The Librarian's Two Catalogues: On the Deliberate Order of the Hierarchical and Relational
Every library needs a catalogue. It’s the fundamental promise a space of knowledge makes to its visitors: that what they seek can be found. In the architecture of a website, this catalogue is our site structure, and the paths through it are our internal links. But there is a quiet, profound schism in how we choose to build this system of discovery. It is the difference between the librarian who files by a strict, hierarchical Dewey Decimal system and the one who creates a web of thematic connections between disparate ideas.
The first approach is one of clear, deliberate hierarchy. It is the architect’s blueprint. Every URL has a designated place, a parent, and perhaps children. A visitor navigates from the broad category down to the specific page, like moving from ‘History’ to ‘20th Century’ to ‘World War II.’ The logic is linear and reassuring. It creates a sense of order and comprehensiveness. For the site owner, it is a model of control, making it clear what belongs where and what the journey of a typical user should be. The internal links here are like signposts on a well-marked trail, guiding you efficiently from the trailhead to the summit without deviation.
The second method is more akin to a scholar’s personal index cards, scattered across a desk and connected by threads of association. This is the relational, or thematic, approach. Here, the primary navigational force isn't a top-down hierarchy but a network of contextual relationships. A page about a specific painter might link directly to a page about their artistic movement, their influences, their hometown, and the techniques they pioneered. The structure is flatter, less about lineage and more about affinity. The visitor isn't guided along a predetermined path but is invited to follow their curiosity through a landscape of ideas.
Neither approach is inherently superior; they serve different purposes and answer different needs. The hierarchical model excels at scalability and user orientation. It is invaluable for large, complex sites like e-commerce platforms or documentation libraries where users often know what they are looking for and simply need the most efficient route to find it. Its weakness is its rigidity. It can feel sterile, and it often fails to surface deep, valuable content that doesn't sit near the top of the tree.
The relational model, by contrast, is a powerful engine for discovery and engagement. It mimics the associative nature of human thought, encouraging deeper exploration and longer sessions. It is the soul of a blog, an online magazine, or any knowledge base where the connections between topics are as important as the topics themselves. Its peril is the potential for chaos. Without some underlying hierarchical logic, it can become a tangled web that is difficult for both users and search engines to comprehend.
The most thoughtful web architects, like the most adept librarians, do not choose one catalogue over the other. They maintain both. They build a solid, logical hierarchy—the official filing system—to provide structure and ensure nothing is lost. Then, within that framework, they weave a rich tapestry of relational links, creating pathways of meaning that transcend the official taxonomy. This dual system respects the user who arrives with a specific purpose and delights the one who came to wander, proving that the true art of structure lies not in adherence to a single philosophy, but in building a home that welcomes every kind of seeker.
Notes & further reading
A few pages I came back to while writing this: