Digital Museum
Distributed Museum Concept for the 21st Century

− Ken Sakamura −

Digital Archive

One of the functions expected of a museum is the preservation of materials, another being to display materials so that they can be used. These two requirements have remained unchanged since ancient times. But, at the same time, they have been contradictory. The objective of preservation is without a doubt future use, but if those materials are used - whether for exhibition or research purposes - a deterioration in the state of preservation is more or less unavoidable.

Therefore, with the background of recent developments in digital technology, the concept of a digital archive was born. This entails digitalizing the information of materials as much as possible, then saving it and using it electronically.

Records have been taken to preserve the information of materials from the past. Typical examples are photographs and movies. The creation of miniature paintings of plants and animals may be described as another method of analog recording. Analog records of information can be made in this manner.

But it is unavoidable that the recording media deteriorate with time. Regardless how optimal the storage environment may be, all tangible items are unable to avoid destruction ultimately, no matter how gradual that may be. In addition, it is impossible for copies of analog records to be made at the same fidelity level as the original, and in this case, deterioration to a certain degree is unavoidable.

In contrast, perfect copies of digital data can be made without any loss of information. Whether the media on which information is recorded be an optical disk or a magnetic disk, the information will be lost or mis-recorded over time. However, by using the digital coding theory and using redundant information, it is also possible to discover such errors and then correct them. This draws the total error rate close to zero. Then, by regularly renewing data by copying that information itself onto a new medium, permanent preservation of data becomes theoretically possible.

Digital archives are clearly superior in terms of preserving data, but digital technology brings into full play many advantages when using the data. By using computers, the fruit of digital technology, it becomes possible to search information quickly, transmit such information, and if necessary, process it. As can also be seen from the Internet, once information has been digitized it can be transmitted around the world at will, and if permission is obtained by the user, it is possible to use the original data in its original, undeteriorated state. Naturally this process is not restricted to digitization of photographs. It is also possible to handle text, three-dimensional information, chemical formulas, a variety of data, sounds and moving pictures. It also becomes possible to organize this information, interlink it and retrieve it from a database.

Although not an example of a museum, in the ruins of "Chauvet-Pont -d'Arc" France, where a wall surface made by prehistoric man was discovered, the cavern was resealed in its original state after a detailed recording inside the cavern was made. People who want to see the wall surface are now expected to view the image displayed on the Internet. Higher resolution data is provided to researchers upon request. Measures such as these are based on unfortunate experiences such in the ruins of Lascaux discovered earlier, which deteriorated badly over a short time by being left open.

In this way, in cases where it is adequate enough to obtain information, not having to take out actual materials is important in terms of preventing deterioration of materials, and digitization makes a contribution to this process. The digitization of materials promoted the effective use of actual materials, while at the same time reducing the necessity of using actual items, thereby protecting the actual items.

From this perception, in order to utilize digital technology more effectively in museums and utilize digitization technology more proficiently, it becomes important to review the way in which digitization will be carried out, and the way information will be classified and organized. For this purpose, the conceptual model with which we view the museum is very important.


"White Lily"
(From Art Magazine Database)


"Laugh"
(From Art Magazine Database)


"Modern Western Pictures"
(From Art Magazine Database)


"Chuuou Bijutsu"
(From Art Magazine Database)

Museums as information devices

The author's specialty is information science. So, I modeled the function of a museum is as an information "device." Using this model, the existing function of museums were organized, and the mutual relationships of functions as modules were reviewed, and the introduction of new functions was considered.

Said, with a little fear of readers misunderstanding, from this information science viewpoint, what is of essential importance of the functions of the museums is the information possessed by the materials that the museums own. The actual materials can be regarded as a primary medium which stores that information.

Antique value and value as a rare item exist in the item itself. Therefore, the ownership of the item is therefore in this regard. In contrast, the academic value of the collection items owned by a museum lies in the ``use'' of information, whether it be displayed and viewed by many visitors, or whether researchers analyze it to obtain new knowledge. Actual ownership therefore is not of much concern.

This conceptual difference between ownership and use was not something people cared much before the advent of the Internet. Before, if only the use of information was desired, for example, the ownership of the media on which that information is recorded was the requirement for accessing that information fully. So it was not necessary to think of the two as distinct entities.

Let's consider the case of a train timetable as an example. In the past, even if the information desired is only for a particular train that one wants to ride, it was necessary to own the large timetable publication. Timetable in order to search for that information freely and at any time. With the appearance of the Internet, however, it has become possible to access the desired information without ownership of the information. If one accesses a train timetable website, it is possible to access only that information which one requires, and by inputting the time that one wants to arrive at a particular station, the website will even tell you what time you should depart and which route to take. Such information can be accessed using a mobile telephone. The place where the timetables exists is a large server computer connected on the other end of the telecommunications line.

No matter how large the hard disk of PCs becomes, it is impossible to store all of the information from all of the Internet servers from around the world. In contrast, if the goal is only to use that information, then it is possible to do so using light hardware such as mobile telephones.

The use of the information is most important. Materials are the primary media that carries the information. Museums are the devices that store these primary media, read off information from them and enable many people to use that information. Users use that information either by visiting the museum directly, or by accessing it via a network. The digital archive is a concept extended along the view outlined above.

The digital archive is a database which receives the input via a digitization function from the primary media (actual item), and accumulates and stores that information and enables users to use it. The digitization function are offered by a variety of different input devices that read information from actual items, such as image scanners and three-dimensional digitizers.

If this approach is extended to the museum function as a whole, it is important to consider the inputs and outputs for all modules. What is carried out between the inputs and outputs is the processing of each module. This approach is fundamental to information science, All the modules build up a system by mutual links of the input and output relationships. Determining the following information for each of the individual modules, it is possible to completely specify the system as a whole. This is the concept of an information systems model in recent years.

Let's apply this modeling concept to a digital archive, for example. The digital archive and actual item are regarded as individual modules, and information is sent from the actual materials to the digital archive via a digitization device. Does information flow in the opposite direction? This would mean that by carrying out so-called "reverse digitization," output that returns to the actual materials is conceivable. Naturally it is unlikely that the modification of actual items to conform to databases should be impossible. Changing one's perspective, however, the production and subsequent displaying of a replica based on data archived in the database may be considered to be illustrative of this reverse flow. In this case, the "reverse digitization" process is made possible by output devices such as printers and light modeling devices (devices which harden light-hardened resin by the flying spot of a laser, automatically creating a three-dimensional model in accordance with the input data).

In practice, there are limitations to reproducibility of output of light modeling devices such as the mouth of a vase, and manual correction is required. It is also possible to obtain image information from the surface of three-dimensional objects using a mapping camera, but with regard to the automatic coloring of replica surfaces in this reverse direction, there is still little hope for the automation of this process, and this process also needs to be manually handled. When such cases occur in information systems, humans are also considered as part of the overall system. In the system, however, humans are referred to as an entity to differentiate them from general modules, because they are regarded not as something that can be completely controlled, but something whose operation can change according to reasons not detected by the system. In fact, actual materials are also entities from a system viewpoint. Furthering this approach, in the following sections museum managers and further, visitors, are incorporated into the system as entities.

If people are introduced into a system as an entity, considering once more the flow from actual materials to the digital archive, for example in the case of historical literature where the recognition of handwritten characters cannot be computerized, then this part is definitely done by human input. In addition, for earthenware, the writing of the explanation, and the addition of that text data to information about earthenware stored in the digital archive, work which is better handled by people, is processing that is carried out by the entity.

Following on this approach, a configuration diagram showing a model of the next-generation museum function in line with this approach is shown in Fig. 1. The following section describes the Digital Museum - A Decentralized Museum Concept for the 21st Century - based on this figure.

Real & Virtual - Then a Digital Museum
Before dealing with the Decentralized Museum Concept for the 21st Century, what must be confirmed here is that after the information is read off, then is there any value in the original materials? With regard to this point, it is necessary to recognize that there is no media that has more complete information than the materials themselves, and that reading off all of the information from the materials (with science as it stands today) is impossible.
No matter how perfect a replica is made, it can never have more information than the original, and in most cases replicas only imitate one aspect of the original. With earthenware, for example, plastic replicas, no matter how perfect their appearance, are useless in analyzing the composition of the clay that is used, or trying to specify the place where it was made. Similarly, no matter how many aspects materials information is digitized using many different types of sensing technology, it is impossible for it to completely cover all information.

*Looking into the future, one cannot completely rule out the possibility to recording in an instant almost all information including the fine structure of the inside of materials using a type of hologram technology that uses strong X-rays. In this case, however, the object being copied is likely to be completely destroyed.


There are still some types of information that cannot be communicated as digitized information. If one were to visit all of the museums throughout Japan and hold and measure each actual item, it would be possible to obtain new information which could be obtained through the sensory pattern recognition unique to humans, something that could not be obtained through normal database searching.

Generally, information is always lost during digitization. That being the case, when asking the question whether a virtual museum made up of digitized information is a cheap cousin of a real museum where actual items are displayed, the answer is no.

For example, there are some cases where digitally archived materials have more information than the actual items. Adding descriptions and related information to exhibits using hyperlinks are such examples.

If one wanted to compare the weights of a large number of earthenware objects, then it would be sufficient for weight measurement data of a realistic accuracy to be available. If a database is searched via a network, it becomes possible for researchers to carry out surveys which would otherwise take many months' work of measuring actual materials in various museums throughout the country, to be performed easily over a short time on a computer. Similarly, if the aim is to investigate graphics patterns, it is sufficient if complete mapping data of surfaces is available. In this way, if the target information is clearly defined, then naturally it is easier to use information that has already been computerized.

Furthermore, reproductions of the past, such as coating with lacquer, reversing the wear and tear and restoring the original state cannot be carried out on valuable actual items. However, as long as the real items that actually exist have aged and are consequently not the same as they were when they were first made, then a difficult question is what to regard as "real." For example, thinking that Jomon earthenware was all a drab earthen color, or knowing that there were many pieces that were coated in lacquer and colored gaudily in red, makes great difference in one's perception of the culture of the time. Showing how the exterior of the real materials change from the "real" of today into the "real" of the past, going back as if on a time-machine, is well-suited to a virtual museum.

If the place where a particular piece of earthenware was excavated is linked to the materials, then it becomes possible to call up information about other materials dug up at the same time: fruit of trees, for example. Furthermore, a virtual museum makes it possible to access climate information for that location at that time, which fosters an understanding of the eating habits of the time, fauna of the location, etc.

This accumulation and expression of information about relationships through a temporal and spatial axis is where a virtual museum has advantages over a real museum.

Digitization is definitely not the technology that renders exhibits of actual items unnecessary, and is not confronting exhibits of actual items as an alternative.

As the configuration diagram indicates, virtual museums which are museums in a pure information space, and real museums which are museums in a physical space where actual materials exist both contrast one another and mutually complement one another. We define the ideal museum for the 21st century as a museum that extends across both spaces and organically integrates these two types of museums using digital technology.

Real Museums

In order to make virtual museums and real museums complement one another, it is necessary to ensure that computers are able to recognize entities such as actual materials and visitors, etc. Electronic tags and sensors acts as the intermediaries of information between real and virtual in this way. On the other hand, the intermediaries of information from virtual to real are displays and speakers in each exhibition venue, and the special small terminals - PDMA (Personalized Digital Museum Assistant) designed for museums and lent to visitors.

By attaching small electronic tags using contactless IC card technology, it is possible to read the tags with a sensor attached to the computer, thereby specifying the materials. Using such electronic tags it also becomes possible to adopt systems which automatically recognize where particular items are in storage. If electronic tags are checked each time exhibits are moved and displayed, then it becomes possible to thoroughly record and track the past of that material, such as when it was last exhibited, as well as the material's current location. If such information is available it also becomes possible to automate scheduling for airing and other maintenance.

Electronic tags can also be applied in exhibits. If a visitor has a PDMA with a built-in sensor, it becomes simple for the PDMA to provide an explanation about the exhibits which the visitor approaches. For large exhibits and exhibits comprising a large number of components, it becomes possible to provide detailed explanations about the areas approached by the PDMA.

The WWW (World Wide Web) was an important idea which resulted in an explosive take-up of Internet use. In addition to being able to view information on computers throughout the world that are connected to a network, the ease with which people can access related information and more detailed information with a click of their mouse is a major factor behind its increasing popularity. This is a structure referred to as hypertext by experts, and incidentally BTRON, of which specifications were developed at the Sakamura Laboratory, the University of Tokyo, incorporates this as an information management structure within the operating system. Hypertext originally appeared as a hierarchical structure for documents, but it is also possible for the things referred to be images or sounds. The term hypermedia is used when we refer to documents that contain such multimedia data.
In the Digital Museum, a structure is adopted which should be referred to as an advanced hyper-environment in which actual materials are added to these hypermedia, and information is linked from actual materials, and actual materials are linked from information. In exhibitions that have existed up until now, explanations have been attached to physical panels located nearby physical materials, and users were only able to read such explanations. Unless an expert fortunately happened to be next to them, it was impossible for them to obtain more detailed information or to explain points about which they were unsure. In the Digital Museum where a hyper-environment has been created, it is possible for users to obtain as much information as they want from the actual materials, and conversely it is also possible to lead users to other actual materials related to that information. Without a doubt such a function is useful not only to general users but also to researchers.

It has become clear, however, that with this system there are some situations where users cannot bring their PDMAs close to an exhibit, perhaps in cases where a large number of earthenware pieces are on display and the visitor is interested in a particular piece at the rear of a shelf, or the visitor is interested in the upper part of fossil remains that is as large as the wall. In order to accommodate such situations, we have developed a system by which visitors use a laser-pointer to indicate the actual materials about which they would like an explanation. This is then detected by a sensor near that material, and an explanation is sent to the terminal. This system is used in this exhibition.


Gate, Nimrud, Iraq (From Photographic database of archaeological sites in Western Asia)


Personal Museum

Recognition of visitors is carried out by human sensors, but it is necessary to specify the individuals in order to personalize the information to meet the characteristics of the visitor. The Personal Museum is another example of supplementing the real with the virtual. In the real world, as illustrated by terms such as mass productions and economies of scale, in many cases it is inefficient to respond separately to individuals.
For example, if efforts are made to match the expressions on museum panels to the visitors as much as possible, we need to pay attention to the following at the least.

It is clear that a wide range of combinations have to be accommodated. The number of combinations is further expanded if foreign languages other than English are required. Considering the combinations of these conditions, it would necessary to attach tens of panels to each material. Museums would become places where real materials could be found within a forest of panels.

In contrast, in the world of computers and information, responding separately to individuals is not a big overhead. If visitors input information about their own characteristics into their PDMAs when they first arrive, then it is possible to display text in accordance with those attributes. The information can be displayed in larger fonts if necessary, and it is possible to have the information read out for visually impaired visitors.

We are currently studying personalization in which exhibits react individually by actively recognizing. For example, if a distant exhibition were linked by a network, it is possible to identify the visitor, and change the commentary according the order in which they are viewing exhibits. Using this function it is possible to adopt organic links between exhibits, such as using information input at exhibit A at exhibit B. It should be possible for the museum exhibits to recognize the visitors, and in principle provide a different response for each individual.

Instead of exhibits of the past where visitors mainly looked into glass boxes, unique exhibits become possible in which exhibits are arranged freely in a panorama (it may be necessary to use replicas, however), and visitors are able to walk around them freely, looking at the commentary on their PDMAs as it suits them.

I led research project in this field and refereed to this type of technology as Dokodemo Computer (lit. Go-Anywhere Computer) and HFDS (Highly Functional Distributed System). Recently it has become established as a leading-edge field and is referred to as Ubiquitous Computing and Cyber Augmented Environment.

The personalization database in the configuration diagram keeps the records visitors even after they have left the museum, so if the same visitor visits the museum again, it is possible to have the PDMA characters set automatically using this personalization database, and suggest a better way for the visitor to view exhibits based on their behavior during the previous visit.

Interaction with visitors is stored as question and response patterns on a FAQ database, separate from the personalization database. By doing so, it is possible to take in the questions harbored by visitors, and reflect these in the exhibits. In addition, by regularly maintaining the FAQ database, a visitor support system can be created in which appropriate questions can be posed by an expert system for those doubts often harbored by visitors in an effort to identify those doubts, and having the FAQ database automatically indicate the appropriate response.

In the future, it is likely to become possible to use an artificial intelligence recognition technology called image recognition to act as the intermediary between virtual and real information. The exterior appearance of materials is digitally archived, i.e. it is in the records of the museum, so it is quite conceivable that it may become possible to specify materials from the exterior appearance by referring to this archive. Similarly, it may also become possible to use face recognition to check for and identify visitors who have already visited the museum before. Furthermore, the commentary texts are generally carefully written, and edited logical commentaries, so as it may be possible to use automatic language translation, the use of artificial intelligence technology will likely become the research theme in museums of the future.

Virtual Museums - Then Distributed Museums


Monolith-Type Kiosk Termin


Automatic Changer of CDROMs

With regard to virtual museums accessed over the Internet, as demonstrated in the configuration diagram, personalization is carried out by the personalization database and the FAQ database. Using the personal ID assigned to visitors, it is possible for them to access the museum over the Internet upon their return home and obtain a list of the exhibits they saw that day, access data they were interested in which they specified as take-home on their PDMAs, and to obtain responses via e-mail for those questions about topics which they were unable to grasp from the commentary alone.

Apart from the provision of web page-based information by general WWW technology, in a MUD three-dimensional environment, it is possible to have visitors visit the museum in virtual space, and have them walk around viewing different exhibits placed within the space.

As demonstrated on the configuration diagram, it is also possible to incorporate a human response by having cooperative members of museums who are experts provide response to questions which cannot be handled by the FAQ database.

In addition, if the personalization database can be shared among museums connected to the network, it is possible to change the commentary on the assumption that visitors have already viewed materials from another museum. If the information on the FAQ database can be shared, it is possible to provide a more stylized commentary in response to more diverse questions.

By virtually connecting MUD virtual spaces, and letting visitors move into different museums by going under a door in the virtual room, the ultimate museum comprising the digital archives of all museums throughout the world can be created on a network. This is the concept of a distributed museum.
The POS (Point of Sale) approach of the commercial distribution network states it is better to input at the place where information is generated, and a similar concept is likely to be introduced into the field of academic materials in the future. For example, currently each time a piece of earthenware is found at an excavation site it is recorded in a ledger. If instead, an input device such as a three-dimensional digitizer were brought in to digitize each piece, the creation of a database would be completed there an then, and reconstruction work could be carried out in a virtual space with computer support. Both of these represent considerable merits. Reconstructed earthenware which has been assembled in a virtual space can then be stored via a network into the museum's digital archive together with excavation information for each piece.

If other museums, different academic databases, and excavation sites where actual materials were discovered are linked together, a distributed museum would come to store an extremely large volume of information. The pressing problem will be how to search for the necessary information with the absolute minimum of work. In searching, the item being not in the database is certainly a problem, but there is an extremely large volume of information, the inability to narrow down information becomes a more pressing problem, as is the case with the problem faced by general search engines currently on the Internet. If the necessary information is buried in several tens of thousands of search hits, then it cannot be found. Even if there is a large volume of information, if its location cannot be found when required, and one does not know how to view it, then the information may as well not exist at all.

The digital archiving of materials entails formulating the information of those materials and storing it on a computer. We must be able to search the wealth of information based on not only simple names and material numbers but on other characteristics as well. When the databases of multiple museums are integrated and connected via a network, and searched as if they were one database, the data structure must be clearly specified, the meanings of individual items defined, so that a computer can understand and process them.

This is generally a problem of distributed databases, but in terms of a decentralized museum this encompasses more difficult problems. When a distributed database is constructed by a commercial business organizations, the main problem is that of synchronization during use. When a particular item is updated, on some databases it appears as it was before the update, while on others it appears as it is after the update, and while the data is being updated, some on another database attempt to use it. In contrast, the problem of distributed databases in museums is that, these are distributed individual entities to begin with, and there are diverse genres, and many classifiers, so it is difficult to decide on how to standardize the attributes of multiple and diverse actual materials. In other words it is a problem relating to database construction and the root of the problem runs deep.

What is important here is to clearly lay down the format that will be used when materials are digitized. For example, on the Internet as it stands today there are no problems as long as English text is used, but there are obstructions to the interchange of other languages, and the only way of sending historical characters is sending them as graphics images. With regard to this point, I have used every opportunity to make recommendations to rectify the character code problem, and have led research in order to bring about actual resolution to this problem. But the information handled by museums is diverse, including three-dimensional data, materials, classification names, chemical structures and map information, apart from simple character data, and stipulations must be clearly laid down to permit these different types of data to be distributed efficiently. This really should be called Museum Information Infrastructure (MII), and forms the basis of digital museum development in the future.

In order to solve the digital format problem, we have developed Museum TAD, which is an attribute description data format for museums. This is a description system that has a flexible attribute description system, and is a framework which ultimately enables the integration of knowledge through the efforts of many organizations and many individuals digitalizing diverse actual materials in an distributed manner.

If this technology results in an organic linking not only of TOKYO UNIVERSITY DIGITAL MUSEUM but also museums and research institutes around Japan and around the world via high-speed networks, it will create a huge network of knowledge in which valuable genuine pieces from various collections and information from around the world are interlinked to one another.