The Architecture of Memory: Eco, Calvino, Borges, and the Digital Archive

Last week, I pursued the general hype and went to our local theatre to see a contemporary Opera inspired by Umberto Eco’s Name of the Rose, one of my favourite books of all time. And though it was mostly inspired by the 1986 movie, it captured many intellectual layers of semiology and historical geekness Eco infused in his masterpiece. From a visual point of view, te Opera was stunning and, while I was sitting there basking in the theatricality of the scenes and the magical environment, it inspired me a reflection on what literature can teach us when it comes to managing knowledge through our digital archives. Maybe I need a vacation.

Anyways, bear with me, if you will. This is dedicated to all the CDE managers out there, and to everyone who agreed with my last piece.

1. Introduction

1.1. Memory as Architecture: Mapping the Invisible

There are places that exist only in our minds, yet feel more concrete than the rooms we wake up in. A hexagonal chamber filled with endless books. A city suspended between possibility and forgetting. A library protected by fire, code, or silence. Literature has long been obsessed with memory—not the tidy kind that fits on a flash drive, but the sprawling, contradictory, dreamlike variety that defies deletion. In the pages of Jorge Luis Borges, Italo Calvino, Umberto Eco whose Name of the Rose inspired this article—and their spiritual successors—memory takes architectural form: structured yet fluid, symbolic yet functional, sacred yet unstable.

What do those fictional architectures of memory reveal about our own very real, very digital attempts to remember everything? What happens when we take the infinite library of Borges and run it through an algorithm? Or ask Calvino’s Invisible Cities to become metadata categories? In other words: what can fiction teach our machines?

As we build data lakes and knowledge graphs, as we train AI models on the entire known internet and store backups in salt mines and lunar capsules, we might pause and ask: are we remembering wisely—or merely hoarding? Are we building temples of knowledge—or just very elegant labyrinths with no exits?

This is not an elegy for lost libraries (although a few might sneak in along the way). Nor is it a technical treatise. Rather, it is an invitation to use fantastic literature as it was meant to be: as a means of reflection to see our digital archives not as a database, but as a dreamscape. We’ll have Calvino’s poetic constraints conversing with cloud storage schemas, and Borges’ infinite permutations brushing shoulders with metadata taxonomies. This is what happens when I go to the theatre. Bear with me.

1.2. From Codex to Cloud: Why Fictional Libraries Still Matter

In a time when every byte of information risks becoming both permanent and meaningless, the imaginative archives of fiction become more than just metaphors—they become warnings, talismans, and perhaps even blueprints. We live amidst the illusion of total memory: every photograph backed up, every message archived, every document versioned into eternity. And yet, meaning dissolves faster than ever in the ever-growing swell of our digital hoards.

Enter the fictional library.

Umberto Eco’s Name of the Rose offers a vision of the archive as a theological labyrinth, where knowledge is deemed dangerous precisely because it is powerful. Knowledge has always been contested, and controlled through encoding. The library’s inner sanctum is both a prison and a puzzle—a reminder that curation is power, and that storage without access is merely containment. Today’s cloud architectures may lack the scent of old parchment, but they share the same double-edged seduction: they promise access, while quietly obscuring meaning through volume, structure, and complexity. Or through an undecodable naming convention.

Robin Sloan’s debut novel Mr. Penumbra’s 24-Hour Bookstore brings this tension into the 21st century. His bookshelves hide a secret society obsessed with decoding a hidden message that may never exist. Digital tools—Google, OCR, data visualisation—become allies, but they never replace the aura of the archive. Sloan’s world suggests that even in the age of algorithms, the archive courts mystery. It’s not enough to know something is stored. We need to believe there is something worth finding.

And then there is Carlos Ruiz Zafón, whose Cemetery of Forgotten Books gives us one of the most emotionally charged depictions of archival memory in modern fiction. In this hidden repository of forsaken volumes, each book is bound to a soul, a history, a secret. The library becomes a spiritual interface—a place where memory is not data, but destiny. How often do our cloud drives, with their folders named “Final_v3_REALLYFINAL”, achieve anything close to this? And all we wanted was a “unified source of truth”. That didn’t seem like asking much.

Cloud archives are, in many ways, the shadow libraries of our age. They are modular, redundant, always accessible, yet strangely elusive. They promise immortality of information, but not intimacy with it. We back up not to remember, but to insure. We search not to understand, but to locate. They are structured not around narrative, but around retrievability, scalability, cost-per-gigabyte.

In contrast, fictional libraries challenge us to consider the quality of memory, not just its quantity. They ask: what if memory has shape? What if it has intent? What if forgetting—judicious, sacred forgetting—is as vital to knowledge as infinite recall? Books in these stories are not passive vessels. They seduce, betray, whisper, even as they burn. They are actors in the drama of knowledge. They remind us that the architecture of memory is not neutral: it directs us, confines us, even deceives us. A search engine may return every instance of a word, but it cannot explain why it matters.

To build better digital memory systems, we might need to look less at how much we store and more at how we choose. And for that, these fictional libraries may offer more wisdom than any technical whitepaper. They teach us that the archive is never just a place—it is a ritual, a relationship, a risk.

Welcome, then, to the architecture of memory. You’ll need no password, just a willingness to get a little lost.

2. Philosophical and Informational Foundations

Long before server farms blinked into existence, before hard drives and cloud backups, memory already had its architects. The Greeks and Romans rethors built palaces not of stone, but of thought—vast mental cathedrals populated by vivid images, placed along imagined corridors. This was the method of loci, the art of memory, the original cloud infrastructure suspended not in the sky but in the mind.

2.1. Mnemosyne’s Architecture: From Memory Palaces to Metadata

In Greek mythology, Mnemosyne, goddess of memory, was not merely a custodian of facts; she was the mother of the Muses. Memory, then, was not static—it was generative, the source of art, story, and song. Her architecture was fluid yet stable, an invisible scaffold on which the richness of human culture could be suspended. The memory palace technique was born within this cultural context, enriched by the Romans’ practical edge on politics, and formalised this principle. Thinkers like Cicero, Quintilian, and later medieval scholars described how a skilled practitioner could navigate through an imaginary building, each chamber holding a vivid, often surreal image linked to a concept or piece of information. The structure itself—the spatial organisation—was crucial. It provided retrieval hooks: a way for the mind to “walk” back through its own archive and find, not just data, but context, meaning, and associative richness.

In our own time, metadata plays a similar role. The “where” and “how” of information are almost as important as the information itself. Without tags, without relational structures, data floats unmoored, inaccessible. Metadata, like the mnemonic architecture of old, provides paths through chaos. It builds corridors where otherwise there would only be mist.

Yet there is a telling difference.

Where ancient memory palaces celebrated symbolic density, the metadata schemas of our age prefer minimalism. That’s the illuminists’ legacy, baby, and our ever-growing obsession for finding the perfect taxonomy, a way to classify stuff, One Metadatum to Rule Them All. A field labelled “Date Modified” or “Author” often lacks the imaginative punch of a flaming phoenix perched on a marble column, signalling the orator to recall the story of Icarus, but does it have to? We have gained precision but lost enchantment. The architecture of memory has become efficient to machines, but not necessarily memorable for people.

And herein lies a lesson: a machine-readable structure is not enough, and a retrieval system that forgets the texture of memory risks becoming nothing more than an index without a story. Codes need to be human-readable, and archive rules must stay within the human ability to comprehend and navigate them because they provide meaning within themselves.

2.2. Ars Memoriae and the Roots of Information Storage

The Ars Memoriae—the art of memory—was not just a set of mnemonic tricks; it was considered, in the classical and medieval mind, a foundational technology of thought itself. To know how to remember was to know how to think. The ordering of knowledge was not arbitrary; it was a spiritual exercise, a discipline of the mind.

In the libraries of the Middle Ages, knowledge was organised according to practical and theological needs, driven by function and religious utility. There was a strong medieval tendency to see the cosmos, knowledge, and architecture as interconnected and reflective of divine order, sure, but stuff was organised based on correspondences, not Dewey decimals. Finding one precise piece of information was as important as encouraging human curiosity to peruse by association. While this approach has fundamentally changed, make no mistake: to store knowledge is always to interpret it, to decide what is worth remembering, how it should be linked, and who should have access.

As Mario Carpo has always stressed in his foundational texts (I talked about them extensively, peruse the archive), modernity brings a dangerous inversion: where the old masters of memory sought compression through vividness—an entire story remembered by a single icon—modern systems sought expansion through redundancy as soon as technology allowed. Backup upon backup, copy upon copy, mirror upon mirror. We trust in storage, but not necessarily in retrieval—or understanding.

2.3. Redundancy, Noise, and Meaning in Systems

If the ancients taught us about memory palaces, the moderns built transmission towers. With the 20th century came a new way of thinking about memory—not as a poetic structure for internal reflection, but as a problem of information transfer. In 1948, Claude Shannon, often called the father of information theory, published a seminal paper proposing that communication could be boiled down to binary signals, redundancy, and noise.

Shannon’s genius lay in separating meaning from transmission. For the engineer, it mattered not what the message said, but whether it could survive distortion. Redundancy, once seen as a flaw or inefficiency, became an essential tool: a way of ensuring that a message arrived intact, even across chaotic channels. In Shannon’s world, duplication wasn’t waste—it was armour against entropy.

Now consider Jorge Luis Borges, writing just a few years earlier. In The Library of Babel—written in 1941, around the same years Shannon was developing his theories—Borges conjured a universe composed entirely of books, every possible permutation of letters across a fixed length. Within that infinity, every possible truth exists—alongside every lie, every mistake, every meaningless jumble of symbols. Meaning is swallowed by abundance. Redundancy, instead of protecting information, obliterates it.

Where Shannon saw redundancy as a way to guard against noise, Borges saw it as a source of existential despair. Too many copies, too many variations, too much noise—and the fragile island of meaning sinks beneath a tide of gibberish. The contrast is striking. In Shannon’s mathematical landscape, redundancy is an engineering solution. In Borges’ literary cosmos, redundancy is an ontological nightmare.

And yet, both were grappling with the same essential problem we’re facing today in our own digital archives: how does information survive? How do we build systems—whether telegraphs or libraries or cloud archives—that protect the fragile signal against the inevitable decay of time, error, and misunderstanding?

Today’s digital archives embody both visions. Metadata tags, error-correcting codes, and blockchain ledgers all apply Shannon’s principles of resilient transmission. Meanwhile, the sheer volume of digital data—the endless backups, copies, revisions—threatens to recreate Borges’ infinite library, where the presence of everything ensures the retrieval of nothing.

In designing memory systems, we must balance these twin poles:

• Redundancy as shield versus redundancy as suffocation.
• Storage as salvation of meaning versus storage as a pit of oblivion.

Perhaps the perfect archive must be neither a sterile broadcast nor an endless labyrinth, but something else—a curated garden where meaning can survive without being drowned. The architect must ask not simply “Can I store this?”, but “Should I?”. And more daring still: “How will it live on, during the project?”

3. The Construct of the Digital Archive

If the memory palace was a guided tour and the Library of Babel a blind wandering, the digital archive needs to support another activity: the autonomous search. Tagging and indexing are its navigational tools—the markers that help retrieve the right piece of information from an expanding sea of data.

Every time we label a file, add metadata to an image, or assign a keyword to a specification, we are not simply organising; we are shaping the pathways of discovery. Taxonomies, categories, and hierarchies are forms of storytelling in their own right, crafting routes that guide future users toward what they seek—or toward what we have anticipated they might seek.

Tagging is rarely neutral. Every keyword implies a perspective, a choice about how a piece of information fits into a broader structure of meaning. Is a document about “architecture” or “urbanism”? Should a photograph be categorised under “construction detail” or “construction flaw”? These decisions influence not only retrieval but also interpretation.

In fiction, Mr. Penumbra’s 24-Hour Bookstore stages this beautifully: Sloan’s protagonist uses both human intuition and algorithmic power to uncover hidden patterns—not in the books’ contents, but in the way they are organised. In doing so, the novel reminds us that how we search often determines what we find, and that the architecture of discovery is as significant as the artefacts themselves.

In our own systems, every search engine result, every database query, reflects a hidden narrative about what has been stored, indexed, and linked. If we didn’t define the system’s architecture with this in mind, thd structure and narrative is totally unconscious, like that of a chicken piling LEGO bricks with its beak. The archive is not a passive container; it is a constructed environment, where meaning emerges not just from content, but from connections.

3.1. Curating vs. Hoarding: How Stories Shape Data

Not all who gather are curators. Some simply hoard.

In the digital archive, the line between collection and clutter grows perilously thin. The temptation to store everything—every draft, every redundant photo, every forgotten email chain—is fueled less by a desire for meaning than by a fear of loss or, worst, the need to get out of hypothetical, future trouble. Every “just in case” file saved, every “final_final_v2” document tucked away in a forgotten folder, is a small act of anxiety: a hedge against forgetting, a talisman against the impermanence of our digital existence.

But accumulation without discernment leads to a slow suffocation of meaning. The more we gather indiscriminately, the more invisible each individual item becomes. Data piles up, thick and unstructured, until the archive becomes a graveyard—not of forgotten things, but of things that were never properly remembered in the first place.

Curation, by contrast, is an act of intentional storytelling. It is an editorial gesture, an active decision to shape memory, not merely to preserve it. To curate is to ask:

• What matters here?
• What deserves to endure?
• What context does this memory need in order to be understood later?

A curated archive says: This matters.
A hoarded archive begrudgingly mutters: Maybe something here will matter someday.

Carlos Ruiz Zafón’s The Shadow of the Wind captures this distinction vividly. In the Cemetery of Forgotten Books, entrance is not automatic. One must choose a single volume to “adopt,” to protect it from oblivion. Even in a library designed for loss, there is selection, stewardship, responsibility. Memory, in Zafón’s world, is not passive; it demands engagement. The books are not warehoused in bulk. They are given meaning by the fact that someone—reader, caretaker, future discoverer—cares enough to choose them.

In our digital systems, we too must learn to choose. And, ideally, not perish.

We must resist the easy drift toward limitless accumulation and embrace the harder, slower work of storytelling through selection. Curation requires time, attention, and a willingness to say no. Not everything needs to be saved. Not every backup needs to be infinite.

Moreover, how we curate matters. Contextual metadata, careful categorization, meaningful tags, human-readable notes—these are the modern equivalent of illuminating a manuscript, adding a marginalia, carefully picking a Drop cap that drives home a message, or binding a codex with care. Without them, files may survive physically but die cognitively. Future readers (including ourselves) will find only shapeless noise where once there could have been living memory.

I know I said I wasn’t going to expand on this, but then again… the more I think about it, the more I feel gardening is a better metaphor than warehousing. The gardener does not try to grow everything at once. They prune, they weed, they plan for cycles of bloom and dormancy. In the same way, the storage portion of a Common Data Environment must be alive, evolving with its purpose and audience, allowing for both permanence and graceful forgetting.

Curation is not an act of nostalgia—it is a commitment to future meaning. It recognises that memory is finite, not infinite, and that its value lies precisely in its limitation: in what is remembered instead of everything else. In the design of digital archives, as in the stewardship of project memory, the ultimate question is not how much we can store, but what stories we choose to tell with what we keep.

Add another aspect on top of this. We act as if our archives can hold an infinite amount of data, but is that actually true and what’s the environmental impact of this belief? While data centres currently account for about 1% of global energy use and contribute approximately 2–4% of global carbon emissions, projections suggest that their usage of global electricity could rise from 3% to 10% by 2030, driven by increasing demand from AI and machine learning workloads.

So, what different approach might we adopt?

3.2. The Archive as Game: Oulipo, Constraints, and Design

Constraints breed creativity. The Oulipo group—Ouvroir de littérature potentielle, or “Workshop for Potential Literature”—embraced this principle with almost mathematical fervor. Writers like Raymond Queneau, Georges Perec (see here), and Italo Calvino experimented with formal limits: novels written without the letter “e” (La Disparition), stories generated by algorithmic permutations, poems unfolding along combinatorial grids. Constraints, for the Oulipo, were not cages but engines. By setting boundaries—rules, limits, formulas—they unlocked new territories of expression that would otherwise remain hidden. Freedom was not the absence of limits; it was the richness found within them.

In the world of digital archives, this lesson is just as vital.
Without constraints, archives grow wild and unstructured, threatening to drown users in undifferentiated bulk. Every file saved without a schema, every dataset indexed without a taxonomy, every uncontrolled vocabulary breeds confusion. Discovery becomes harder, not easier, as options proliferate without guidance. And while we are listening to the courting songs of AI, promising to create flexible scaffolds of meaning around our digital mess, what are we prepared to lose when we relinquish curation? Have we really given it enough thought? Can we train a machine to do something we don’t actually understand?

Constraints in archive design—through metadata standards, controlled vocabularies, access protocols, modular storage rules—turn accumulation into intelligible structure. They guide users along designed paths without scripting every move, and offer navigational scaffolding without dictating the journey’s meaning. That’s the work of an architect and a manager of a Common Data Environment. Not assigning reading vs writing permits in folders.

In this sense, Calvino’s Invisible Cities offers a profound model.
The novel’s intricate architecture—a precisely ordered series of dialogues, cities grouped and repeated across thematic axes—does not confine the reader’s imagination. Instead, it amplifies it. The strict formal structure creates resonances, echoes, and hidden symmetries that invite readers to participate, to notice patterns, to build their own internal maps of Marco Polo’s shifting cities, and to read the novel in a non-linear fashion, pursuing those patterns.

An effective digital archive might work the same way.

Rather than presenting data as a flat inventory, it can be structured like a game board, a matrix of possibilities where users follow relational references, encounter unexpected juxtapositions, and assemble meaning through interaction. Constraints, far from being limitations, shape user experience and define the grammar of interaction. Faceted search, curated pathways, dynamically generated clusters of related materials—all these design choices echo Oulipian principles, turning the archive from a static warehouse into an active exploration, and it’s entirely within the possibility of our current technology. We’re just too busy renaming files to pursue it.

Moreover, constraints act as memory’s protectors. By deciding what types of data are admissible, how they must be described, and in what relationships they must stand, the manager preserves the possibility of future understanding. An archive without structure risks becoming, once again, Borges’ Library of Babel—vast, infinite, but ultimately unreadable.

Thus, the future of the Common Data Environment might demand a shift: from a container of information to a designed experience, a balance between rigidity and openness, order and imagination. The most enduring memory systems will not be those that store the most, but those that invite the richest and meaningful exploration.

4. Designing the Archive of the Future

Now, what do I mean when I say that the technology exists? Let’s take a closer look.

4.1. The Archive Status. A.k.a. Can a Library Be Infinite and Useful?

The dream of the infinite library persists, and I won’t be the one to banish it from your imagination. It flickers in Borges’ Library of Babel, resurfaces in the boundless databases of Google, and echoes in every ambitious digital repository that promises to preserve everything. In theory, an infinite archive offers limitless discovery: every fact, every version, every forgotten trace made eternally retrievable. As we have seen, discovery doesn’t seem possible without structure. And infinity with structure is a paradox.
As Borges understood all too well, a library containing every book ever written—including every possible combination of letters—is not a paradise but a nightmare.

A useful archive, no matter how vast, must therefore impose limits of intelligibility. It must structure infinity into zones of relevance. It must prune, prioritize, and organize—not to diminish knowledge, but to make it reachable.

But I can’t decide to delete stuff.

I know, darling, I know.

One possible compromise, that doesn’t address the environmental issue but at least tries to recuperate meaning for our Common Data Environment, is this: an infinite archive is possible only if it is also finite at the point of encounter, offering the right document at the right time to the right seeker. Otherwise, the user is left adrift, shipwrecked on the endless shores of possibility.

That’s why we have the Common Data Environment’s archive status, right? To keep out of the way whatever’s no longer relevant.

And yet, the absence of arrows in that well-known scheme of the international norm speaks volumes: how do we get there? DO WE get there? And why? With whose clearances? Under which circumstances? An archive that’s really dynamic will not merely scale that section horizontally by adding more data or by keeping piling superseded versions upon superseded versions. It will scale vertically by adding layers of meaning, filters of interpretation, and maps of guidance to retrieve how we got to the current point in the project and—maybe—what went wrong. Usefulness is not the enemy of infinity; it is its steward.

4.2. The Classification Systems. A.k.a. From Dewey to Deep Learning

Classification systems have always reflected philosophies of order. I speak at length about it at the beginning of pretty much every lesson on Levels of Development and Work Breakdown Structures, involving people like Raymond Lull, Pierre de la Ramée, Conrad Gessner and Paul Otlet (see here).

Melvil Dewey’s Decimal System, born in the 19th century, is one of those milestones that divided human knowledge into rigid, numbered categories—a structure that, despite its endurance, reveals a very particular worldview: hierarchical, Western-centric, and increasingly outdated.

Librarians everywhere, please don’t hate me. For everybody else, the Dewey Decimal System remains the most widely used library classification scheme globally, used in over 200,000 libraries in at least 135 countries, and I know it’s being continuously updated to better accommodate new fields and perspectives. I know you’re trying.

Today’s information architects face a new challenge. The rise of deep learning and AI-driven retrieval threatens to make traditional categories obsolete and, as we have seen, that might not be a bad thing. In theory. Neural networks group information not by predefined taxonomies, but by fluid patterns of association. A photograph, a text snippet, and a drawing may cluster together not because they belong to the same “subject,” but because some latent vector hints at a shared essence.

Those systems, however, require designing, else the risk of delegating to the spontaneous definition of meaning inherent in neural networks. Also, designing systems isn’t enough: systems will have to be used by people. These new tools demand a new sensibility—a poetics of interface—that moves beyond rigid folders and static indexes. Interfaces must invite exploration, not just retrieval. They must allow users to connect across categories, to glimpse unexpected affinities. They must offer not just answers, but encounters with knowledge. Imagine, if you will, an archive where searching for “memory” brings you not only scientific papers, but also paintings, personal letters, city maps, and forgotten myths. Where meaning is discovered relationally, dynamically, through an interplay of forms. This isn’t any different than the concept of the Medieval library, if you’ve been paying attention.

Conclusion: Literary Templates for the Ethical Archive. Ak.a. What Fiction Teaches Tech

Fantastic and sci-fi literature, far from being a relic, offers powerful templates for how we might design and govern our technological solutions.

In The Name of the Rose, Eco warns us that access is power, and that curating knowledge carries ethical stakes.
In Mr. Penumbra’s 24-Hour Bookstore, Sloan shows how digital tools can extend, but not replace, the thrill of human curiosity.
In The Shadow of the Wind, Zafón reminds us that archives are repositories not just of data, but of personal histories and emotional debts.

From these literary visions, we can extract guiding principles:

1. Curation matters as much as storage. Not everything needs to be saved; some things must be chosen.
2. Context must be preserved. A document divorced from its origins is a memory without meaning.
3. Discovery should remain an adventure. Archives must nurture curiosity, not just efficiency.
4. Access must be equitable and ethical. Not all memories are harmless; not all knowledge should be stripped of its consequences.

Fiction shows us that an archive is never neutral. It is a place where values are encoded into every organisational choice, every retrieval algorithm, every forgotten folder. How’s your Common Data Environment structured these days?

---

Bookish References

• The Library of Babel by Jorge Luis Borges
• Invisible Cities by Italo Calvino
• The Name of the Rose by Umberto Eco
• Mr. Penumbra’s 24-Hour Bookstore by Robin Sloan
• The Shadow of the Wind by Carlos Ruiz Zafón
• The Archivist by Martha Cooley
• The Library at Mount Char by Scott Hawkins