One of the major objectives of using a Collaboratory is information gathering, whereby a user can access relevant information dispersed throughout the Internet, without having to travel extensively in Gopherspace or Webspace. Note that this does not imply that there should be only one such central point. That would go against the very spirit of the Internet, and pose an overwhelming administrative load on the site maintaining the Collaboratory. Later, in section 7, we envision a set of such central points distributed throughout the globe, serving as gateways to the field of interest and related information through appropriate organization and context specific search engines. What are the determinants of the level of information access? We focus on three factors, information organization, search engines and dynamic linking capability.
Information organization
As discussed earlier, there is already an information overload on the Internet, as exemplified by the results of queries with popular search engines. While this is definitely a positive indicator of Internet usage and its popularity, it poses the problem of ``Infoglut''. It is therefore important for the information to be organized in some structured form, so as to be easily accessible to the user. While the physical accessibility of data has been well defined in terms of addresses and protocols for various files and formats, there is still an identification problem of how one can obtain access to this information. The traditional way of identification has been keyword based searches (e.g., Wide Area Information Servers (WAIS)), which index words in documents based on efficient algorithms. This still poses the problem of numbers, where large volumes of documents (with or without any relation amongst themselves) have to be searched. To narrow this search further, subject or author based indexing is adopted. While we will shortly discuss some desirable characteristics of search engines, we can make the existing keyword based search work much more efficiently by reorganizing information according to a shared paradigm for the discipline of interest. Therefore, we suggest an intellectual indexing of documents and other information resources on a Collaboratory, whereby a set of related documents will be linked to each other. For example, in the world of MIS, an issue such as end-user computing will have many related facets involving technological, social and economic factors. A Collaboratory which organizes information on end-user computing without grouping relevant documents under the above categories with appropriate links between them is likely to be less useful to a user who would like to obtain information on, say, social issues in end-user support. Such a categorization and linking scheme will help a user efficiently locate a set of related documents on a given topic.In the design section, we explain our approach to this intellectual indexing of information for the field of MIS.
Better search engines
Search is central to information access and retrieval. Since we envision a distributed architecture for a set of core MIS Collaboratories across the globe, a document search would involve both location and content. Typical Internet search engines at specific sites maintains WAIS-based indexing schemes. With the proliferation of Web based documents, location now plays a central role in search. Web crawlers or WWW Worms address this problem by traveling in Webspace and searching for documents which meet certain criteria specified by users. However, the efficiency of search can be further enhanced by making them context specific.
Linking documents on Collaboratories
Apart from better information organization and search engines, information on different Collaboratories must be linked together, to enable search engines (like Web crawlers) to travel through the virtual network of Collaboratories. While the Web allows the creation of hypertext links, Collaboratories will have to support the capability to create such links on a dynamic basis, whereby a user who finds a relevant document can immediately link it to a Collaboratory, thereby making it immediately available to other users. In the absence of such a capability, there is the potential danger of creating pockets of hard-to-access information on different Collaboratories.