Comparison with Other Systems and Methods

The previous section described how the MRA-HE implementation provides functionality to support the scenario CSFs. In order to assess the success of MRA-HE in terms of providing an improved document management environment over currently available methods, this section describes how existing methods may, or may not, achieve similar functionality and assist in the CSFs previously identified. Although MRA-HE was designed specifically to benefit the scenarios in chapter 4, there is value in assessing how other mechanisms may perform against the critical success factors identified through those scenarios.

Conventional Desktop Computing

Traditional office software products focus on support for the individual in accomplishing work-based tasks.

They provide advanced functionality for document production, manipulation and storage on a per-document basis. General-purpose operating system buffers permit ’copying’ and ’pasting’ of existing materials into a new context. Mechanisms such as Microsoft’s Object Linking and Embedding (OLE) protocol assist in the inclusion of one document inside another while retaining a live link to the original document. Software products such as recent versions of Microsoft PowerPoint allow the application of a ’master slide’ that defines an overall presentation to a set of created slides, providing a limited separation between documents and their presentation. The resulting document is stored in a folder structure dictated by the owner of the file space; some office applications permit the addition of document meta-data and of annotations to the main document. Current file systems permit the definition of shortcuts or symbolic links that permit the definition of cross-folder relationships, allowing for multiple classification of documents and for the reuse of folders within a number of other folders. Users with access to the same local area network may access certain parts of a shared file space, or search across them, allowing them to access documents of relevance, and access to different drives provides a low level of support for multiple classification schemes. Across a local area network, permissions can be set for read and write access to individual files.

Conventional desktop mechanisms provide a means to implement a number of the features and qualities required by the use scenarios. There are, however, a number of fundamental drawbacks in using a traditional desktop solution

• Documents and ’resource rooms’, which may be implemented as sets of work-specific folders, are accessible only over proprietary networks and proprietary operating systems, precluding the possibility for widespread access and discovery.

• Although documents and folders may be used elsewhere in the file structure, there are no rules by which to check the integrity of the file store, nor to ask the question ’where else is this document stored?’

• The provision of document meta-data is ad-hoc, and there is no explicit association between a document or folder and the person or organisational element to which it belongs

• There are no mechanisms for the separate application of document-related entities such as document annotations or hyperlinks.

• There is no notion of an event system to support the discovery of user history trails or to register interests and receive notification updates

• Although resource-level reuse is possible through ’copy and paste’ and OLE methods, there is no structure or control over the widespread and distributed use of such mechanisms.

• There is no explicit support for different user types, providing the same interface to all users

Virtual Science Park Resource Rooms

The application of the current VSP Resource Room architecture to the identified use scenarios was discussed in chapter 4. The strength of current resource rooms in assisting the identified scenarios is the ability to define purposeful resource rooms for each VSP tenant that store documents in groupings that represent a valuable resource classification for a work purpose. The VSP short falls in pursuit of the use scenarios described in chapter 4 are reiterated below.

• Within the VSP resource rooms, each resource is connected tightly with the folder it was originally stored in, tying the document to its original purpose. The nature of the resource room information model hinders large-scale repurposing of documents because of the tight binding of a document to a specific (primary) resource room folder.

• The VSP currently supports the addition of lightweight hyperlinks between resources and the registration of interests in a particular resource or folder. However, the VSP provides no architectural support for hypertext approaches such as document history trails, typed hyperlinks and annotations. These are fundamental omissions from the current architecture.

• The VSP resource room architecture does not provide a document-model based approach.

Beyond the meta-data that the document owner must supply at document check-in, reading room resources are considered as Binary Large Objects (BLOBs). There is no distinction between a resource’s content, structure and presentation. It is therefore difficult to promote reuse of the document resources for more than their original purposes and to perform full- text searches across the documents.

• There is no direct link between the resources held within the VSP and the VSP information directory. It is impossible, therefore, to locate information about a document’s author except through a disconnected text search across the information space. This connection is essential to facilitate collaboration once a document of interest is found.

Comparison using Critical Success Factors

Table 6-4 lists the critical success factors identified at the start of this chapter and shows those which are satisfied by the approaches discussed in this section – the MRA-HE implementation, conventional desktop computing and VSP resource rooms. In the following table, each CSF is given a three-stage rating according to how it is satisfied by the relevant approach.

the approach is unable to satisfy the identified CSF

the approach is able to partly satisfy the identified CSF

the approach is able to more fully satisfy the identified CSF_

Critical Success Factor (abbreviated) MRA-HE Desktop

Computing

VSP Resource

Rooms 1A) manage more than one classification scheme

1B) associate keywords and descriptions with resources 1C) multiply classify and manage the same document 1D) copy resources without duplicating documents 1E) access from a web browser

1F) automatically re-link resources

1G) view who has seen a particular document 1H) view who has expressed an interest in a document 1I) find out where a document is classified

1J) control access permissions for reading and editing 1K) find out what documents other people have seen 2A) search the meta-data for all documents in the store 2B) view information about the owner of a document 2C) register an interest in a folder or document 2D) search an organisation for document owners 2E) perform full-text search through documents 2F) create pointers between classification folders 2G) reuse document fragments

2H) retain a dynamic link to document fragments 2I) specify a new document presentation style 2J) move resources between classification folders 2K) include one folder structure within another 2L) view information about a folder owner 3A) create and follow links between documents 3B) create and view document annotations 3C) restrict the application of links and annotations 3D) view creator of links and annotations

3E) adapt the interface for different user types

Table 6-4: Satisfaction of CSFs by Three Approaches

Microcosm - A Modern Hypertext System

The Microcosm project and range of commercial products from the University of Southampton (UK) represent many years of research in the area of large-scale hypertext management. As such, it is able to offer robust solutions to a number of CSFs identified in this chapter. Microcosm started by focussing on the provision of a hypertext system that was able to automate linking between electronic documents within large heterogeneous document collections. In recent years, during the development of this thesis, the scope of Microcosm has extended to supporting more social aspects of a hypertext system - the mining of document trails (MEMOIR [DeRoure98], [Pikrakis97]) and the grouping of documents in application-specific collections (Microcosm TNG [Goose97]). Microcosm is able to provide functionality not available within MRA-HE, notably the construction of automated hyperlinks and other hypertext mechanism such as the provision of document ’tours’ that end users are able to follow. It is clear that a system that incorporated all the ideas from Microcosm and its extended mechanisms would be able to serve many of the critical success factors from the scenarios identified in this thesis, but a few notable omissions follow:

• Microcosm provides an ’umbrella environment’ for documents within its store, which may be of any type. As such, it provides only a limited document model-based approach, which does not emphasise opportunities for resource-level reuse.

• Microcosm provides no explicit association between a document and the organisational unit to which it belongs. Its emphasis is on making all information available, not on maintaining an ownership of that information or restricting access to unhelpful information such as inappropriate document links and annotations.

• Microcosm does not offer a working environment to support day-to-day working, although its work with Pirelli on specifically assisting maintenance tasks [Hall96] points to a single application for a work purpose. Microcosm’s primary aim is on the development of large-scale hypermedia stores, not on supporting the day-to-day work of the individual system users.

• Microcosm is not a native Internet product, although a read-only option is available using a conventional web browser.

Due to the inability to fully assess a Microcosm product, and because it is not targeted as a document management system per se, it was neither possible nor advisable to rate its success against the identified critical success factors. Microcosm would rate highly on the hypertext CSFs, but the absence of support for focussed work and lack of emphasis on document classification and reuse renders a direct comparison meaningless. It is included here, however, as a well-developed hypertext system to highlight the way in which traditional document management and hypertext appear to be converging upon the same problem

domain. Document management systems are developing understanding of document content, while hypertext systems are beginning to address the social aspects of document classification and use.