#1194 2001 · Ordnance Survey · Cartography / government mapping
Ordnance Survey proved map copying by matching a cartographer's habits, not the roads
问题
The AA was reproducing OS mapping, but facts like roads and place names cannot themselves be copyrighted
背景
Ordnance Survey, Britain's national mapping agency, suspected for years that the Automobile Association was building its road atlases and touring maps by tracing OS survey data rather than surveying independently. The dispute first surfaced in court in 1996 and reached back to allegations dating to 1990, but proving it legally was harder than it looked: a road's location, a village's name, a river's course are facts, and facts are not protected by copyright. OS could not win simply by showing that AA's maps and its own agreed on where things were — every accurate map of the same country necessarily would.
What copyright does protect is the expressive choices a cartographer makes rendering those facts: how a bend in a minor road is smoothed, where a place-name label is nudged to avoid clutter, which symbol conventions are used for a footpath versus a bridleway. OS built its case by comparing its own maps against AA's line by line for exactly these arbitrary stylistic habits — idiosyncrasies with no geographic reason to appear in two independently drawn maps, but every reason to appear if one had simply been traced from the other.
换别人会怎么做
The instinct was to argue the maps matched — but a court could simply respond that both maps were drawing the same real country, so of course the roads and towns lined up; matching facts proves nothing when the facts themselves aren't owned by anyone, and any cartographer redrawing Britain from scratch would land on largely the same reality.
他们看到了什么
Two teams surveying the same terrain draw the same roads by necessity, but won't coincidentally curve a bend the same idiosyncratic way. So the meaningless stylistic details, not the true ones, carried the proof.
那一手
OS termed the technique 'fingerprinting': cross-referencing dozens of small, functionally meaningless cartographic choices — a curve's exact smoothing, a label's exact offset — that a cartographer makes idiosyncratically but consistently. Because these choices carry no geographic information, two cartographers surveying the same real terrain from scratch essentially never reproduce them by chance, so when AA's output matched OS's fingerprint again and again across dozens of maps, it proved tracing rather than independent survey.
为什么管用
Stylistic fingerprinting exploits the fact that arbitrary choices have no external reference to converge on independently — unlike a road's real location, a label's exact pixel offset has no 'correct' answer that two cartographers would both discover by surveying honestly. Any match on these irrelevant details is therefore evidence of one specific causal path: copying. The mechanism needs no confession, no leaked internal files and no cooperating witness — the copied product carries the proof of its own origin baked directly into its rendering.
值了多少
AA paid OS a £20 million settlement in 2001, after 26 AA maps were found to be direct or substantial copies of OS mapping.
什么时候会失灵
It only works if the original genuinely contains enough idiosyncratic, non-functional choices to fingerprint, and if the copier reproduces them faithfully rather than redrawing over the traced base — a sufficiently careful copier who alters styling while keeping the underlying facts leaves no fingerprint at all, since facts alone remain legally unprotected.
后来呢
Ordnance Survey's fingerprinting method became the standard evidentiary approach across the mapmaking industry for proving copying of factual data that copyright law does not otherwise protect, cited in cartographic infringement cases well beyond Britain.
资料来源
- [1]AA copyright breach settlement (FOI231143)Ordnance Survey (UK government), 2023ordnancesurvey.co.uk
- [2]AA loses way in map disputeOsborne Clarke, MarketingLaw, 2001marketinglaw.osborneclarke.com