Introduction
Even though the Nordic countries are small, it would be preposterous to pretend to map all the activities in the Nordic countries in this field. Instead, this paper will try to outline some lines of development and present some examples of implementations. In the last five years, a number of projects combining materials from archives, libraries and museums have been initiated in Denmark and other parts of Scandinavia; all involving different new technologies in order to present the material in innovative and immediately accessible ways. It is interesting to examine the approaches underlying the different projects and the resulting differences in user experience.One group includes mind-mapping, topic maps and other similar technologies in order to present large bodies of material in an ordered manner independent of the classification of the individual institutions. Other projects have relied on more traditional approaches using a limited set of common data to enable aggregated searches. A different group of sites have a more limited scope. Some of these try to tell coherent stories based on limited but thoroughly researched material; others depend on presenting large bodies of material on a single subject, or on a smaller area of subjects.
Challenges
One of the principal obstacles for all attempts at presenting cultural heritage information from the ALM-area is the different outlook of the different kinds of institutions. These differences have conditioned development of individual world-views and accordingly affect decisions on what types of information are necessary and relevant. Furthermore, information that is common to all institutions, such as names, places and dates, is seldom registered following the same standards; even more important, classification systems are radically different from one institution to another.
One of the principal challenges for all types of cooperative projects in this area is therefore the merging of the different ontologies of the institutions. Based on a preliminary survey, the projects seem to fall into a few rather distinct categories. One is projects aiming at presenting the material in “pure” form, thus leaving interpretation to the user. Another tries to present the material together with context, in order to make it more intelligible to users not inherently familiar with the subject matter. Yet another tries to develop the individual information snippets into coherent narratives, thus making it interesting to a wider audience. This paper analyses advantages and drawbacks of the specific solutions.
Possible Solutions
It is evident that the digital presentation and utilization of the immense number of materials is to some extent at a parting of the ways in a Scandinavian context. On the one hand, there is a policy of letting “a thousand flowers bloom”; on the other hand, a strongly perceived need to make all these diverse sources of information more easily available, to allow aggregated searches in some form.
This last approach comes in two distinctively different flavors. One attempts to make everybody use the same system, or at least report a set of core data to a centralized database. The other tries to build a transparent interpretation layer between the individual databases and the users.
The schools of thought lying behind these two different approaches are clearly linked to different notions of how easy it is to extract meaning from natural language, and to different ideas of the semantic Web in general. There are those who believe that this is not only theoretically possible but also practicable, while others think that this possibility is at best only attainable in a not-too-certain future. The former tends to advocate ontology-based solutions, while the latter favors centralized repositories using controlled vocabularies.
In order to understand the policies adopted in the cultural heritage area, a little background on the history of Nordic museums in necessary.
Background
In Denmark and the other Scandinavian countries, the overwhelming majority of the local museums are sprung from popular initiatives in the decades around 1900. Partly because of their origins, there has been a marked propensity for independence. In this context it has taken the form of numerous individual databases and presentations of different bodies of material. In all fairness, it must be noted that many of these were done by first-movers, long before any centrally organized system could have been operational. The drawback is that they magnify the task of aggregating searches and effectively limit them to a very basic set of data.
Fig 1: Opening page of the historical atlas of Funen
Even today, many well-functioning systems are based on the work of a small body of individuals. A case in point is “the historical atlas of Funen” (Figure 1). In the opening page of the museum of Odense, the third largest city in Denmark (185.000 inhabitants), presents geographically organized historical material on the island of Funen (http://www.historiskatlas.dk). As can be seen from the screenshot, this is done using a map as background, grouping the information physically around this. It is clearly a well-functioning site, but it is not immediately usable in other connections.
A number of initiatives try to induce the individual institutions to report to central repositories, often using controlled vocabulary and a limited amount of information (parameters). An example of this is the Danish museernes samlinger (collections of the museums) which makes it possible to search in the aggregated database of the ministry of culture (see figures 2 and 3).
Fig 2: The initial page of http://www.kulturarv.dk/mussam/index.jsp
Fig 3: A search result from “mussam”
This database is populated partly through museums using a common interface "Regin", and partly through XML import from individual databases. All Museums are required by law to deliver data to this database, using either of the two methods. Not all museums are at present contributing, but this will shortly be the case.
Fig 4: Data entry page of “regin”
As is seen, the interface is very similar to the external interface of mussam, which of course indicates the temporary nature of the current GUI.
As is apparent, the user friendliness of the current interface leaves a lot to be desired. At present a large project to create common tools for the presentation of these data is underway. The museum of Copenhagen is involved in this development which includes, among other tools, a timeline. The idea behind these tools is primarily to enhance the usability of the common database, but also to make it possible to present data from the database together with information gathered for specific purposes. An example of the use of this could be to integrate information on Danish artifacts together with important dates in the history of other cultures. In this way, the immigrants to Denmark would get easier access to Danish cultural heritage to see its relationship to other cultures. This project is planned to be operational by the end of this year.
In Norway, the ALM area has been integrated at the state level in a government body called ABM-utvikling, which among many other initiatives has initiated a process of consolidation of the museums, trying to merge smaller units into larger, more professionally run units. This process has of course met with some opposition, as it necessitates unification of a lot of routines formerly done individually.
This leads to another line of separation in the area of cultural heritage presentation.
What is Presentation?
In Denmark at least there have been many attempts to qualify the term presentation, in order to distinguish between just making material available, and disseminating knowledge with the material as the vehicle.
The former is, of course, much easier, and, it can be argued, is a necessary condition for the latter. This is to a large extent correct, but in actual practice many projects remain at the first level. While they may be very useful for “the trade”, they are practically of no use to the general public. Therefore, they are not a good argument in favor of increased grants, and ultimately not a good argument for Web-based presentation/learning.
A New Approach
The site http://www.Absalon.nu which I demonstrated last year is an attempt at another approach; a third way, so to speak Instead of waiting for the perfect foundation for the presentation, we have initiated the aggregated search, hoping that enhancement of the material will go hand in hand with enhancement of the GUI. This approach is at the site complemented by the inclusion of an encyclopedia, in which the information is controlled by the institutions. In this way we hope that the two sides will complement each other, and create a both reliable and innovative experience.
User Reactions
User reactions to this have been twofold. On the one hand, reactions have been very positive, apart from users directing our attention to specific errors. On the other hand, the use of the site has been much less than anticipated. It must be noted, however, that we have deliberately not advertised the site, for two reasons. First of all, we wanted to have a period without too much traffic, in order to be able to solve performance issues. Secondly, it must be said, parties in the institutions have been hesitant towards directly advertising a site with content that was less than 100% controlled by the institutions. The compromise was that we started off gathering experiences and user reactions.
Drawbacks
One of the drawbacks to this solution has turned out to be that in the period of collecting experiences, the iconography of the Web changed so much that the design and look-and-feel of the site have become an obstacle to its acceptance. We have as a consequence been forced to revamp the site, a modernization we hope to introduce in April of this year.
Another is the disadvantage of living in a non-Anglophone country: we cannot easily take advantage of large international thesauri, ontologies, etc. Before we can utilize ULAN, TGN, etc., we have to translate labels. This involves much labor, and invariably gives rise to discussions about whether it would be better to create our own, and this is in turn even more time-consuming …. It quickly turns into the old discussion of the best being the enemy of the good.
Another Approach
Another approach is followed by sites where the aim is to present resources in a more portal-like way. The Norwegian site http://www.kulturnett.no presents events, temporary exhibitions and the like. More similar to the Danish mussam is the Norwegian http://www.primusweb.no .
Fig 5: Primusweb: Opening page
Fig 6: Primusweb, result, showing a traditional Norwegian costume
The Swedish, site run by the office of the “state antiquarian,” is also similar.
Fig 7: The Swedish http://www.kms.raa.se/cocoon/kulturmiljosok/enkel_sok.html
Fig 8: Result from a search on pictures
In the same vein but with an even more austere interface is www.arbimus.dk (“arbimus” is an acronym based on the initial letters in arkiv, bibliotek, museum (archive, library, museum).
This last system does indeed make the material available, but the search tools are limited to free text, offering a choice from 45 categories in 5 groups and/or from a list of 5 different types of material.
Fig.10: The arbimus list of topics
An approach somewhat similar to absalon is the Swedish "samsök" (aggregated search) which is based on creating a transparent layer between the individual databases and a common search interface.
Fig. 11: http://samsok.museum24.se Opening page
Behind the scenes it is planned to collect the content of these mappings and in time use them to turn the site into an ontology-based search-engine. This has the advantage of being useful from day one, and evolving into a more and more advanced system. Current plans are underway to test the system in a Scandinavian setting, to try to internationalize the search.
Fig 12: Part of result, showing cupboard
One of the principal decisions required before building a site of this character is whether to include all material or limit content to materials that are deemed “fit to print”. The case in Scandinavian museums, as probably in most museums, is that many of the registrations are made for internal purposes. That means that there has traditionally been neither the need to proofread the registrations nor the need to express oneself in a manner immediately understandable to everyone.
The Institutional Culture
It seems that the approaches to solving these types of problems are very much linked to the general outlook of the person(s) in charge, and to the “ethos” of the institution in question. I can’t help recalling the differences explained at last years MW between the Smithsonian and Brooklyn museum, regarding how much control was felt necessary in the area of blogging.
In Denmark, for example, some museums have felt that making the material known at all has priority, and have accordingly made a subset of their recorded information available, even if it at times it is very scant indeed. Examples of this are the already-mentioned NOKS and Arbimus. Other museums have made material available only after proof-reading and enhancing the data.
Absalon has tried to devise automatic and semiautomatic ways of rectifying the more basic omissions and obvious errors. To illustrate the problems facing public use of the material, the question of dating is a case in point.
The Problem of Dating
One of the problems with poor quality data, at least in many Scandinavian museums, is the lack of dating, especially of archaeological artifacts. As everyone working with the artifacts knows the rough dating of the object in question, there is no need to include this in the registration. Dating is included only if it is a closer dating based on context or the result of scientific dating like radiocarbon dating or the like. The consequence of this is, however, that these registrations are useless if the museum wants to present them on a timeline. This is an area where semi-automatic, algorithmical methods might be employed to enhance the data. In Danish museums, the concept of cultural-historic periods is used to give this dating; e.g. Viking age, middle age, etc.
In the case of datings, it would probably be possible to add outline dating of the individual types based on cultural relationships to allow presention of the artifacts with a reasonable chronological precision.
The Problem of Personal Data
Another problem which is not so easily solved concerns personal data. In many cases, persons are named in connection with the acquisition of the artifacts. Mention can be in the form of donors or comments concerning former use. At times the staff have made comments, including personal remarks. These types of data are naturally not fit to publish, if for no other than legal reasons. The snag is that these remarks are not always confined to the field where one would expect them, and it is therefore difficult to be absolutely sure that they do not slip out unless every record is thoroughly checked before publication.
The Problem of Proofing
A perhaps less serious debate, but one that nevertheless causes profound discussion, arises over the amount of proofreading that should go into digitized texts. This is an area where there have been huge differences between institutions, and even greater between public institutions and private enterprises; that is, between museums, archives and libraries on one hand and user-initiated projects on the other. The latter have to a much larger extent made the “raw” OCR’ed texts available and left it to the users to proofread, while the majority of institutions have felt that the impossibility to make exhaustive searches on texts with many errors made them unusable for their purposes, and have felt that the importance of the reputation of the institution prevented it from presenting such faulty material.
There have, of course, simultaneously been those advocating the other approach, again arguing that “the best is the enemy of the good”. In other words: if we have to wait for the texts to be proofread, we’ll all be dead before we see them on the Web. More to the point, even the limited usability of the uncorrected texts is an improvement over no digitized text at all, until proof-reading can be completed. As with most questions of this type, there are political aspects as well as professional. There is widespread fear that if just the faulty text is on the Web, funding for the much less visible task of correcting will be harder to get.
There may be some truth in this last argument, but last fall I heard the Royal Library in Amsterdam hinting that they may be lowering their level of proofreading on digitized tests in order to boost volume. This seems to be a powerful voice in favor of the immediate diffusion of less-than-perfect texts.
An example of a totally different approach is www.historiefyn.dk. This site has taken the radical approach to the question of control, and presents only material that is arranged specifically for this purpose. The result is a number of very coherent stories and a very organized way of navigating. The downside is that you very soon exhaust the material on the site, and economic constraints have prevented participating institutions from increasing the amount of material on the site.
It is very obvious that presentation of cultural heritage material on the Internet is still very much in embryo in Scandanavia.
New Steps
A number of avenues are being pursued, and with profound difference from institution to institution. The attitude towards presentation on the Internet as opposed to, or as a supplement to, other forms of presentation is radically different in different places.
This is illustrated by the current debate at the main library in Copenhagen, where the newly appointed leader, who wants to boost presence on the Internet and allocate resources to this end, has met with fierce resistance from a number of sides. Not all of this opposition is directed the right way. There is little doubt, however, that if cultural heritage institutions designed for the good of the common population, as most of them historically were, were to be reinvented in the 21st Century, they would have a very large focus on Internet-based presentation.
Some of the so-called Web2.0 functionalities are being taken up by museums. The museum in the town of Truer has built a site where users can send in their stories and upload pictures. The site is, however, designed so that each “story” becomes its own static page in a kind of CMS-system.
A similar approach has been adopted by The ALM-institutions in the Danish town of Viborg who are cooperating on www.viborghistorie.dk. This is a larger undertaking, as it involves much more diverse types of material, but the concept is basically the same: each piece of information is allocated its place, and internal linking is “hard coded”.
Fig. 13: Viborghistorie, page shoving details of a street
It is obvious that this keeps very close control over the user experience, and in an “encyclopedic” setting this is obviously an important point. It is just as obvious that this approach is not scaleable. The economic resources involved soon become prohibitive, as there are almost no large-scale production advantages, apart from work-flow etc.
It seems that if we want to aggregate the material and reap the insights that can be gained from such aggregation, we have to accept loss of control to a certain extent. A number of projects primarily in the text genre are relying on users to proof-read, and this seems to be working. It is surprising to some, how many people want to use a bit of their time to share their expertise.
In the last few years, new theories of learning stressing the importance of student involvement have boosted interest in Web-based learning, including learning in the field of cultural heritage. A number of projects trying to utilize these insights are beginning to take form; they areas yet so new that there is no significant experience with the results.
References
Mäkelä, E., O. Suominen and E. Hyvönen: Automatic Exhibition Generation Based on Semantic Cultural Content. Proceedings of the Cultural Heritage on the Semantic Web Workshop at the 6th International Semantic Web Conference (ISWC 2007), Busan, Korea, November 12, 2007. Available on-line at http://www.seco.tkk.fi/publications/
http://www.kulturarv.dk/tjenester/publikationer/emneopdelt/boernekultur/digitalformidling/index.jsp
http://www.historiefyn.dk/test/projektet.htm
http://www.abm-utvikling.no/publisert/abm-skrift/abm-skrift-32
http://www.abm-utvikling.no/publisert/brosjyrer/veien-til-kulturarven.html
http://www.kulturarv.dk/tjenester/publikationer/emneopdelt/kulturarv-it/abm2006/index.jsp
