Author: Tracy Gardner, UKOLN
Date: 1st July 1999
Public Version (Created after meeting on 29th June between Rosemary Russell, Paul Miller and Tracy Gardner)
In the UK a number of organisations and projects are currently in the process of making public information available via some form of 'hybrid information environment'. In some cases organisations may wish to make their own information available to the public (e.g. the Natural History Museum), in other cases the aim is to provide access to existing resources to a particular user community (e.g. the National Grid for Learning).
In all but the simplest cases the resources are distributed across a number of repositories using a variety of communication protocols and resource description formats. The goal is to allow the user to search or browse across such resources in a uniform manner and to have results presented consistently. Additional goals include tailoring the presentation of information to the user's profile and to the current context. Further services such as conversion between document formats and attaching comments or ratings to resources may also be required within the same environment.
Currently a number of organisations provide hybrid information environments for their user community, other are at the planning stage and yet others are at the stage where they realise that such a system would be valuable. While particular hybrid information environments will vary considerably in their detail they also have much in common. The immediate aim of this work is to capture the commonalities that exist at a high-level and to make that information available in the form of a generic requirements analysis that can be used as the basis for new systems. This should prevent similar work being carried out repeatedly by different organisations and should also lead to a degree of consistency across the systems thereby increasing the chances of interoperability in the future.
Current systems rely on proprietary protocols and it is not possible to simply plug in new data sources or new services. In the future it should be possible to buy a hybrid information system from one vendor and 'plug in' components from other vendors, or develop specialised components in-house. A forward-looking goal of this requirements analysis is to move a step closer to hybrid information systems with pluggable components based on open standards.
The starting point for the requirements analysis is the MODELS Information Architecture (MIA) which describes hybrid information environments at a high level. The requirements for a MIA-like hybrid information environment will be developed along with a set of interoperability requirements for such systems. The MIA logical architecture will be developed further to provide the high-level framework within which these requirements must be met. The MIA architecture has already proved to be a useful tool for understanding the behaviour of hybrid information systems.
An analysis of hybrid information environment concepts and their interrelationship will be included, this may include concepts such as digital objects, search result sets, user profiles, and so on.
A set of 'scenarios' will also be developed to capture the basic functionality of hybrid information systems. The scenarios will be at a fairly high level since over-specialisation will prevent the generic application of the scenarios. The scenarios will cover the areas of searching and browsing, result collection and combination, context sensitive adaptation, and also the introduction of new functionality such as adding a new data source.
Having developed a set of generic scenarios a number of providers, or potential providers, of hybrid information environments will be approached in order to elicit requirements for particular hybrid information environments. Note that we are concerned with providers that have a requirement to offer a hybrid information environment to a particular user community, not the providers of hybrid information systems or the providers of content (although in some cases these may be one and the same). The requirements analysis document will be used to guide the collection of requirements - specialised requirements and scenarios will be developed to produce a tailored requirements analysis. It is hoped that the tailored requirements analysis will be of direct use to some of the providers, particularly those at an early stage in the provision of a hybrid information environment.
The experience gained from gathering detailed requirements from providers will be used to improve the requirements analysis document and produce a revised version.
It will be possible to draw conclusions about the value of the approach described here and to make recommendations to the hybrid information environment community. This work will be presented at MODELS 10 in October 1999. The desirable interoperability points for such systems have been identified. In some cases appropriate standards for interoperability may already exist in which case they will be recommended, in other cases areas for future standardisation work will be recommended.
A further task that may be approached if time permits is the comparison of existing hybrid/digital information system architectures (AGORA, BUILDER, MALIBU, AFS, etc) against the generic architecture. It should be possible to identify the extent to which each supports the recommended interoperability points. It should also be possible to determine the level of support for each of the generic scenarios, providing a structured comparison method for digital information systems.
Responsible: Tracy Gardner (with input from Rosemary Russell, Paul Miller and Lorcan Dempsey)
Produce a document describing the MIA building on existing MODELS work including a logical architecture and domain model (task 2 will feed into this activity so tasks 1 and 2 need to be carried out it parallel). Document should also identify `interoperability points'.
Responsible: Tracy Gardner, Paul Miller, Rosemary Russell
Generic scenarios should cover basic search and retrieve functionality, authentication and authorisation, user profile management, service provider profile management, provision of/access to non-search related services (document conversion/translation; rating of resources, resource annotation) and interoperability. A limited set of functionality should be covered in the initial version of the document with aditional functionality being added later in the study.
In their simplest form, scenarios are enumerated lists describing the steps required to perform a particular task.
For example, the following is a generic scenario for a search-retrieve action:
User initiates search.
User is prompted for search terms.
System presents results to user.
User processes results:
a) User performs actions on results (e.g.
rank, refine, save), or
b) User selects resource to view.
Repeat step 3 until user exits search.
Responsible: Paul Miller and/or Rosemary Russell (with Tracy Gardner)
The generic scenarios developed in Task 2 and the architecture developed in Task 1 should be used as a basis for gathering high-level requirements from organisations with a requirement to offer a hybrid information environment. Depending on the organisation the requirements may describe a fully or partially existing system or a system that is currently being considered.
For each client organisation each of the generic scenarios should be discussed to determine whether the organisation has a current or future requirement to support that scenario and if so to derive a specialised scenario (or scenarios) describing the required behaviour.
Interoperability requirements for each of the interoperability point identified in Task 1 should also be considered. The focus should be on functionality rather than specific protocols.
Responsible: Tracy Gardner
The MIA and generic scenarios should be refined based on finding from Task 3 to ensure that the architecture is generally applicable and that the scenarios provide sufficient coverage of common requirements.
The resulting document will be of use to organisations considering building a hybrid information environment.
Responsible: Tracy Gardner, Paul Miller, Rosemary Russell
Final results in the form one or more presentations at the final MODELS workshop.
The exact nature of results will depend on the outcome of earlier tasks but is expected to recommend areas in which interoperability standards need to be identified or developed and to identify scenarios that can be used in requirements gathering for hybrid information systems.
Responsible: Tracy Gardner, Paul Miller, Rosemary Russell
The architecture and scenarios developed in earlier tasks provides a structured basis for the comparison of existing systems and their architectures.
Systems can be compared both for their functionality (based on scenario support) and their potential for interoperability (based on existence of interoperability points in the architecture).