DEV Community: Dietmar Schoder

What year is it?

Dietmar Schoder — Sat, 13 Jun 2026 13:12:17 +0000

What year is it?

How an inexplicable decision 2,000 years ago baked one calendar into every database on Earth

Since the widespread use of social media and the automation of most parts of our lives, we have reached a phase of humanity in which we constantly produce historical evidence. In comparison: we might have a handful of sources telling us about the most important events in a whole ancient kingdom over a decade. But today, we have not only data about a vast majority of all citizens, but also an enormous amount of data produced by many individuals themselves – data about things no historian would ever consider important enough to write into a history book.

By the structure alone of how databases, tables, and data fields are usually designed, all this data is fairly consistently chronologically ordered. If we want to reconstruct how an international chain of events unfolded, we can usually do that, because most dates are stored as UTC nowadays, or in a format that can be translated into UTC. And the Western dominion over computer systems and standards has led to the fact that the Gregorian calendar and the counting of years as "Common Era" (CE) is far more used than it would be culturally alone. Even the data of people who use, for example, the Islamic calendar, the Jewish calendar, or the Coptic calendar is turned into Gregorian UTC when they post anything on a Western social media platform, use an international bank for transactions, book an international flight, or buy anything on an international online marketplace.

But when did our globally dominant standard year counting actually start? What happened that made a random year into "year 1" for the whole world? And when did people begin dating events using this special counting system over all other previous, well-established systems?

To answer this, we need to understand how the ancient city of Rome fed its citizens, how a short series of unassuming individuals made – from around 30 CE onward – a chain of very unique decisions, each leading to the next, and how finally one ambitious king in Europe around 800 CE standardized our norm of counting the year forever, without ever knowing it.

The grain, the trade route, and two healers

Egypt had grain. The Romans needed bread. That had two consequences. First, the grain had to be transported on ships from Alexandria in Egypt during a long and safe journey along the coast to Italy. Second, this trading route had to be controlled by the Roman Empire to make sure no one could interrupt the trade. Eventually, a Roman puppet king named Herod Antipas ruled one region of Palestine – which lies on this trading route – then called Galilee. At that time, two separate healers in his kingdom gained a lot of popular attention. Herod Antipas had one of them killed, while the other went with a group of his followers north into the kingdom of Herod Antipas' brother, a tolerant man. There, in the villages near the capital – a place called Caesarea Philippi – the second healer could have stayed, healed people, and done as much good as he ever wanted.

But he did not stay there.

His name was Jesus of Nazareth, and he made a strange decision that no historian can explain to this day: he walked all the way south to Jerusalem in Judea, with the sudden intention and precise plan to overthrow the Roman Empire. His plan failed. The Roman prefect in Jerusalem executed Jesus. This inexplicable decision of one man – to try to bring the Roman Empire down – is the root cause of our modern calendar. Yet his action was seen as so unremarkable and common at the time that it left no entry in any history book of that era.

The first written hints

Forty to seventy years later, some people began to write the story down, but they had neither a specific date for Jesus' birth nor his death. Still, one of these authors wrote two crucial hints. He stated that these key events took place in the 15th year of the reign of Emperor Tiberius, and that Jesus was about 30 years old at that time. This was a very common way of dating events, because people back then knew how to count Roman regnal years, but did not need precise historical timelines in daily life.

By the time we reach 100 CE in modern calendar terms, all eyewitnesses were dead. Since then, no one has ever added any new evidence that would help us find a more precise dating of Jesus' life.

Sextus Julius Africanus – the first calculator

Around 200 CE, our next important character entered the stage. He was called Sextus Julius Africanus, and he had access to the text of Luke with its reference to Tiberius and Jesus' age of about 30 years. Sextus Julius Africanus was a well-traveled Roman citizen from Palestine who spent a lot of time in Alexandria and in Rome – the two opposite ends of the grain trading route mentioned above. Based on all the relevant knowledge he gathered, he was able to calculate the birth and death date of Jesus of Nazareth. Of course, there was no concept of our modern year counting back then, so he calculated it in terms of the Roman calendar system. His result was that Jesus must have been born in the 42nd year of Augustus' reign, and that he died in the year of the two consuls Gemini.

But Sextus Julius Africanus was most interested in the age of the entire world. In our modern counting of 221 CE, he defined – based on Jewish scripture and its interpretation – that the universe was 5,500 years old when Jesus was born. Much later, this anno mundi (year of the world) counting nearly became our modern standard.

Diocletian and the stable counting

In the meantime, in 284 CE, Emperor Diocletian started his reign, and as usual the common counting was reset to the "1st year of Diocletian" – as had been done many times with every new emperor. But Diocletian himself then established a rule of four people ruling the vast empire. As a result, the counting was not reset anymore after him. The Diocletian counting became a common standard that stayed long after his death.

Then around 400 CE, a man named Panodorus not only studied the calendar issue intensively in Alexandria (where he lived and had access to the famous library), but also did the heavy work of computing the timespan between the 42nd year of Augustus' reign and the 1st year of Diocletian's reign as 284 years.

Now, following all the years and numbers mentioned so far, at that point in time – around 410 CE – for the first time in recorded history, the birth of Jesus was reliably anchored to an absolute, well-established common calendar. Not only was Jesus placed at "284 before Diocletian", but also every date that could be expressed in the Diocletian counting became identifiable as Diocletian plus 284 years since the birth of Jesus. Still, no one began to use any dating method from this new angle – our modern angle.

The collapse of Rome and a monk named Dionysius

Then the Roman Empire in the West – in Rome itself – collapsed as a political power. In the aftermath, the bishops of Rome, still part of an international network of Christian communities, one after another walked into the political vacuum in the West. This brings us, in an astonishing turn of events, to a monk named Dionysius Exiguus, a Greco-Roman from the East who eventually worked for a series of bishops in Rome over decades. One day around 525 CE, he received a task: to provide new so-called Easter tables. He started with the existing ones, which would run out soon, and added a new extension – something like a modern Excel sheet.

But this monk did not like Diocletian. (Diocletian had been a fierce persecutor of Christians.) Therefore, in a fascinating turn of events, he decided to rearrange the "year" column of the table. He copied the old one, added the new years as new rows, and recomputed the Diocletian years into years since Jesus' birth. He renamed the header of the column Anno Domini – "the year of the Lord". Then he even sent the table to a monk in the East as an attachment to a letter. And this letter, written in 525 CE, he dated as 525 Anno Domini. It is the first letter in the history of humanity that was dated using our modern counting of years.

Then Dionysius died, and with him the whole new idea died.

The Venerable Bede – the unexplained switch

It took another 200 years for the idea to resurface. Some Anglo-Saxons had moved over into Britain, merged with the existing people there, and had become Christians. They built monasteries and by that joined the international network of "the Church". An ambitious example of a new monastery was built in Northumbria, and there another monk gained access to a very modern library filled with copies of books – especially from Alexandria, Eastern Rome, and Rome itself.

Around 700 CE, this new character entered the stage: the Venerable Bede. He, in this chain of events, made a crucial decision that no one can explain. He spent decades studying, describing, and fiercely debating how to correctly calculate the age of the universe. As a result, in 725 CE he published an encyclopedia explaining in depth why the world – the universe – must be exactly 3,952 years old. He also added a history of the world in which he dated all events based on this calculation of anno mundi.

But then, surprisingly, from 726 CE to 731 CE he wrote another book covering the history of the Anglo-Saxons from a Christian perspective. In this book, without any explanation, he suddenly dated all events in the Anno Domini counting – the years since Jesus' birth. It became the first book that ever dated events in our modern CE reference.

Why did he change his mind? He left a lot of texts, but in none of them does he explain his switch from being a fanatical defender of an anno mundi system – spending decades of work on it – to using Anno Domini counting from one book to the next. This remains an unsolved mystery in the history of the calendar.

Side note: At this point, the Islamic expansion had reached its largest geographical and political extent, from India to France. That empire had its own calendar. It started with year one in modern 622 CE, and was already a common standard for a vast part of the world.

Charlemagne, Alcuin, and standardization

Bede's book became niche-popular, but again nothing much happened to popularize the new counting of years going back to Jesus. In the same century, an ambitious king of the Franks met a later monk from Northumbria – a student of Bede's works. This monk, named Alcuin, became the king's advisor and helped him become Charlemagne, the Holy Roman Emperor. Eventually, they established a new Roman Empire in the West, with a pope in Rome, a political center in Aachen, a mixture of people with all sorts of inheritances in a new European order, and a church centered in Rome as the border‑crossing glue of a very fluid European political scene.

Alcuin helped Charlemagne set standards – a common currency, for example – and also the Anno Domini date. But again the effect and spread were rather slow. The first inscription on a building in the AD format can be found on the facade of the Cathedral of Pisa. It reads: Anno Dominicae Incarnationis M. LXIII – "In the year of the Lord's Incarnation 1063".

From Roman numerals to global standard

Around 1190 CE, a man from Pisa became consul in Algeria, and his son learned the Arabic numerals there. He became famous as Fibonacci, and with him Western culture changed to expressing years as "1063" instead of "MLXIII". (It remains a peculiar tradition that at the end of movie credits, the year is often still expressed in Latin form, so 2026 is shown as MMXXVI.)

Around 1450 CE, the printing press began to have its effect. Anno Domini dates in texts were copied and distributed en masse. Our modern understanding of the year became the norm in the West. The copyright notice – literally present in every single publication – uses this year.

From Anno Domini to Common Era: the Kepler transition

But there were two more twists to come. Yes, the Gregorian calendar reform happened in 1582 CE – that changed the structure of the calendar but did not change the year itself.

The first new twist came in 1615 CE, when an astronomer named Johannes Kepler published a book in which he called the year annus aerae nostrae vulgaris – "year of our common era". This phrase was later translated into English as "Common Era", now abbreviated as CE. Kepler demonstrated that the reference of counting years from Jesus' birth had evolved from something uniquely Christian into something completely common – a civil standard now used by believers and non‑believers alike.

This was the intellectual bridge: AD remained the ecclesiastical term, while CE became the inclusive, global, secular equivalent. Today, databases, logs, and international standards use CE precisely because of this shift that Kepler articulated.

As we know, the system could have ended in 2000 CE when the so‑called "Y2K bug" caused a lot of headaches. And – if computers and databases stay in their core the same until the year 9999 – the whole CE system will face another crisis on the 1st of January, 10000.

The chain of decisions

Overall, the chain of decisions is traceable: from Jesus of Nazareth in the villages of Caesarea Philippi, to Dionysius extending an Easter table in Rome, to Bede writing a history book in Northumbria, to Alcuin helping Charlemagne in Aachen to standardize a new Holy Roman Empire in Europe. The individual decisions on that journey are more or less mysterious, but the result is global and now deeply baked into humanity.

The year on your screen right now – 2026 – exists because a healer made an inexplicable decision to walk south, a monk disliked a dead emperor, a scholar switched calendars without explanation, and a king needed a standard.

And that is what year it is.

Timeline of key decisions

~30 CE – Jesus walks to Jerusalem
~200 CE – Africanus calculates birth in Roman terms
284 CE – Diocletian's reign starts
525 CE – Dionysius writes first Anno Domini letter
731 CE – Bede switches to AD without explanation
~800 CE – Charlemagne and Alcuin standardize AD
1190 CE – Fibonacci introduces Arabic numerals
1450 CE – Printing press mass‑distributes AD dates
1615 CE – Kepler coins "Common Era"

First published on dev.to. This article is CC BY‑NC 4.0.

AI in Software Engineering: Why Understanding Still Matters

Dietmar Schoder — Tue, 09 Jun 2026 17:53:25 +0000

The Limits of AI in Software Engineering: Why Understanding Still Matters

A software company delivers a product to its customers. Then a new government regulation appears, mandating changes to that software by a specific deadline. The company creates backlog items, breaks them into tasks, and developers start coding. The question is: how much AI should they use to write that code?

Three Levels of AI Adoption

Level one. The developer gives the AI access to the task description, the existing codebase, the version control system, and the development database. The developer then asks the AI to create a new branch, write the code change, and open a pull request.

Level two. The company also uses AI to read the new regulation, compare it with the existing system, write the backlog item, and break it down into smaller coding tasks. After the AI has handled the coding work from level one, the company asks it to perform the pull request review, merge to master, write test scenarios, run all tests, and deploy the change.

Level three. The company's own customers use AI to perform everything the software previously did for them.

How far is this a good idea? How realistic is this vision?

Comparing Human and Artificial Intelligence

To answer these questions, we must compare what human engineers do with what AI does. If we understand both, we can decide how much human work AI can realistically replace. At minimum, we can see what improves and what worsens as we replace more human work with AI.

So how do humans work? This question has generated fierce debate and endless scientific investigation. But for our purpose, a rough sketch of human cognition is enough. That sketch differs so fundamentally from how AI works that it alone reveals the answer.

The Human Mind: Building Models of Reality

A human being experiences the world. Whenever they want to understand something, they build a model of that domain. They construct this model as a system, meaning it has both structure and behaviour.

Take tennis. Anyone who knows what tennis means can easily create a mental model: two players with rackets, a ball, a net, a court with borders, and a barrier in the middle. These components represent the structure of the real-world system. Then the human adds serve, return, volley, fault, game, set, and match as behaviours or processes.

Every human has a model like this when they think of tennis, though not exactly the same model. And humans can do remarkably useful things with such models. They can observe a real tennis match and predict what happens next. They can refine the model when they learn new facts, such as that balls are yellow now instead of white. They can run mental simulations. What if there are four players? What if matches are too long and we introduce a tiebreak? They can even experiment with absurd hypotheticals. What if both players serve at the same time and must return the serves in parallel, meaning two balls are in play at once?

Some humans are better at this modelling than others. We usually call this skill "imagination," because we long believed that we hold images in our minds. But in fact, this skill is what human brains excel at, and it is what enables intelligence. Humans can invent mental models, explore them, use them for testing, adapt them, and constantly check how closely they match reality. Humans can write these models down, draw them, communicate them, and exchange enormously valuable knowledge through them.

The power of this ability cannot be overstated. Humans can listen to a fairy tale and, from listening alone with no further explanation, immediately establish what is valid in the fairy tale versus what is valid in real life. They can instantly recognize a fairy tale as such. They watch a film like Avatar, immerse themselves in a new world, and derive a new model of that world's rules. Humans are so good at this that they can detect flaws that break the rules in any fictional world, even when that world is complete fantasy.

Here is a certainty from daily life. Whenever someone has a highly realistic model of a particular domain in their head, we recognize that this person understands what they are talking about. Understanding is having a mental model of a real system, its structure and behaviour, that is close to reality. And it includes the ability to operate with this model in many useful ways, including learning.

How Large Language Models Work

Now, how does AI work? The AI the world is currently discussing consists of large language models. How do they work? During training, they read everything ever published on a given topic, such as tennis, on the internet. Then, when a user asks a question, the model searches through the vast mixture of words from its training and returns the most statistically likely words as an answer.

In other words, the AI selects human symbols based on probability, rearranges them, and returns them as an answer. Consider a simple example. Suppose we gain access to all written texts of an alien civilization, but we cannot translate them. The aliens send us questions. We do not understand anything they say. Nevertheless, we send back their own symbols, selected and arranged according to probabilities derived from their texts using a particular algorithm.

In this scenario, it is clear that we do not understand the aliens a little bit and gradually improve. No. We do not understand them at all, and we never will. And because we mirror their own texts back to them, the aliens will believe these are intelligent answers, as long as their texts are generally intelligent. This becomes even more impressive to them when a single user receives back the most matching answers from the entire alien knowledge base.

Even if this simple model of human intelligence and this simple model of LLMs are not remotely close to reality, one thing becomes crystal clear. Humans can understand portions of reality. LLMs cannot and never will.

The key difference is always what we call understanding. Understanding is the playful, joyful, realistic modelling of the world in the human mind. LLMs have none of this.

Implications for Software Engineering

Now we return to software engineering and AI. The full picture becomes visible immediately. If a software developer uses AI to gain a better understanding of their work, AI is certainly useful. But the more a developer or any user lets AI "do the work" as a black box, the more that user skips the understanding part. And AI will never replace that understanding.

Consequences follow directly. When the new code works in production, no one understands why it works. When it fails in production, no one understands why it fails either.

AI can then be used to fix the bug. But because the AI does not understand the software user's world, the new regulation, the existing code, or anything at all, it will likely fix some bugs while introducing new ones. AI will eventually encounter the same experience every software developer has when they do not truly understand what they are doing. The total number of bugs, vulnerabilities, and flaws in the software system will increase with every new attempt to fix it.

Conclusion

As long as humans use LLMs to gain more understanding, LLMs are helpful tools that make life easier for intelligent people. But the moment they are used to complete tasks faster by excluding humans, those tasks must not require the tiniest form of understanding. If they do, the work will be done with significantly lower quality. That will inevitably become costly and dangerous in the long run.

Leave a comment below. Let's talk.