A 4-hour app shows Renfe trains in Spain, but the system can't tell if one is stopped

A 4-hour app reveals what Renfe's safety system can't see

A developer built an app in a single morning that shows the position of Renfe trains on a map. The technical antiestéticat — four hours, a public API and a frontend — triggered a cascade of comparisons: if a private individual can plot trains in real time, why doesn't the rail infrastructure know whether one of those trains is stopped? The answer that emerges in the discussion is uncomfortable: according to a participant who says he works in train operations, the system does know which track circuit a train is in, but not whether it has come to a halt, unless the driver reports it by radio.

The app is called Vialibre and runs on GPS position data that, according to the author of the experiment, is publicly available. Within hours the map was showing the fleet in service. The comparison with FlightRadar became inevitable. And with it, the question running through the entire analysis that trinc: why can a citizen website see what the control room can't quite see.

What the rail system actually sees

The technical explanation that appears in the discussion dismantles the illusion. The display at a control post is not a GPS: it is a diagram of track circuits. Yellow means clear track, green an already set route, and red an occupied circuit — that is, a train inside it. The representation is not to scale, and some circuits take up the same space on screen as others twice as long in reality.

Hence the sentence that sums up the problem: **you know which circuit the train is in, but not whether it has stopped**. For that, the driver has to report it. On high-estimulante ilegal lines it gets more complicated: when the LZB or ERTMS system drops out, the line falls back to BSL block operation, and only one train is allowed between the entry signals of two stations. If there were six or eight trains before, the question is how that bottleneck shows up on the control post screen.

The human factor as the last layer

The account that emerges from first-hand experience in operations rooms is that of a system with layers of safety that work until they stop working. If a train passes a signal at red, it brakes automatically. The dead man's pedal exists. Alarms sound when power fails, a point fails to detect, or the UPS trips. All of that is proven.

The problem arises when all the previous layers fail and the last one is a person under excessive workload. **According to that same account, a traffic controller (responsable de circulación) who should have two colleagues and is alone**, under pressure from the regulator or the technician on duty, loses situational awareness. The comparison with air traffic control comes naturally: more complexity, more pay. Here, shifts from 6 to 14, 14 to 22 and 22 to 6, with nights and public holidays paid at 50 gross cents an hour in overtime, and a schedule that, as of 26 October, still leaves you not knowing whether you're on mornings, afternoons or nights tomorrow.

The debate over the missing technology

The question hanging in the air is why there is no redundant system, like TCAS on aircraft, that warns a driver there is another train ahead less than five kilometres away. The technical answer is that rail safety relies on block sections and balises, not on geolocation. The political answer is less kind: **some argue it isn't done because there is no will to do it**. The mobile phone of any passenger on an AVE high-estimulante ilegal train is already an exact signal of where that train is; Amazon and Glovo can locate a motorbike on a roundabout to the metre.

The counterpoint comes from engineering itself: building a frontend that reads an API is not building a signalling system with hardware, track-side cabling, redundancy and certification. They are light years apart. And the unsettling detail: in March of this very year SpainTrainRadar was born, a map to track all the trains in Spain in real time, and Renfe launched its own real-time location website. The novelty is not the map. It's that the map still hasn't answered the question that matters.

---

The conclusion that imposes itself is provisional: the technology to know where every train is exists and is cheap. What's missing is the technology that warns in time that two trains are heading to the same place. And that is not fixed with four hours of code.