A Journey of Modern OS Construction From boot to DOOM

Wonkyo Choe*, Rongxiang Wang*, Afsara Benazir*, Felix Xiaozhu Lin · 2025

Read A Journey of Modern OS Construction From boot to DOOM by Wonkyo Choe*, Rongxiang Wang*, Afsara Benazir*, Felix Xiaozhu Lin (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-17984.txt.

Introduction

Building an operating system is the classic rite of passage in computer science, and textbooks usually stop at a terminal. Wonkyo Choe, Rongxiang Wang, Afsara Benazir, and Felix Xiaozhu Lin present Proto in 2025, a teaching system built up to a working window manager and games like DOOM.

Proto runs on ordinary portable hardware and carries modern parts: separate address spaces per app, threads, real filesystems, USB, multicore support, and self-hosted debugging. Rather than teach every mechanism first, the course grows the system as a chain of small prototypes, each adding the least machinery a target app needs. Students end with media players, a blockchain miner, and 2D and 3D games running on a system they built.

Catalog record

CREATORWonkyo Choe*, Rongxiang Wang*, Afsara Benazir*, Felix Xiaozhu Lin
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE53 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-17984

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

A Journey of Modern OS Construction From boot to DOOM Wonkyo Choe*, Rongxiang Wang*, Afsara Benazir*, Felix Xiaozhu Lin Abstract . VOS is a first-of-its-kind instructional OS that: (1) Runs on commodity, portable hardware. (2) Showcases modern features, including per-app address spaces, threading, commodity filesystems, USB, DMA, multicore, self-hosted debugging, and a window manager. (3) Supports rich applications such as 2D/3D games, music and video players, and a blockchain miner. Unlike traditional instructional systems, VOS emphasizes strong motivation for building systems—supporting engaging, media-rich apps that go beyond basic terminal programs. To achieve this, we design VOS to strike a careful balance between essential OS complexity and overall simplicity. Our method, which we call inverse engineering , breaks down a full-featured OS into a set of incremental, self-contained prototypes. Each prototype introduces a minimal set of OS mechanisms, driven by the needs of specific apps. The construction process (i.e., forward engineering ) then progressively enables these apps by bringing up one mechanism at a time. VOS makes it accessible for a wider audience to experience building a software system that is self-contained and usable in everyday scenarios. 1 1 footnotetext: Co-primary authors. Figure 1 . VOS running various apps, across different prototype stages, and in different hardware settings. (a)–(c) from Prototype 1, (d) from Prototype 2, (e) from Prototype 3, (f–g) from Prototype 4, (h–m) from Prototype 5. 1. Introduction Over the past several decades, the OS course has been a cornerstone of computer science education. While the specific objectives of OS courses may vary, they commonly aim to convey core system concepts: abstractions, virtualization, privilege separation, concurrency, and software/hardware interactions ( ebling2024resources ) . Beyond these technical concepts, OS courses are also expected to offer a unique experience: building a working system that students can both demonstrate and use themselves. To this end, the OS course is the apex of an undergraduate CS curriculum – where students integrate their knowledge to construct a complete computer system. 1.1. A Missed Opportunity Unfortunately, we observe a declining interest in OS—and systems software in general—among college students. Several factors contribute to this trend: the dominance of AI, the perception that systems work is overly complex, and the frustration often associated with modern systems software. One telling indicator is that many premier North American universities no longer require an OS course for undergraduate CS majors (see Section 2 ). Today’s OS education—and more importantly, the instructional OSes that support it—tend to prioritize teaching technical concepts ( ebling2024resources ) , while often overlooking the motivating power of real-world apps. A unique strength of systems software, especially OSes, is that they can produce tangible artifacts that enable interesting, interactive apps. These artifacts can be demonstrated, played with, and even connected to students’ other passions—whether in art, entertainment, design, or creative expression. This, in fact, aligns with the historical roots of many influential systems projects, which were originally created to serve personal or creative needs. Examples include UNIX , developed to run the game Space Travel ( salus1994quarter ) ; the World Wide Web ( jacksi2019development ) , created for sharing research information; and TeX , written for typesetting the author’s own books ( knuth1984texbook ) . Sadly, the spirit and joy of creating a usable system to support one’s own goals is often absent in today’s instructional OS projects. 1.2. Principles Our response is an instructional OS called Proto , which embodies a guided process for modern OS building. Our first priority is to establish clear and compelling motivations for system building — motivations that go beyond running simple terminal programs. To this end, we target (P1) appealing apps : engaging, interactive programs that ideally evoke an emotional response from the OS builders themselves. This motivation directly influences our design choices, such as treating the framebuffer as a first-class peripheral and prioritizing support for USB and audio IO. We further emphasize (P2) demonstrability : The OS should be able to run on real commodity hardware in addition to emulators. Ideally, it should be portable enough to carry around for live demonstrations, enabling students to showcase their work long after the course ends. Our second consideration is managing complexity: balancing the technical depth needed to support engaging apps with the limited effort students can realistically invest (typically 17 weeks and no more than 150 hours). To achieve this, we adopt (P3) incremental prototyping : The OS is delivered to students in multiple self-contained snapshots, each augmenting earlier ones and is fully functional. Compared to a “big bang” approach, incremental prototypes provide clear baselines for continued development. More broadly, incremental prototyping is a proven method in complex engineering projects such as the Ford Model T, Boeing 747, and SpaceX rockets ( burchardt2020development ; wu2012rapid ; spaceX ) . We also require (P4) minimum viable implementations : Every new OS feature—whether it’s a syscall, threading, or multicore scheduling—must directly support a target app or an essential component of it. We will evaluate the efficacy of these principles in Section 6 . 1.3. Design Through a process dubbed inverse engineering , we first fully implement VOS , before decomposing it into five incremental snapshots, referred to as “prototypes 1–5”. Further through inverse engineering, we map each prototype to one or multiple target applications, ranging from simple 3D animations (Prototype 1) to video games (Prototype 3) and video playback (Prototype 5). See Table 1 for prototype features and their target apps. The forward engineering (i.e. OS construction by students) is then driven by enabling one app after another, re-architecting earlier prototypes and adding new OS features including per-app address spaces, multimedia output, USB peripherals, filesystems, DMA, and multicore execution. At a higher level, this process resembles interactive designs — akin to video games or theme park experiences — by keeping students as OS builders motivated with clear, meaningful goals, rewarding them with tangible progress (making apps functional), and maintaining engagement through timely transitions between different prototype stages. 1.4. Results & Contributions First, VOS is lightweight. Its kernel core is less than 10K SLoC, comparable to existing instructional OSes (§ 6.1 ). Second, VOS is demonstration-ready. It runs on both software emulators and low-cost ARMv8 hardware. As shown in Figure A Journey of Modern OS Construction From boot to DOOM , VOS supports a variety of apps: it can run games (e, i), play music and videos (g, j), and even project presentation slides for classroom use (f). Third, VOS is fast and efficient. Through a careful balance of optimization and simplicity, VOS achieves strong performance on low-cost ARMv8 platforms. On kernel microbenchmarks (§ 6.2 ), it incurs overheads similar to xv6 ( cox2011xv6 ) ; it performs comparably to Linux and FreeBSD on most benchmarks. On app benchmarks (§ 6.3 ), VOS runs DOOM at around 60 FPS and plays back video at around 26 FPS – on par with Linux and FreeBSD running the same apps on the same hardware. VOS also scales to quad-core processors. When running on a handheld device (Figure A Journey of Modern OS Construction From boot to DOOM (h)), the system typically draws less than 4 watts and operates for several hours on a single battery charge (§ 6.4 ). Finally, VOS is pedagogically effective.

Read the complete text