29 November 2025

We often describe software obsolescence as though it were the inevitable by-product of technological progress—an unfortunate side-effect of “innovation.” It isn’t that simple. The cycle of updates that renders yesterday’s files unreadable is not merely accidental; it is also sustained by commercial models that benefit from continual replacement.

As I prepare for my SAHANZ presentation next week, I’ve been revisiting the deeper structural patterns behind this phenomenon. Over the past two decades, architects have gradually outsourced their technical memory. Like many creative professionals, they now rely on corporations whose business models depend on continual upgrade cycles and, increasingly, subscription access. This dynamic parallels what Nick Srnicek describes in Platform Capitalism.¹ Economic value is increasingly generated through the capture and control of digital infrastructures rather than simply through the improvement of tools.

In this environment, access replaces ownership as the commodity. Subscriptions replace licences. Continuity becomes a cost rather than an asset.

This is why Jussi Parikka’s term “zombie media” is so apt.² It describes technologies that appear to persist: files still exist on a hard drive or server, yet can no longer perform the functions for which they were created. As data, they remain visible; as working artefacts, they are dead.

Architecture is increasingly haunted by these undead files—models, drawings, scripts and renders sitting inert in digital storage, unopenable without the long-gone software environments that created them.

A recent experience made this problem very clear. I tried to open a set of Autodesk files from 2004. They had been created using fully licensed software and were intended to be opened within the same software ecosystem. Instead, the application returned a “fatal error”—not something you want to see on your computer.

I contacted Autodesk support to ask why I could not access files created in their own archival formats. The reply was: “We apologize for the inconvenience.”

No legacy viewer. No compatibility mode. No sanctioned pathway for recovering the older work.

This is not simply a matter of technical impossibility. As Wendy Chun argues in Programmed Visions,³ digital systems are not inherently ephemeral; the infrastructures and policies governing them help make them so. The question is not whether backwards access can be supported, but whether the economic structures surrounding software provide sufficient incentive to support it.

McKenzie Wark’s description of the vectoralist class is useful here.⁴ Value increasingly arises from controlling access to information rather than from producing the information itself. Architectural memory is therefore becoming governed not primarily by architects, archivists or educators, but by corporate licensing systems and product cycles.

In this sense, digital obsolescence becomes part of a revenue model, and the cost is our collective intellectual labour. Architectural drawings, models, BIM files and visualisations are not incidental artefacts. They are among the core materials through which architecture is designed, taught, researched and understood. Yet they can become inaccessible the moment a platform deprioritises its past.

For a discipline grounded in documentation, lineage and precedent, this should be recognised as a serious threat. If we cannot open our own digital history, we cannot properly study it, critique it or preserve it.

This will be the core argument of my SAHANZ talk next week: digital architectural memory is being quietly reshaped by economic systems that privilege monetisation over preservation. If the cultural record of the digital turn is to survive, architecture needs to treat digital custodianship as a professional and institutional responsibility.

That means confronting these systems directly—contractually, professionally and institutionally. Otherwise, architectural heritage will continue to erode one corporate “We apologize for the inconvenience” at a time.

1. Nick Srnicek, Platform Capitalism, Reprinted, Theory Redux (Polity, 2019). ↩︎
2. Jussi Parikka, “Zombie Media in Leonardo,” Machinology, September 5, 2012. ↩︎
3. Wendy Hui Kyong Chun, Programmed Visions: Software and Memory (The MIT Press, 2011). ↩︎
4. McKenzie Wark, Capital Is Dead: Is This Something Worse?, Paperback edition (Verso, 2021).
   ↩︎