to learn about.

Each week we'll have two ninety-minute sessions. The sessions themselves will be a mix of lecture, discussion, and occasional student presentations. Ninety minutes is a long time to sit in one place, so you can expect there to be a ten-minute break somewhere near the middle of class each day.

On Thursdays we will introduce new material in class. Students will spend some time thinking about this material afterwards, reading some things, doing some reflections, and then we will have a broader discussion on the following Tuesday. (In this sense, you can think of Thursday + Tuesday as a single week in 6.1850.)

30% participation
40% final project
30% other writing assignments

We expect you to work hard in this class, but we don't expect you to be perfect. There will be lots of opportunities to revise assignments, and lots of guidance from the course staff along the way. If you are willing to put in the work, you can expect to do well in this class.

Participation

Participation will be graded very similarly to how we handle it in 6.1800_[6.033], where we look at whether students are coming prepared to class, actively listening, speaking up, asking questions, etc. We do not base your participation grade on how often you give a "correct" answer in class. And we understand that different people participate in different ways; there's no one right way to participate.

Part of your participation grade will come from short, weekly, written reflections about what you're learning in the class, and what connections you're making to other topics.

Final Project

The final project will be done in teams of 2-3; we may allow some students to complete the project individually depending on their circumstances. Although the final project is worth 40% of the final grade, there will be multiple deliverables along the way, and multiple opportunities to revise. You can expect something roughly along the lines of:

• Phase 1: Generate multiple ideas for projects; get feedback from staff (and possibly other students)
• Phase 2: Narrow down to one idea with a formal project proposal
• Phase 3: An interim oral report delivered to staff and possibly other students
• Phase 4: A final, written report

Individual Writing Assignments

The driving force behind our policies is that we care about you as people more than we care about any deadline or grade. This will be a small class (15-25 students), which means we are able to be relatively flexible with our policies.

Late Policy: Assignments will have due dates. If you are unable to turn in an assignment by its due date, get in touch with the course staff; you don't need to give us a reason for your request. As long as you reach out ahead of time, you can assume we will give you at least a 24-hour extension, and we'll work with you to figure out whether a longer extension is appropriate. We want to keep you on track as the class progresses, but you should never pull an all-nighter because of this class.

Attendance Policy: Because we grade for participation, we expect you to be attending class regularly. But, things come up. If you miss three or fewer classes during the semester, your participation grade won't be impacted at all as long as you let us know ahead of time that you'll miss class (we'll make an exception for things like medical emergencies, of course). If you need to miss more than three class periods, for example due to illness or a crisis outside of class, get in touch with the course staff. You don't need to explain the situation to us, but we may ask that you work with an S3 dean so that we can figure out the best course of action.

Note that these policies don't mean that we can accommodate every possible situation. But we will do our best to make sure that you have as positive experience in this class as possible.

Who is this class for? Anyone (who has taken 6.1800_[6.033]; see below) who is interested in thinking about the impact of computer systems on individuals, society, and the environment. If you vaguely enjoyed the content of 6.1800, even if you don't think of yourself as a "systems person", and if you especially enjoyed some of the 6.1800-in-the-news segments, or some of the ethical issues that we brought up in recitation, this might be a class for you.

What should I know going into this class? Don't worry about reviewing specific technical content in preparation for 6.1850. If you've taken 6.1800, you'll be fine; we'll review necessary details of certain system as we go along.

What's more important is your attitude towards this class. 6.1850 is still relatively new, and some of the material might change on the fly; you'll need to be able to roll with some changes as things progress. We'll do a decent amount of discussion each day; you should expect to be an enthusiastic participant in those discussions.

You should not expect this to be a high-stress class. There are no exams, the final project has lots of opportunities for feedback and revision, and the small size of the class means that we can be pretty accommodating to student circumstances. That said, we expect you to work hard. If you aren't willing to put in the work to preparing for this class and participating in it, don't take it.

I'm interested in this class, but I didn't get an A in 6.1800; should I still take it? Most likely, yes. 6.1850 is intended for students who were enthusiastic about the material in 6.1800 and are interested in thinking critically about the impact of the systems we build on the world at large. Your contributions will be valuable regardless of your expertise; getting an A in 6.1800 is not a prerequisite.

I haven't taken 6.1800; can I still take this class? No; the 6.1800 pre-requisite is strictly enforced. The material will build directly off of the content and structure of 6.1800, and will hopefully feed back into next semester's offering of 6.1800. Even if you are a systems expert, you cannot take this class if you haven't taken 6.1800.

Will class sessions be recorded? No; 6.1850 is more discussion-based than lecture-based, so recordings aren't ideal. For students who need to miss class (due to illness, e.g.), we'll help you make up the material in a different way.

What degree requirements does this class fulfill? As far as Institute requirements go, 6.1850 is a CI-M for 6-1, 6-2, 6-3, and 6-4 students. For departmental requirements in Course 6, 6.1850 fulfills both the EECS and the II requirement; in the new 6-3 requirements, it is part of the Computers and Society track. It is not an AUS.

Will Broccoli or Junebug ever come to class? No; that would be a disaster. Rest assured I'll show you plenty of photos.