Imagine you could surf Facebook ... from the Middle Ages. It may not be as unlikely as it sounds. In a fun and interesting talk, researcher and engineer Frederic Kaplan shows us the Venetian Machine of Time, an 80 kilometer digitization project to create a historical and geographical simulation of Venice during 1000 years. (Video filmed at TEDxCaFoscariU.)
The translation was done by Dimitri Frangoyanni and the editing by Chryssa Rapessi for TED Talks.
This is an image of the planet Earth. It looks very much like Apollo's photographs that are well known. There is something different, you can click on it and if you click on it, you can zoom almost anywhere on Earth. For example, this is a panoramic view of the EPFL campus.
In many cases, you can also see how a building looks from a nearby street. This is awesome. But something is missing from this wonderful tour: Time. I'm not sure when this photo was taken. I'm not even sure if it was pulled at the same time as the panoramic image. In my workshop, we create tools to travel not only in space but also in time. The question we are asking is is it possible to make something like Google's past charts?
Can I add a slider on Google Maps and change only the year, seeing how it was 100 years ago, 1000 years ago? Is that possible? Can I rebuild the social networks of the past? Can I bark a Facebook of the Middle Ages? So can I build machines of the time? Perhaps we can just say, "No, it is not possible". Or to think about it from the point of view of information. That's what I call the mushroom of information.
Vertically, you have the time, and horizontally, you have the amount of digital information available. Obviously, in the last 10 years, we have a lot of information. And the farther we go in the past, the less information we have. If we want to build something like the Google Maps of the past, or Facebook of the past, we have to grow this space, we have to make it rectangular. How do we do this? One way is digitization. There is a lot of material available - newspapers, printed books, thousands of printed books. I can digitize all that. I can extract information from them. Of course, the less we go in the past, the less information we have. So it may not be enough. So I can do what historians do. I can extend. This is what we call computer science. If I get a diary, I can see, it's not just a diary of a Venetian captain making a specific trip. I can regard it as a calendar representative of many trips of that period.
Προεκτείνω. Αν έχω έναν πίνακα μιας πρόσοψης, μπορώ να θεωρήσω ότι δεν είναι μόνο το συγκεκριμένο κτίριο, αλλά πιθανώς να έχει την ίδια γραμματική με κτίρια για τα οποία έχουμε χάσει κάποιες πληροφορίες. Έτσι, αν θέλουμε να κατασκευάσουμε μια μηχανή του χρόνου, χρειαζόμαστε δύο πράγματα. Χρειαζόμαστε πολύ μεγάλα αρχεία, και χρειαζόμαστε εξαιρετικούς ειδικούς. Η Μηχανή του Χρόνου της Βενετίας, το program which I will talk to you about, is a joint program between EPFL and the University of Venice Ca'Foscari.
There is something very strange about Venice that its administration was too bureaucratic. They have recorded everything, almost like Google today. At Archivio di Stato, you have 80 mileage records that record every aspect of life in Venice for over 1.000 years. You have every ship that comes out, every ship that comes in. You have every change that has been made in town. It's all there. We are setting up a 10 year digitization program designed to transform this vast file into a giant information system. The kind of goal we want to achieve is digitizing 450 books a day. Of course, digitization is not enough, because these records are mainly written in Latin, in Tuscany in the Venetian dialect, so we have to copy them, translate them in some cases, register them and that is obviously not easy.
Συγκεκριμένα, η παραδοσιακή οπτική μέθοδος recognitionς χαρακτήρων που μπορεί να χρησιμοποιηθεί σε τυπωμένα κείμενα, δεν λειτουργεί καλά σε χειρόγραφα κείμενα. Οπότε η λύση είναι να εμπνευστούμε από έναν άλλο τομέα: την αναγνώριση φωνής. Ένας τομέας που μπορεί να φαίνεται αδύνατος, αλλά που μπορεί να γίνει, απλά βάζοντας επιπλέον περιορισμούς. Αν έχετε ένα πολύ καλό μοντέλο της languages used, if you have a very good model of a document, how well structured it is. And these are administrative documents. They are well structured in many cases. If you split this huge file into smaller subsets where a smaller subset shares similar elements, then there is a chance of success.
If we reach this stage, then there is something else: we can extract events from these documents. And in fact, about 10 billion events can be extracted from this file. and this giant information system can be searched in many ways. You can ask questions like, "Who lived in this palace in 1323?" "How much was a fagri on the Realto market in 1434?" "What was the salary of a glassmaker in Murano over the course of perhaps a decade?" You can ask even longer questions because they will be semantically encoded. And then you can put them in space, because a lot of this information is spatial. And from that, you can do things like reconstructing this surreal journey of that city that managed to have a sustainable development for over a thousand years, managing to have a form of balance with the environment all that time. You can reconstruct this journey, visualize it in many different ways. But of course, you can't understand Venice by just seeing the city. You have to put it in a larger European background. So the idea is also to record all these things that were working at the European level. We can also reconstruct the journey of the Venetian maritime empire, how it gradually controlled the Adriatic Sea, how it became the most powerful medieval empire of its time, controlling most of the sea routes from east to south. But you can still do other things, because in these sea routes, there are repeating patterns. You can go a step further and create a simulation system, create a Mediterranean simulator that is able to reconstruct even the missing information, which would allow us to ask questions like if you were using a route planner. "If I am in Corfu in June 1323 and want to go to Constantinople, where can I get a boat?" We can probably answer that question to within two or three days. "How much will it cost;" "What's the chance we'll encounter pirates?"
Of course, you understand, the central scientific challenge of such a project is the characterization, quantification and presentation of uncertainty and inconsistency at every step of the process. There are everywhere bugs, bugs in the document, the name of the captain is wrong, some ships never went out into the sea. There are translation errors, interpretive biases, and moreover, if you add algorithmic processes, you will have errors in recognition, export errors, so you have too much uncertainty. So how can we discover and correct these inconsistencies? How can we present this kind of uncertainty? It's hard.
One thing you can do is to record every step of the process, not only by codifying historical information but what we call the post-historical information, how to construct historical knowledge, documenting each step. This does not guarantee that we are converging on a unique history of Venice, but we may be able to reconstruct a fully documented history of Venice. There may not be a single map. There may be several maps.
The system must allow this, because we are dealing with a new form of uncertainty, which is really new for this kind of giant database. And how should we communicate this new research to a large audience? Again, Venice is notable for this. With the millions of visitors that come every year, it really is one of the best places to try to invent the museum of the future.
Imagine, you see horizontally the reconstructed map of a given year. and vertically you see the document that served the restructuring, tables for example. Imagine an immersion system that allows us to dive and rebuild Venice of a given year, an experience you can share within a group. Conversely, imagine starting from a document, a Venetian manuscript, and showing what you can make of it, how it is decoded, how the document box can be recreated. This is a picture of a report currently being held in Geneva with this kind of system.
In conclusion, we can say that research in the humanities is about to undergo a process probably similar to what happened in the life sciences before 30 years. It's really a matter of scale. We see work that is far beyond what a research team can do, and this is really new to the humanities, who are very often used to working in small groups or only with one or two researchers. When you visit the Archivio di Stato, you feel that it is beyond what a single team can do, and that it should be a joint effort. So what we need to do about this change of perception is to grow up a new generation of "digital humanists" who will be ready for this change. Thank you very much.
