Istanbul's Digital Archive Crisis: The Key Decisions Ahead on Duplicate Image Replacement

Istanbul's municipal archivists have a problem that is quietly getting worse. Across at least three major public repositories — the Istanbul Metropolitan Municipality's digital library, the SALT Galata research centre on Bankalar Caddesi, and the Istanbul Archaeology Museums' photographic collection in Sultanahmet — duplicate and mislabelled images have accumulated over decades of piecemeal digitisation. The question now is not whether to clean house, but who decides which version of a photograph survives, and which one goes.

The timing matters. Istanbul's broader urban documentation effort has accelerated sharply since the February 2023 Kahramanmaraş earthquakes, which destroyed irreplaceable physical archives in affected provinces and forced heritage institutions here to confront how fragile analogue records really are. The Istanbul Metropolitan Municipality launched an expanded digitisation push in the months after that disaster, pulling in tens of thousands of additional images. Volume increased faster than quality control could follow.

The Technical and Political Tangle

Duplicate image replacement sounds like a database maintenance task. In practice it involves judgment calls that carry real consequences. A photograph of the Galata Bridge taken in 1955 might exist in four versions across different collections — each with slightly different metadata, different crop, different attribution. Choosing one as the canonical record and discarding the others means deciding whose institutional label takes precedence. At a moment when the CHP-run municipality and the AKP-aligned central government are contesting control over everything from budget allocations to construction permits along the Bosphorus shore, even archival decisions carry a political undercurrent.

SALT Galata, which holds one of the most significant collections of late Ottoman and early Republican-era urban photography in the country, has been developing its own deduplication protocols for several years. The institution uses perceptual hashing software to flag near-identical images, then routes flagged pairs to human reviewers. That process works within a single institution. The harder problem is cross-institutional deduplication — agreeing on a shared identifier standard that allows, say, the municipality's portal and the Atatürk Library database in Taksim to recognise that they hold the same photograph under different file names.

Turkey does not yet have a binding national standard for cultural heritage metadata interoperability, unlike the European Union's Europeana framework, which has been mandating common data models among member-state institutions since 2011. Istanbul's archivists are effectively building cross-institutional rules from scratch, institution by institution, agreement by agreement.

What the Next Six Months Will Decide

Several concrete decisions are now on the table. The Istanbul Metropolitan Municipality's cultural affairs directorate is expected to finalise its archive data governance framework before the end of 2026 — a document that will set retention rules for duplicate files and establish who has authority to mark an image as the primary record. Whether that framework will include binding cooperation clauses with non-municipal institutions like SALT or the private Yapı Kredi Culture Arts archive on İstiklal Caddesi remains unresolved.

Storage costs are part of the calculus too. High-resolution TIFF files — the standard for archival-quality scans — run to roughly 100 megabytes per image. A collection with 200,000 duplicate pairs is carrying around 20 terabytes of redundant data. At current cloud storage pricing for public-sector contracts in Turkey, that represents a meaningful and recurring line item at a time when municipal budgets are under strain from lira depreciation.

For researchers and heritage professionals working in the city, the practical advice right now is straightforward: download and locally preserve any archival images you are actively using for published work. Cross-institutional deduplication exercises, even well-managed ones, occasionally result in the accidental deletion of the higher-quality version. The Beşiktaş Municipality's smaller local archive underwent a deduplication exercise in late 2024 and spent several months reconstructing records that were removed in error.

The bigger stakes are not technical. Istanbul's visual history — of its neighbourhoods demolished for urban renewal, of communities displaced from Tarlabaşı or along the Bosphorus waterfront — exists in photographs that may survive in only one institutional copy. Getting the replacement logic wrong, even once, is permanent. The decisions made in the next six months will set the template for how the city manages that risk for the next generation.