Gespeichert in:
| Hauptverfasser: | , , |
|---|---|
| Format: | Recurso digital |
| Sprache: | Englisch |
| Veröffentlicht: |
Zenodo
2026
|
| Schlagworte: | |
| Online-Zugang: | https://doi.org/10.5281/zenodo.19359883 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Inhaltsangabe:
- <p><strong>Episode summary:</strong> For over forty years, the digital world has been organized like a physical filing cabinet: folders inside folders. But the human brain doesn't think in hierarchies; it thinks in associations. In this episode, Herman and Corn dive into the history and future of operating systems, asking why we haven't yet moved to a graph-based model. They trace the lineage from Vannevar Bush's 1945 "Memex" concept to Microsoft's ambitious but failed WinFS project in the early 2000s. The duo discusses the technical hurdles of the past—like POSIX compatibility and hardware limitations—and why the rise of AI, vector databases, and tools like Obsidian suggest we are finally ready for a shift. Is the era of the file path ending? Join the conversation as we explore how semantic computing and modern storage architectures might finally let us navigate our data as a constellation of ideas rather than a stack of digital paper. It's a deep dive into the very ground we walk on in the digital world.</p> <h3>Show Notes</h3> <p>In the latest episode, hosts Herman Poppleberry and Corn take a deep dive into the foundational architecture of modern computing, questioning why, in 2026, users are still tethered to a hierarchical folder structure that dates back to the 1970s. The discussion centers on the tension between how computers store data and how the human brain actually processes information. While humans think through association—linking a smell to a memory or a person to a project—operating systems still force users to navigate "trees" of folders.</p> <p>### The Legacy of the Filing Cabinet Corn opens the discussion by pointing out that whether one uses Windows, macOS, or Linux, the experience is essentially the same: a digital recreation of a physical filing cabinet. This metaphor was popularized in the early 1980s by the Xerox Star and the original Macintosh to help office workers transition to computers. However, Herman notes that the vision for something better existed long before the first PC. He cites Vannevar Bush's 1945 essay, *As We May Think*, which proposed the "Memex." This theoretical device would have allowed users to create "associative trails" between documents, mimicking the web-like nature of human thought rather than a rigid index.</p> <p>### The WinFS Ambition and the POSIX Problem The conversation turns to the most significant attempt to modernize the file system: Microsoft's WinFS. Intended for the "Longhorn" project (which eventually became Windows Vista), WinFS aimed to replace the standard file system with a relational database. Instead of static files, every piece of data—an email, a contact, a photo—would be an "item" with rich, interconnected relationships.</p> <p>Herman explains that WinFS ultimately failed for two primary reasons: performance and compatibility. In 2004, the overhead of running a SQL-based engine beneath every file operation slowed hardware to a crawl. More importantly, the tech world is beholden to the POSIX (Portable Operating System Interface) standard. Most software is written to expect a "file path" (e.g., C:/Users/Documents). If an operating system replaces paths with a graph of nodes, every existing application breaks. This "legacy trap" has kept the hierarchical model alive long past its expiration date.</p> <p>### The Hardware Catch-Up Despite past failures, Herman argues that we have reached a technological tipping point. Modern hardware, specifically NVMe drives and Compute Express Link (CXL) architectures, can now handle the rapid, random-access patterns required by graph databases without the performance lag that doomed WinFS. Furthermore, the duo discusses how "tags" in modern operating systems act as a primitive bridge, allowing a file to exist in multiple "places" at once, though they remain an overlay on top of the traditional cabinet.</p> <p>### Semantic Computing and the Rise of AI One of the most compelling segments of the episode explores how AI is quietly building the graph-based future through the back door. Corn highlights the rise of "semantic computing," where AI models turn files into mathematical vectors. In this vector space, files aren't organized by where they are stored, but by what they *mean*.</p> <p>This shift is already visible in productivity tools like Obsidian, Roam Research, and Tana. These applications allow users to build personal knowledge graphs where notes are linked by context rather than location. Herman points out that Tana's "supertags" are perhaps the closest realization of the WinFS dream, operating at the application level to define relationships between atomic bits of data.</p> <p>### The User Interface Challenge: Navigation vs. Containment A significant portion of the discussion is dedicated to the psychological aspect of data organization. Corn expresses concern that a purely graph-based system might lead to a high "cognitive load." Hierarchies, while limited, provide a sense of containment and predictability—knowing exactly where a file "lives."</p> <p>Herman counters this by introducing the concept of "perspectives." In a future graph-based OS, a user wouldn't be forced to view a chaotic web of billions of nodes. Instead, the system would provide "slices" or "views" based on the current task. If a user is working on taxes, the system highlights those specific nodes and their connections, hiding irrelevant data like vacation photos.</p> <p>### Toward a Hybrid Future The episode concludes with the idea of a "poly-hierarchical" system. In this model, the graph is the "source of truth," but the system can generate a traditional folder view whenever the user needs the comfort of a hierarchy. Herman suggests that "saving a file" will eventually become more like "publishing a post," where the act of saving involves the system automatically identifying and creating links to relevant people, dates, and projects.</p> <p>While the "filing cabinet" has served us for forty years, Corn and Herman agree that the combination of AI-driven semantic understanding and high-performance hardware is finally making the dream of an associative, graph-based operating system a reality. The transition may be slow due to legacy software, but the era of the rigid folder is finally nearing its end.</p> <p>Listen online: <a href="https://myweirdprompts.com/episode/graph-based-operating-systems">https://myweirdprompts.com/episode/graph-based-operating-systems</a></p>