Museums and the Web 1999

Best of the Web

Archives & Museum Informatics
2008 Murray Ave.,
Suite D
Pittsburgh, PA
15217 USA

Join our Mailing List.

Published: March 1999.


Designing Web User Interfaces Adaptable to Different Types of Use

Fabio Paterno , CNUCE - CNR, Italy
Cristiano Mancini , CNUCE-CNR, Italy


The advent of the web has made easy to develop hypermedia with museum information that allow a wide variety of users from all parts of the word to access it. However little attention has been paid to the problem that usually different type of users access such a information with different purposes. Thus it is important that web interfaces are able to adapt themselves in such a way to better support the achievement of different goals from different types of users. More generally, adaptation of hypermedia systems to each individual user is increasingly needed. Adaptation can solve the problem of hypermedia systems which are used by different classes of users. Users can seriously differ in their goals, background and knowledge covered by the hypermedia system. Besides, adaptation can prevent the user from being lost in hyperspace; with the growing size, complexity and heterogeneity of current hypermedia systems it becomes harder and harder for the existing navigational tools to provide orientation on where to search for the needed information.

Adaptable systems are systems that allow one to modify some parameters of the system and then adapt their behaviour accordingly. If the system adapts to the user automatically it is called adaptive. Adaptive systems tailor information to the user and may guide the user in the information space to present the most relevant material, taking into account a model of the user's goals, interests and preferences. Though most applications that make use of adaptive hypermedia are currently in the area of educational hypermedia, where it might be easier to acquire a detailed user-model, these techniques might prove very suitable for the museum domain too. A museum web site can be structured to allow free roaming but it should also facilitate learning, and, especially for this purpose, adaptive hypermedia can be helpful. To this end it is important to take into account the many studies that have been conducted to understand the way in which people visit museums and the way it is best to build an exhibit.

The main problem in the development of adaptive hypermedia systems is to structure the information in such a way as to allow adaptations. The structure should include a characterisation of users and the interface should be organised to allow the system to infer from user actions performed the characteristics request. In the paper we want to discuss different techniques to design adaptable and adaptive interfaces for museum web sites and provide examples of applications of some of them by considering the type of impact that they can have in a web interface for the Marble Museum located in Carrara (Italy). Currently in the application for the Marble Museum, at the beginning the user can select a visitor profile (e.g. expert, art student or tourist), and during the virtual visit the user has the opportunity to change this profile. Thus the system is adaptable because the user can modify the profile parameter and the system changes its behaviour accordingly.

More precisely, in this paper we first briefly discuss the various possible general approaches that can be adopted to obtain user interfaces able to adapt to the different types of users. Next, we describe the web support for the Marble Museum, located in Carrara, Italy, that we have designed and developed. In the current version this museum site can be classified as a adaptable hypermedia and we show the specific solutions that have been adopted for this case. Then further possible improvements, such as including adaptive features, are more specifically discussed taken this specific web museum as reference point. The possibilities and the conditions under which this can be done will be carefully considered.

Adaptability versus Adaptivity

When a hypermedia is used by users with different goals and levels of knowledge an important aspect is to support adaptation: different users may be interested in different parts of the information contained and they may want to use different links for navigation. Most of current hypermedia systems, on the other hand, are independent from the type of user: they provide the same hypermedia pages and the same set of links to all users. Adaptive and adaptable hypertext and hypermedia systems attempt to bridge this gap trying to use knowledge about a particular user, represented in the user model, to adapt the information and links being presented to that user.

Generally speaking, user models can be static or dynamic. In static models, user knowledge is represented as 'topic-value' pairs but the values are not completely independent. The user can be assigned to one or more stereotypes (for example novice - intermediate - expert). Each stereotype is characterised by a fixed set of pairs 'topic-value' and the user assigned to a stereotype inherits all these properties. This modelling is reliable enough and works well for a system that needs to adapt to different classes of users. In the dynamic model for each domain model topic there is some estimation of how well the user is familiar with this topic and this estimation takes into account the history of previous accesses to the application. In dynamic modelling it is possible to measure user knowledge with more flexibility. The static models are simpler than dynamic modelling, but less powerful. Good results can be achieved by combining stereotype and overlay modelling. One possible way to combine these two techniques: static modelling is used to determine the class of the user and to assign initial values for the dynamic model. Then dynamic modelling is used to keep the model updated.

Adaptive systems tailor information to the user and may guide the user in the information space to present the most relevant material, taking into account a model of the user's goals, interests and preferences. One solution to this problem is to use the data collected on a user which are contained in the user model to adapt the information and the links which have to be presented. By knowing the goals and the knowledge of the user an adaptable hypermedia can support users during navigation by focusing on the information provided, providing comments on the visible links, or suggesting what links are the most relevant thus obtaining an adaptable navigation (Brusilovsky, 1996). In (Brusilovsky, Eklund, Schwarz, 1997) techniques to support adaptable navigation are highlighted and their application to the WWW context is discussed. For example, the technique of adaptable navigation proposed is interesting since it associates links with a comment informing the user about the current state of the nodes which can be reached. The annotation can be provided in textual form or in the form of visible elements, for example using different icons or colours or different types of fonts.

Adaptation can occur at three levels:

  • Presentation, where it is possible to differentiate the type of media, the layout, the attributes of the perceivable elements (such as font type and size) depending on the type of access;

  • Information, where the information content can be changed, sometime drastically, even if related to the same topic, depending on the type of user and the use foreseen;

  • Navigation, where different links are provided, in some cases in different locations and with different appearance.

Adaptive presentation techniques

The idea of various adaptive presentation techniques is to adapt the content of a page accessed by a particular user to current knowledge, goals and other characteristics of the user. Existing adaptive presentation techniques deal with text adaptation. Text adaptation implies that different users at different points in time may get different texts as a context of the same page. We group all these techniques into one technology that we call adaptive text presentation technology. Another possibility for adaptive presentation techniques is to change the layout of the page instead of the text. One technique which is concerned with changing the text as well as the layout is stretchtext. It is one example of an adaptive presentation technique which is a special kind of hypertext. In a regular hypertext, a result of clicking on a hot word is moving to another page with related text. In stretchtext this related text can simply replace the activated hot word (or a phrase including the hot word), thereby extending the text of the current page. If required, this extended or 'uncollapsed' text may be collapsed back to a hot word. The user can override the adaptation by opening and closing any desired piece of information. A higher level technique which is based on stretchtext is suggested in MetaDoc (Boyle and Encarnacion, 1994). Each node in MetaDoc is a stretchtext page which may contain many uncollapsable hot words. The idea of adaptive stretchtext presentation in MetaDoc is to present a requested page with all stretchtext extensions that are non-relevant to the user being collapsed, and all extensions relevant to the user being uncollapsed.

Adaptive navigation support

The idea of adaptive navigation support is to help users to find their paths in hypermedia by adapting link presentations to the goals, knowledge and other characteristics of an individual user. The most popular technologies for adapting link presentation are direct guidance, adaptive ordering, hiding and annotation (Brusilovsky, Eklund, Schwarz, 1997).

Direct guidance is the most simple technology of adaptive navigation support and can be applied in any system which can decide what is the next 'best' node for the user to visit according user's goal and other parameters represented in the user model. A problem of direct guidance is that it provides no support for the users who would not like to follow the system's suggestions. Direct guidance is useful but it should be used together with a 'more supportive' technology.

The idea of adaptive ordering is to sort all the links of a particular page according to the user model and some user-valuable criteria: the closer to the top, the more relevant the link is. Adaptive ordering has a limited applicability: it can be used with non-contextual links but it can hardly be used for indexes and content pages and can never be used with contextual links and maps. Another problem with adaptive ordering is that this technology makes the order of the links non-stable: it may change each time the user enters the page. Recent research shows that the stable order of options in menus is important for novices.

The idea of navigation support by hiding is to restrict the navigation space by hiding links to irrelevant pages. A page can be considered irrelevant for several reasons: for example if it is not related to the users current goal or if it contains material that the user is not yet prepared to understand. Hiding protects users from the complexity of the unrestricted hyperspace and reduces their cognitive overload. Hiding is also more transparent to the user and provides different effects from adaptive ordering where links are usually added incrementally, but not removed or reordered. However seeing a different selection of outgoing links every time a node is visited can also be confusing, Therefore, the pages with a varying link structure must be limited and well chosen. Hiding has however another problem; as noted by some psychologist hiding can provoke formation of incorrect mental models of the hyperspace.

The idea of adaptive annotation technology is to augment the links with some form of comments that can tell the user more about the current state of the nodes behind the annotated links. These annotations can be presented in textual form or in the form of visual cues using for example different icons colours, font sizes or font types. The typical kind of annotation considered in traditional hypermedia is static (user independent) annotation. Adaptive navigation support can be provided with dynamic user model-driven annotation. Adaptive annotation in its simplest history based form (outlining the links to previously visited nodes) has been applied in some hypermedia systems including several WWW browsers. Even this simplest form of adaptive annotation which can distinguish only two states of links (links to visited/not visited nodes) appears to be useful. Current adaptive hypermedia systems can distinguish and annotate differently a higher number of states on the basis of the user model.

For example the ISIS-Tutor system (Brusilovsky, 1997) uses adaptive navigation as a primary technique for adaptive navigation support. Isis-Tutor uses direct guidance, hiding and annotation. In this system the teacher can set for each student a sequence of learning goals. A goal is a set of concept nodes of the network that must be learned. Concepts that belong to the same goal are expected to be learned together and then mastered by solving a number of problems before a student moves to the next goal. ISIS-Tutor uses two methods to adapt to the learning goal: first, to attract the students attention, it can outline links to the concepts belonging to the current goal; second (as an option), to decrease the students cognitive load, it can hide concepts which belong to future learning goals.

Adaptive annotation supports stable order of links and avoids problems with incorrect mental maps. Annotation is generally a more powerful technology than hiding: hiding can distinguish only two states for the related nodes - relevant and non-relevant- while annotation as mentioned above, can distinguish a higher number of states. Hiding can be quite well simulated by the annotation technology using a kind of "dimming" instead of hiding for "not relevant" links. Dimming can decrease cognitive overload to some extent (the user can learn to ignore dimmed links). Dimmed links are still visible though (and selectable, if required) which prevents the user from forming wrong mental maps.

Most existing adaptation techniques use exactly one of these ways to provide adaptive navigation support. However, these technologies are not mutually exclusive and can be used in combinations. In particular the direct guidance technology can be naturally used in combination with any of the other three technologies.


The Adaptable Web Interface to the Marble Museum

The hypermedia for the access to the information of the Marble Museum has been designed following a model-based approach (Patern?, Bucca, 1997). This approach has raised the interest in various research sites interested in design of interactive software application (Puerta, 1997). The basic idea is to use some declarative description to drive the implementation of a software artefact. In this case we have used task models to support such a design. A task model is a logical description of the possible activities that a user wants to perform when interacting with the considered application. In order to obtain meaningful task model it is very important to involve in the discussion related to their development all the relevant people, in the case of a museum application: manager and employees of the museum, possible end users, artists, software developers, application designers, user interface experts. In our case the resulting task model has been specified using the ConcurTaskTrees notation, a notation that we have developed (Patern?, Mancini, Meniconi, 1997). It is a precise notation that provides a rich set of temporal operators thus allowing designers to describe interactive, concurrent dynamic activities. The purpose of such a notation is to support designers in understanding how activities should evolve and then to obtain a design of the user interface consistent with the indications given in the task model following some specific criteria (Patern?, Mancini, 1999).

The hypermedia contains information on the works of arts placed in the museum, on tools that have been used to quarry and process marble, and to pieces of works located in the historical centre of the town that can be considered as a natural extension of the museum as it is particularly rich of artistic works done by marble, such as marble icons, sculptures, monuments and so on.

When we designed such an application we soon recognised that different types of users have different related task models because when they approach a museum application they have different goals, different basic knowledge of the application domain, interest in different information and so on. We decided to group the possible users into three types: tourist, student (of art), and expert. Tourists are characterised by the need for basic general information, expressed and presented clearly. They like to access information by spatial representations (for example a museum or a city map) because this gives implicitly information to them concerning how to organise a physical visit. Students have a better knowledge of the application domain so they want to access a wider range of topics, receiving more detailed information. Finally experts want to have full access to all the information available. They need minimal support in formulating their request and should be allowed to formulate their requests in a flexible way.

When there is a classification of possible users it is always possible to find specific cases that do not fit in any of them but we have found this classification suitable for most users.

In the design of the hypermedia we have an initial part that provides general information on the museum. It supports a few tasks which are independent from the user profile: the choice of the language and some general information on the museum that do not involve access to the information related to the works of art (opening time, how to get to the museum, fares) and it is independent from the type of user. Then we give the possibility to select a user profile (Figure 1). Users can have access to the information that describes the main features characterising each user type. Then the user will be able to navigate in the hypermedia by the navigation styles associated with the user profile chosen.

the available profiles

Figure 1: the available profiles.

The resulting hypermedia was designed according to the indications contained in the users task model. We developed three task models, one for each main type of user identified (expert, tourist, student of art). More specifically they differ for three main aspects:

  • Initial access to the museum information, for example, the expert can specify directly very specific requests whereas the tourist access mainly by spatial representations of the museum and the town;

  • Presentation of the information related to the works of art, which takes into account the basic knowledge of the type of users and the different tasks that they are likely to wish to perform;

  • Navigation in the hypermedia, which is more structured and pre-ordered for tourists whereas more navigational freedom has been given to expert users.

Initial access to the information

As you can see in Figure 2 the initial access to the information in the web museum depends on the current user profile. The expert (right side in Figure 2) can specify directly very specific requests by typing in the fields available, selecting lists of values or combining both techniques, whereas the student can access only information by lists (Figure 2 - centre) indicating the information available and tourists access information mainly by spatial representations of the museum and the town that can be activated from their initial point of access (Figure 2 - left side);

The initial access for the different users

Figure 2: The initial access for the different users.

Presentation of information related to a work of art

As we mentioned above, different views of the same information can be possible depending on the type of users. For example in the left part of Figure 3, we can see the tourist view of information related to a sculpture. As you can see it is possible to be guided to go to the next work. In this case the next work means the next work in the Modern Sculpture Section of the museum. It is also possible to access the list of works made by this material, the works performed by using different materials (wood, bronze, and others) and the museum map which drives the visit of the tourist in the museum hypermedia.

In the case of an expert user (Figure 3 - right part) the information can be accessed more immediately (for example, by just giving the name of the author), it is more detailed (for example, precise dimensions and precise date of creation are given), and further information on the material, the author, the biography or other information can be accessed.

An example of different views of the same information

Figure 3: An example of different views of the same information

Navigation in the hypermedia

There are different styles of navigation depending on the type of user. In the case of experts or student of art they can formulate a query (experts have a wider set of possibilities) and then navigate among the information that satisfy their requests. In the case of a tourist the navigation is more structured and driven by spatial representations (the map of the town or the map of the museum). Thus they can navigate among the works that are in a section for the museum or that are located in a street of the town. Usually such works are grouped following some criteria (such material, type of work, and so on).

Users can change the current user profile (tourist, student, and expert) interactively during the session. Thus they can access the different views on the information available and the navigation styles without having to start a new application session. It is important to note that at any time the user can change the current profile by selecting an item in a pull-down menu and from then on the hypermedia will show the information differently providing new ways to navigate through it.

Possibilities of Using Adaptive Techniques in the Museum Hypermedia

We have found the level of support provided by the current hypermedia for the Marble Museum satisfying because on the one hand users have shown to be able to easily understand it and exploit it in the navigation and on the other hand they appreciate this possibility that allow them to have a more supported access to the information contained in the hypermedia.

We have considered the possibility provided by an adaptive solution. However, some research experiences have shown that a complete adaptive solution where the system, depending on the user selection, automatically change the type of presentation and navigation proposed often fail to provide a usable behaviour because users may not understand why the system changes its behaviour and the underlying rules driving such changes can be valid for some cases but not for others.

One more interesting solution is to include some adaptive features that while still leaving full control to the user give the possibility to provide more dynamic support, taking into account the interactions performed by users and their preferences.

When to provide adaptive support?

As we said before we prefer to leave to the user full control of the adaptive support. This can be performed in two ways:
  • the user explicitly indicates when to activate such a support, for example when users are not satisfied of the current information they may want to ask for further information and the system should understand what kind of further information may be of interest in that specific context, taking into account the previously performed interactions;

  • the system, taking into account the user interactions, indicates that it has some suggestions and additional information to provide and the user decides whether or not to allow it to show this information; in the latter case an agent, with some rules incorporated, is included in the hypermedia to allow the system to understand how to dynamically present information (some information can be dynamically removed because the user does not seem interested in it or other information can be added because the user seems desiring to receive more detailed information).

What information should adaptive support provide?

There are various type of information that can be dynamically provided. They can be classified in these types of information:
  • Summarising information, thus after presenting a series of piece of information it can be useful to summarise the main aspects that have been considered; the system can dynamically elaborate some information depending on the data available. For example, if the user examines multiple works of the same author it can be interpreted as a strong interest for such an author and this can trigger the presentation of summarising information on such an author, such as a discussion of the preferred techniques and materials used

  • Highlighting issues, in this case the purpose is to highlight elements that can raise curiosity, discussion or comments from the users;

  • Compare different items, giving some information that can allow the user to make the comparison or to provide directly a comparison; there are many types of information that can be analysed to provide comparisons or highlight differences, for example the dimensions or the date of creation.

  • Highlighting differences, to show how a piece of information differs from the previous information considered or from a specific piece previously considered. Another possibility is to provide information by differences with respect to what has been already considered. Thus, taking into account the history of the interaction and the related information presented the system can propose comparisons and contrasts to the previously viewed information while omitting any background information that the visitor has already been received. This means that the system has a logical description of the information contained and it is able to keep the history of what the user has already seen.
It is also possible to give some automatic support to suggest users to change among the profiles predefined or to allow them to configure dynamically one specific profile intermediate among those predefined. We still prefer that the system suggests a possible modification explaining the reasons for this suggestion rather than to change the current profile dynamically. For example, if the current profile is the expert that gives the possibility to access additional information whenever a work of art is presented but the user never selects the additional information then we can assume that probably there is no much interest in detailed information and another user profile can be more suitable for the current user. Another example is when the user starts with the tourist profile but after a certain number of accesses they may be interested to get more detailed information and the system should be able at least to propose this possibility.

There are many aspects in the presentations of hypermedia that can be adapted. We have given examples concerning text-level adaptation, another possibility, for example, is to change the layout of the page according to the user's needs. One example is to always have an overview of the current room present as to help the user navigate through the museum.

Adaptive navigation support in the museum application

It is important to be careful to cognitive overloading and to the possible problems for user navigation. An excessive use of adaptive techniques can generate confusions in end users and get them lost during the navigation.

A useful support is to give feedback of the links to nodes already visited. In the current state of the hypermedia application the works of art are viewed in circular form (once the user has seen all works, the next-work button brings her again to the first work), it is important to provide at least the simplest form of link annotation seen/unseen: any link to a page that has already been viewed should change colour. Another possibility for link-level adaptation already mentioned above is to give links to pages already visited a different colour. There are also some possibilities for adaptive ordering. For example in pages where a list of works is presented, works of art not visited before could be placed on top, while links to works of art already seen by the user could be placed below and have a different colour. In this way the layout of the page is changed to make navigation easier for the user without increasing the cognitive load. These link-level adaptations are all very simple but have the power to greatly increase the user-friendliness of the application.

Attention should be paid to how to support the access to the "next" page. In our case we give the possibility to sequentially scan the list of works grouped by author or by historical period or by material. The order of scansion should reflect a logical order. For example, if the user is examining works of a certain author these could be ordered by date of creation or by material. The criteria for defining an order of scansion can be established by the system or by the user, selecting one of those supported by the system. Another possible example is that the hypermedia order the works of an artist by the number of accesses thus the scansion can follow the level of interest raised in end users.

Another important aspect is to take into account the restrictions concerning the technical environment of the user accessing the web site. Information with high data volume (e.g., videos, high-resolution pictures) is replaced by less demanding equivalents with low-bandwidth network access (e.g. via a slow modem) in order to reduce download times if required. The response time of a hypermedia system is extremely critical from the point of usability.

Examples of the use of adaptive presentation techniques in the museum application

This section will provide some examples of the possibilities for adaptation described in section 4. Some examples will be given on the use of adaptive text presentation technology. The main goal of these extensions is to supply the user with more information and to make the overall presentation more attractive. By referencing to works earlier visited the presentation can be more coherent and less monotonous.

The goal is to give the user the feeling he is shown around the museum by a real human tour-guide. This feature should stimulate the user to spend more time with the application and consider the application as more attractive. Now some examples will be given on how exactly this earlier information can be taken into account. The first few examples are based on the works of Felice Vatteroni, using the "tourist" visitor profile. The Vatteroni donation consists of some 20 works of art. To the tourist all are described using the following format.

<description of the work>
<material used>

There are several dimensions on which the current work-of-art can be compared to the ones previously viewed, for example:

  • differences concerning the sculptor
  • differences concerning the material
  • differences concerning the historical period
We now show an example of each of them. We have identified a space in the presentation of information on a work of art of the current hypermedia that can be used to provide the additional information provided by the agent. It is surrounded by a rectangle and in bold in order to highlight it.

Differences concerning the sculptor, when the visitor has only seen work of other sculptors so far and is now entering the virtual Vatteroni room, some general information on the sculptor could be included in the first Vatteroni-work viewed. For example (bold text represents the text added in the adaptive version)

example of differences concerning the sculptor

Figure 4: example of differences concerning the sculptor.

Differences concerning the material, when the visitor encounters a work of a new material some remark could be made about that. For example

differences concerning the material

Figure 5: differences concerning the material.

Differences concerning the historical period. When a user moves from works from one historical period to works from another some remark could be made about that. For example when a user virtually enters the modern sculpture part, some remark could be made about the differences between modern and ancient sculpture. For example

Differences concerning the historical period

Figure 6: Differences concerning the historical period.


We have discussed how the design of web sites containing museum information can be improved to take into account the different needs of different types of users. We have indicated the different levels of automatic support that can be provided. We have also presented the solution that we have provided in web site for the Marble Museum ( and we have taken this system as reference point to discuss further solutions.


We wish to thank the C.N.R. Project on Safeguard of Cultural Heritage, the Carrara City Council and the staff of the Museum of Marble for their support.


Boyle, C. and Encarnacion, A.O. (1994) MetaDoc:an adaptive hypertext reading system. User Modelling and User-Adapted Interaction 6(2-3), 87-129.

Brusilovsky P. and Pesin L. (1994). ISIS Tutor: An adaptive hypertext learning environment In H. Ueomo & V. Stefanuk (eds.), Proceedings of JCKBSE?94, Japanese-CIS Symposium on knoeledge-based software engineering, Tokyo: EIC.

Brusilovsky P, Eklund J., Schwarz E., Adaptive Navigation Support in Educational Hypermedia on the World Wide Web, Proceedings INTERACT?97, Chapman & Hall, pp.278-285.

Paterno F., Mancini C., Meniconi S., ConcurTaskTrees: A diagrammatic Notation for Specifying Task Models, Proceedings INTERACT?97, Chapman & Hall, pp.362-369

Paterno F., Bucca M.F., Task-Oriented Design for Interactive User Interfaces of Museum Systems, selected papers from ICHIM?97, ed. Bearman & Trant, pp.23-31

Paterno F., Mancini C., Developing Adaptable Hypermedia, Proceedings Intelligent User Interfaces?99, ACM Press, 1999.

Puerta, A., "A Model-Based Interface Development Environment", IEEE Software, pp.40-47, July/August ?97.