Creating a Collaboratory in Cyberspace: Theoretical
Foundation and an Implementation
Anitesh Barua
Ramnath Chellappa
Andrew B. Whinston
{barua,ram,abw@cism.bus.utexas.edu}
Center for Information Systems Management
Department of MSIS
Graduate School of Business
The University of Texas at Austin
Austin, Texas 78712
http://cism.bus.utexas.edu
Short Title (running head): Collaboratory in Cyberspace
Direct all correspondence to:
Andrew B. Whinston
Center for Information Systems Management
Department of MSIS
Graduate School of Business
The University of Texas at Austin
Austin, Texas 78712
Tel: 512 471-8879
Fax: 512 471-0587
2
Abstract:
Internet applications such as the World Wide Web (WWW) have created
the possibility of developing global collaborative platforms for
supporting interactions between professionals and academics in various
disciplines. While Web browsers such as Mosaic and Netscape have
revolutionized the way we use The Internet, we envision the need for a
theory-based approach to the development of Collaboratories on the
Internet. Based on complementarity theory, we provide a conceptual
foundation for designing Collaboratories which maximize users' value
through the judicious choice of complementary design factors. We
emphasize the need for developments in the area of ``open''
collaborative systems, and suggest that analyzing the design problem
from a complementarity theoretic standpoint can lead to useful
insights regarding the value users derive from the system. We also
describe the design and an early implementation of an MIS
Collaboratory, which uses this theoretical foundation to organize
information and to provide a forum for document-centric, multi-media
interactions between users. While the prototype focuses on the MIS
discipline, we believe that the general principles of our design are
applicable to other areas as well.
Key words: Collaboration, rich interactions, information access,
Internet, World Wide Web (WWW), complementarity, and open systems.
Recent Internet applications such as the World Wide Web (WWW) and its
front-end browsers such as Mosaic and Netscape opened up a vista of
opportunities for collaboration among academics and professionals of
any discipline on a global basis
. The absence of proprietary standards on the
Internet make it an ``open'' system, where users from around the globe
can communicate and interact with each other without the need to have
specific, proprietary end-user applications. Since the introduction of
Gopher, the first Internet application which integrated diverse
capabilities such as content and location search for information
resources, document transfer and remote access, it has become
theoretically feasible to create a world-wide platform for rich
interactions between a large number of members of any giveg and group
activities, we envision a Collaboratory more as a global platform
where people with common interests in a field can interact to enhance
knowledge and practice within the discipline. Whether a feature like
anonymity has a positive or negative impact on the interaction
outcomes is an empirical issue, the answer to which really depends on
the context of use. However, having features which provide a
face-to-face interaction capability, and then selecting only a subset
of features depending on the context of interaction will not
negatively affect the outcomes. For example, if a research article is
being reviewed electronically through the Collaboratory, it may be
necessary to mask the identity of the author(s), depending on the
guidelines set forth for the review process.
What are the design features which can provide this level of
interaction richness? We suggest that interaction richness is
affected by dynamic linking capability, synchronous and
asynchronous communication, multimedia support, and
interaction management.
Dynamic linking capability
Dynamic linking capability is critical to support interactions based
on sound arguments, rationale and empirical or anecdotal evidence. For
example, in a discussion of client/server effectiveness on an
electronic forum, a user may take a position that client/server allows
better integration of diverse computing platforms than a mainframe
environment. On an ideal electronic forum (a part of the
Collaboratory), the user will be able to create links to documents on
the Internet to support his/her stand on the issue or to counter
someone else's claim. That is, the user will embed a Universal
Resource Locator (URL), Gopher or File Transfer Protocol (FTP) link to
some relevant articles, comments or cases (located somewhere on the
Internet) within his/her posting about the issue of integration in a
client server environment.
Asynchronous and synchronous communication
While the Web does not support real time interactions, and has been
primarily used for disseminating information (by storing documents on
Web servers), the ability to interact in both asynchronous and
synchronous modes should be an integral part of a Collaboratory.
While GDSS and CSCW systems achieve this capability within a small
geographical region (e.g., on a local area network), such features
have to be developed on a global basis.
It should be noted that while other Internet applications such as
Newsgroups and Internet Relay Chat (IRC) provide the asynchronous and
synchronous communication capabilities respectively, they do not support
multimedia or hypertext linking capabilities like the Web.
Multimedia support
A Collaboratory should be able to provide full multimedia capabilities
to support a wide variety of documents and interactions among
users. While the Web does support multimedia documents, the challenge
is to bring the same capabilities to a global interactive forum.
Of course, the relative importance of multimedia support will vary by the
discipline of interest. For example, ability to handle high quality
images may be more critical in fields
such as Astronomy and Biology than in MIS.
Interaction management
Collaboratories must provide mechanisms to manage interactions in
an orderly fashion. For example, if a working paper is put up for
discussion, and if users make comments and suggestions, their inputs
will be less meaningful in the absence of a way to organize such
interactions. If the comments are organized according to their
position on the central theme and the methodology of the paper, a new
user will find it relatively easy to know what others think about the
article. Similarly, the article itself could have been written in a
structured form, whereby the authors clearly state their stand on an
issue and provide empirical or analytical support for their
position. Without such explicit mechanisms for managing discussions
and argumentation, it may be impossible to obtain the full benefits
that a Collaboratory may offer.
It is also important to be able to link related issues and ideas
discussed on a Collaboratory. As the number of discussions grows, this
feature will allow a user focusing on one discussion to examine
comments and viewpoints that have been expressed on a related
topic. It calls for a database environment which can link related
discussions, and dynamically reorganize such links based on certain
semantic criteria. Note that this linking involves automatic creation
of links between topics or issues of discussion based system specified
rules. This is different from Collaboratory users embedding links
within their comments or postings.
In our model, the cost and the benefit elements are considered as
distinct from each other. To justify this modeling approach, consider
the fact that different search engines or approaches to organizing
information may lead to the same set of documents being retrieved
(resulting in the same benefit to the user). However, one engine or
information organization approach may be more user friendly and more
efficient than the other, whereby the user will be able to perform the
search faster and with greater ease. In other words, two different
designs can lead to the same benefit (or cost) but different user
costs (or benefits).
There is a potentially high cost to the user resulting from
information search and interactions. For example, the user may invoke
a WAIS-like search engine to look up information on a topic, and be
presented with an overload of possibly irrelevant articles, which
simply happen to contain keywords specified by the user. Similarly,
meaningful interactions involving papers or projects over platforms
such as Usenet groups or Internet Relay Chat (IRC) can be tedious for
the inability to reference information dynamically, and the absence of
a shared workspace like a Multicast Backbone (MBONE) Whiteboard. This
is not necessarily a real dollar cost; rather, it is the opportunity
cost associated with the time and effort put in by the user.
With the isolated technologies that exist on and off the Internet, it
may be possible to provide a high level of information access and
interaction richness (the benefit side); unfortunately, the time and
effort (and hence cost) for the user may be prohibitively high. The
users' cost increases with poor information organization, less
efficient search engines, lack of multimedia support, the inability to
interact in asynchronous or synchronous manner, and with poor
management of the interaction processes. We argue that by combining
features and capabilities in a coordinated fashion within a single
system we identify as the Collaboratory, we can simultaneously
increase the level of information access and interaction richness
while reducing the users' cost. In the subsection below, we address
this issue of choosing appropriate combinations of design factors.
We have already discussed how the design factors are individually
beneficial for both the benefit and the cost sides in the value
equation. Now we provide examples to show how the above features
become much more valuable to the users when they are provided in an
integrated manner.
The ability to dynamically link to an information resource to
support one's stand on an issue or a proposed model is a key feature
of a constructive and scholarly interaction. For example, in a
discussion involving the business value of IT, someone taking a
position that measurement problems have led to the ``IT productivity
paradox'' could create a link to one or more relevant articles within
his/her comments. While the Web allows hypertext links, it does not
come with a platform for asynchronous or synchronous
communication. Herein lies the complementarity for the benefit
side. The benefit of having the platform and the linking capability
together is higher than the sum of values derived from having two
systems, one with the linking capability and the other with the forum.
As a second example of benefit side complementarity between pairs of
design factors, a Collaboratory that supports multimedia but does not
have an interactive feature (e.g., asynchronous or synchronous
communication) or vice versa will be less valuable to a user than one
which supports both. Consider a traditional Web page, which has
multi-media capabilities, but which does not support interactions like
newsgroups. On the other hand, newsgroups lack the multi-media
capabilities of Web based applications. The benefit of having both
features together is very high from the users' standpoint.
In a similar vein, we could take pairwise combinations of various
design factors and establish that the value of having them together is
higher than the sum of values derived from having them in
isolation. Such pairwise comparisons establish the complementarity
between all the design related factors.
To understand the cost side complementarity, we note that
theoretically it is possible to have all or most of the above features
on The Internet in an isolated manner. For example, Web and Gopher
sites already have a large volume of relevant information on virtually
every topic. The Web also supports multimedia documents. USENET
bulletin boards allow asynchronous interactions on a global
basis. Internet Relay Chat provides synchronous interaction
capabilities. WAIS, Archie, and their derivative search applications
allow users to search for documents and their locations. However, from
a user's perspective, it is tedious to use one application to locate a
document, to open another application to post a comment. In other
words, the cost of time and effort will be too high to use separate
applications for each aspect of an interaction. On the other hand,
having all these capabilities within one system can significantly
reduce the users' cost of spending time on a Collaboratory. To be
more specific, take the case of information organization and search
capabilities. Having better information organization and search
capabilities together will reduce the user's opportunity cost much
more than either of these alone. Similarly, being able to dynamically
link to, say, FTP sites, but not retrieve diagrams or pictures
associated with a document will be much more tedious (and hence
costly) than using a system which allows both dynamic links and
multi-media documents. As with the benefit side, we can continue this
pairwise comparison to argue that when all the capabilities are
present simultaneously, the cost to the user is the lowest.
In the above examples of benefit side complementarity, we focused on
how various combinations of design choices improved information access
and interaction richness (and hence benefited the user). However,
there is complementarity between the level of information access and
interaction richness as well. A Collaboratory which only disseminates
information in an organized way or only supports interactions will be
less valuable than one which provides both capabilities. Thus we find
complementarity at each level in the value model shown in
Figure 1.
It should be noted that while choosing the design factors in a
coordinated fashion in the right directions will increase the net
benefit to a user, it is not feasible (due to technological and
resource constraints) for the designers and developers of a
Collaboratory to provide the highest possible levels of these
factors. Although we do not provide a mathematical formulation of the
Collaboratory designer/developer's choice problem, s/he must consider
the tradeoff between the development cost and the users' net
value.
Given that there may be thousands of MIS Web sites on a global basis,
we envision information search to become a major obstacle to the
efficient dissemination of information. While indexing schemes will
clearly be helpful, we believe that the way we organize information on
an MIS Collaboratory should be a direct reflection of our
conceptualization of the field of MIS. We believe that MIS creates new
knowledge and value for both academics and professionals by combining
technological and managerial viewpoints. The managerial viewpoints come through
multiple reference disciplines such as cognitive psychology,
economics, organizational behavior, marketing and technical
developments (such as GDSS, knowledge representation, etc.). In other
words, the technological and managerial viewpoints are complementary,
and new knowledge is created at the intersection of IT and the
reference disciplines. This visualization is shown in
Figure 2.
*********************Put Figure 2 about here*******************
IT refers to computing and communications hardware and software. The
reference disciplines enable us to understand how managers invest and
use IT, and how IT impacts various cognitive, social and economic
aspects of the organization. In other words, for the world of MIS, the
area represented by the intersection in the above diagram is more
valuable than the sum of the values obtained from pure IT and pure
reference disciplines. Through some examples provided below, we
suggest that organizing MIS and related information on a Collaboratory
according to the above conceptualization will lead to more efficient
dissemination of information.
Suppose a user is interested in issues arising in the course of
migrating to a client/server environment. What are the technical
issues (e.g., database servers, network operating systems, and
front-end tools) to be dealt with? What are the organizational
implications (e.g., user acceptance, learning, etc.) of moving to a
client/server environment? What are the economic factors involved in such a
transition? Are there any studies that document the cost-benefit
analysis of such a move? Thus, one single issue like migration to
client/server could pose questions that refer to a variety of
different fields. The concept of complementarity and its application
in the form of organizing the MIS domain provide a solid basis for
addressing such questions.
Let us consider another issue that has received wide attention in
recent times: Commercialization of the Internet. This aspect of
Internet involves a wide variety of topics and sub-topics from
advertising to legal issues, from technical challenges to economics
analysis of pricing and resource allocation. Academics and
professionals interested in this area could possess diverse
backgrounds ranging from Economics and Computer Science to Law. If
every aspect of the commercialization of the Internet were to be
documented, it would be impossible to search through various documents
to find one of interest and relevance with a reasonable amount of
effort. However if the same issues were to be organized on the basis
of our conceptualization, then commercialization of the Internet would
fall under the MIS domain with documents annotated with links to
commercial sites and academic articles on pricing and resource
allocation on the Internet. If there is a reference to some auction
or usage-based pricing, it would be easy to categorize such
information under the pure Economics discipline. A link to such a
document from the MIS domain will shed light on the details of such a
pricing scheme. So the value to an user from such an arrangement will
be far greater than that from traditional indexing of pages by author names
or organizations.
In organizing information on our MIS Collaboratory, we found that
almost every aspect of MIS problems could be accommodated under
such a conceptual design. Academic and business articles,
announcements (e.g., conferences, call for papers, etc.), MIS
personnel information, etc. can be organized according to this
complementarity paradigm. There is one additional benefit of
organizing information in this manner. Since one of the objectives of
having MIS Collaboratories is to increase the level of interaction
with industry professionals, it is important to ensure that the latter
find useful information on a Collaboratory with minimum effort. For
example, listing research articles under faculty names can increase the
effort required on part of professionals in accessing useful
articles. It is unlikely that they will know the names of all MIS
academics and their areas of interest. It may be more useful to list
articles according to subject areas, and to have links to author resumes
from the articles themselves.
One of the most important instruments for advancing a relatively young
field like MIS is an electronic forum for generating research ideas
and stimulating debates involving researchers and practitioners. We
saw the Web as a common ground for supporting
discussions and providing pointers to
existing information resources. The richness of the Web in terms of the
data types can be exploited to the fullest by creating forums centered
around the Web as opposed to other Internet discussion systems like
the Newsgroups. Based on the theory of technology convergence and
complementarity, we conceptualized an ideal public CSCW system on
the Internet, which will allow person-to-person and group-to-group
communication in a multi-media format. The relevant tasks involved in such a
communication would include mailing, posting, and viewing rich documents, as
shown in Figure 3.
The Internet supports many applications (e.g., Usenet, email, IRC,
Multi-User Dungeons (MUD), conferencing) for interaction among
users. However, each of these tools has existed in an isolated manner,
and typically users have had to access them separately. But for
productive interactions, it would be ideal to have an integrated
application which provides users all the above capabilities in a
seamless manner.
*********************Put Figure 3 about here*******************
In December 1994, we created an MIS Collaboratory by organizing information
according to our conceptualization of the MIS field. Further, in
conformance with the complementarity theoretic foundation described
earlier, we added an on-line forum for supporting interactions between
MIS academics and professionals. While our prototype only has a subset
of the features outlined in the conceptual model, to the best of our
knowledge, this is the first development allowing global dynamic
interactions on the Web. An important feature of the forum is not only
its ability to display issues and comments (a traditional Web function
as a presenter of information), but its unique capability to allow
users to post comments on these issues and discussions. One can also
post a new issue for discussion by others.
The key feature of the WWW and its browsers that has made it so
popular is the ability to combine multiple tasks. A
browser like Mosaic or Netscape can be made to handle news and email, or
be used to invoke other applications seamlessly. To illustrate this
point in the context of collaboration, consider a setting
where a particular document is to be
annotated. Suppose an MIS professional visiting our site is interested
in making comments on a particular research article, and also wishes to
follow up on others' observations. In a traditional Web environment, this
single action involves at least two parties, namely the user and the
Web administrator. For such a user interested in annotating a
document, the typical tasks would include the following:
1. Reading the document to be annotated
2. Writing comments about the document
3. Mailing back comments
For a systems administrator, the tasks involved in putting up the document
would include:
1. Saving his/her mail as a file.
2. Converting it to a document in the desired form
3. Attaching it to his/her annotation system
All of the above tasks can be done with the available tools on the
Internet. The user could have read the article using a
newsreader/presenter, composed his/her comment/reply using an editor,
and then mailed back the article. Furthermore, if s/he were
interested in receiving comments on his/her annotation, the Web
administrator would have had to post the document back on the Web
server. However to the user, these separate tasks constitute one
related group of actions. Indeed, on our forum (which uses an
annotation system), the above scenario is just one seamless process.
The user looks at a Web page, reads the article, clicks on a hypertext
link that leads him/her to a form; then s/he fills out the comments
and clicks on a button, which automatically executes a Common Gateway
Interface (CGI) script to post the article in the relevant form on the
web page immediately. In this manner, we achieved the interactive
nature of newsgroups, but with the multi-media and linking
capabilities of the WWW. In other words, our forum simulates the
features of Usenet on the WWW, while still allowing users to embed a
variety of different URLs (e.g., Hyper Text Transfer Protocol (HTTP),
Gopher, FTP, etc.) links in their comments. These features are, of
course, in addition to incorporating inline images, audio or even
video clips. Some snapshots of the Forum are provided in
Figures 4 and 5.
*********************Put Figure 4 about here*******************
*********************Put Figure 5 about here*******************
In our implementation, the MIS Collaboratory server handles user
requests, and provides a ``back-end'' process for information
organization and interaction. As with any Web architecture, the two
main communication points are at the server and client ends. A HTTP
daemon (httpd) at the server side listens to requests, and a browser
like Mosaic or Netscape at the client end initiates communication and
displays information. We have three key processes at the server end
providing the necessary functionality: CGI processes, search engines,
and organized information repositories.
CGI mechanisms allow HTTP requests to
execute processes on a remote server through the use of a normal
browser and form inputs. These are processes which manage both
information dissemination and interaction.
The CGI processes have two core duties: document display and
request handling. Consider an MIS user who visits the Collaboratory
for the first time. The document displayer (a core part of the
httpd structure) responds to the browser's request and sends back the
main document with embedded links to various other relevant documents
and forums. If the user chooses a link to any kind of interaction, a
form would be displayed with a button to activate a back-end CGI
script. For example, if the user wishes to add a conference
announcement, s/he enters the requisite input and submits the
document. The request handler deals with this submission request and
evaluates it to determine the appropriate repository to place
the document. Once the document is placed in the appropriate
repository it sends a message to the document displayer unit to update
the index document on conference announcements. Thus, both the
information positioning and dynamic indexing are handled by this unit.
This unit is responsible for searching repositories for
relevant information.
While we have used a simple keyword based search in our prototype,
other search mechanisms like WAIS and Archie with further additions could
be used depending upon the repository type and document formats.
While our search engine only searches locally, in future refinements,
the search engine must be enhanced to
search other Collaboratory sites on the Internet
as well. Its capability should extend beyond simple textual document
search to include searches based on URLs, document types, etc.
We propose that existing search engines can be enhanced with
addition of MIME capabilities to include document types so that in
addition to a specific key word search, additional
qualifications such as document type, location, etc. can be
included as
meta-fields.
Thus, any request for search will be handled by the CGI process,
which will then invoke the search engine by passing relevant input
parameters. While the user enters input in a normal fashion, a
CGI script parses it to provide a suitable string to the search engine,
and further redisplays results in a formatted fashion.
This is the unit that deals with actual storage of documents (text,
audio, video, references to links, etc.). All core information,
information added by users and logs of interactions are stored here.
Though we have used a simple, flat file system with Unix directory
structures to store documents, it is possible to use commercial database
servers as well. Such a database would provide better indexing
capabilities and also allow construction of optimal query scripts
for the CGI processing unit. Dynamic modifications are needed in this
repository not only to add new information but also to reorganize from
time to time the ``intellectual index'' and referencing information.
There has been a significantly positive response to our MIS
Collaboratory from academics and professionals from all over the
world. Responses range from private email to postings on the system.
The users of the Collaboratory have been unanimous regarding the need
to bring together not only features of the Internet tools but also
relevant information and documents. A noteworthy observation was the
global nature of interactions on the Collaboratory. For example, more
than 25% of the discussions were initiated by users from countries
other than the United States, apart from the responses to issues
raised here.
As more users joined the forum, there were many requests for
additional features. One such important feature was to ``preview''
postings. The need for this capability can be better understood by
analyzing features of the browsers and Hyper Text Markup Language
(HTML) itself. Since the system allows embedded links, inline images
and other URLs, the posting itself has to contain tags for such
links. Since HTML is a markup language with tags, it is easy to make
mistakes; therefore a mechanism for testing the links and previewing
the formatted document is needed to ensure that the administrator will
not have to correct syntax errors. This was done by filtering the
posting to a temporary buffer before actually writing it to the
server.
An alternative to the above approach would have been to allow users to
edit their comments directly. This would require maintaining a
database, so that only the original user could alter his/her posts.
At this point, we deemed this to be unnecessary, as it would add to
the complexity of the system and might hinder a first time user from
experimenting with the system. However, if we had to customize the
Collaboratory for a specific organization, a database environment
would be essential. If an organization prefers to conduct its internal
meetings and discussions over a Collaboratory, then separate databases
would be maintained to handle such discussions with privacy protection
and restricted access. The extensions such a system can take are
remarkable. For example, we created an announcement system for group
interactions of an MBA level class. This allowed students not only to
read announcements from the professor, but also enabled them to share
information (such as recruitment schedules) with other students by
automatically posting to the announcements page.
From our experience, we found that inline images and embedded links
were popular only with Web-savvy users, who were already familiar with
HTML tags. The primary reason could be the absence of integrated
tools like a browser and an editor which supports a
What-You-See-Is-What-You-Get (WYSIWYG) environment. With browsers
gaining popularity, it should be possible to create a true
word-processing type of environment that would eliminate the need to
know the markup language on part of the end users. Another
significant issue in our implementation is the organization of
discussion or rather the need thereof. As the number of issues being
discussed grew over time, the need to organize the discussions based
on semantic criteria became more apparent. At present, the discussions
are arranged simply by their subject along the lines of Internet
newsgroups.
To better understand the needs of the Forum users, we conducted a Web
based survey where the users were requested to answer questions on
their general Internet usage and opinions on the MIS Forum. We present
some preliminary summary statistics of user profiles and opinions,
without analyzing causal relationships between the Collaboratory
design factors and user value. A sequel paper will report a detailed
analysis of such relationships. A profile of the 126 users who
responded to our survey shows a mix of academics (4.8%), students
(32.8%) and professionals (58.4%). 36.8% of the respondents have
been using the Internet for 0-1 year, 28.0% for 1-2 years, 14.4% for
2-3 years, 11.2% for 3-4 years, while 9.6% said that they have been
using the Internet for more than 5 years. It was also evident that
more people are using the Internet for work-related activities. For
example, 37.6% said that they used it ``often'', while 33.6%
indicated that they used the Internet ``very often'' for work. In
general, the profile of the respondents who used our forum was
representative of the sample we wished to test, i.e., a mix of
academics, students and professionals, who are not necessarily
Internet experts, but who have begun to use the Internet for work
related activities. Since our focus is on interactions for the
purpose of advancing knowledge and practice in a discipline, casual
surfers on the net and those who use it purely for entertainment
purposes are not of particular interest to us.
While 40.8% of the respondents felt that our Forum was ``somewhat''
of an effective medium to discuss MIS issues, 35% were of the opinion
that the Forum was ``quite a bit'' or ``very much'' an effective
medium for collaboration. However, when
we asked the respondents to compare our Forum with Usenet newsgroups
as a potentially constructive platform for interactions regarding MIS
issues, 56.8% rated our Forum as ``quite a bit'' or ``very much''
better. It is interesting to examine what existing and planned
features of the Forum the respondents considered useful. 63.2% and
65.6% of the respondents indicated that the ability to post
issues and comments automatically from any browser and the
ability to link to documents respectively were ``quite a bit'' to
``very important'' features of our Forum. These features enable
interactive and document-centric discussions on our Forum. The
respondents were not particularly enthusiastic about the ability to
embed voice and images in their comments (e.g., 30.4% thought it was
only ``somewhat important'' to embed images, while 48.0% indicated it
was ``not very much'' to ``somewhat important'' to annotate with
voice). We may infer that the importance of multimedia capabilities
such as voice and images are influenced by other factors such as
available bandwidth, domain of discussion, and the nature of
collaboration. For example, one user commented ``too much graphics,
my modem cannot handle'', thereby indicating a bandwidth
constraint. It may also be reasonable to assume that a Collaboratory
for a discipline like Biology might see more use of pictures than the
field of MIS.
For future enhancements to the Forum, it was clear that a search
capability similar to that offered by a relational database was very
much needed. 65.6% of the respondents felt that this was an important
feature and would increase the value of our Forum. Also, 60.8% felt
that there was a need for discussion management, while 60% felt the
ability to vote on issues was ``somewhat'' to ``quite a bit''
important. These are clearly some directions to be pursued in terms
of enhancing the Collaboratory. Only 8% saw the need to integrate
synchronous conferencing tools like MBONE, Cu-SeeMe with the Forum.
Having considered the design and preliminary implementation of a
single Collaboratory, let us now consider a more holistic view of the
MIS world using the Web as a collaborative tool. We envision that
there will be centers of collaboration which will act as gateways to
the linked repositories of MIS related information. This is shown in
Figure 6. The key idea is to maintain the distributed nature of the
Web, while avoiding excessive duplication of repositories. Since this is a
collaborative environment, there is bound to be some amount of
duplication of material by servers that wish to maintain a local copy
of a certain repository or even a just a single document. Also, if
there were to be only one central Collaboratory, there would be an
undue load on the server, to be of use to MIS academics and
professionals all over the world.
*********************Put Figure 6 about here*******************
While it may be difficult to determine an optimal repository distribution
strategy, the entire MIS Collaborative environment could be made up of
five to seven core Collaboratories distributed across the continents.
Further, depending upon usage load factors, there could be ``mirror
sites'' to replicate the same repositories and the ability to execute
similar search agents. While it is a topographical problem to
ascertain the geographical locations of these servers, the users are
not affected due to the transparent nature of access to these
Collaboratories.
As discussed in the motivation section, creating a collaborative
environment on the Web adds an ``openness'' to the system, i.e., a
system can be designed with flexibility and integration as key
features without being restricted by proprietary standards and
protocols. While our paper focused primarily on creating a platform
for global interactions, we did not concentrate on customizing the
Collaboratory for specific organizational requirements or business
needs. However the open nature allows for such integration; in fact, a
collaborative system on the Web can be developed to handle the
integration of workflow with messaging systems and databases. One
fruitful avenue for future research involves the use of ``normative
scripting'' in Web based collaboration to demonstrate workflow
automation both within an organization and on a global basis.
``Normative'' implies conformance to a standard or a norm. Normative
scripting refers to scripting workflow and processes in an
organization. A script in the context of the Web refers to shell
programs or CGI processes, which can be triggered by different
actions. By scripting or documenting these actions on the basis of
certain rules, an entire process could be automated. As mentioned
before, actions on the Web are primarily dependent on two dimensions,
i.e, a document and its location. A normative script adds a third
dimension to this action, by defining when, how and why a certain
document is created and where it is located. For example, consider a
Web based system for a marketing group that is collaborating on a
marketing plan for a product. This plan may have two facets, one
involving a comparison of similar products in the market, and the
other an internal decision making process which approves or rejects
the plan. A normative script will send out ``smart agents'' like Web
crawlers to look for references in the plan to existing products and
retrieve the information for the marketing group.
If a business rule of the firm requires that at least two-thirds of
the group must agree on a certain plan, then a CGI process can be
created to verify the group members' stands, and automatically message
the document to a senior manager if there is at least a two-thirds
approval. We note that while commercial groupware can handle messaging
and workflow automation, they also require adherence to proprietary
standards. By normative scripting of processes on a Web based
Collaboratory, the same functionality can be achieved on a world-wide
basis. Such scripts can be particularly useful for the on-line review
of electronically submitted research articles, where they can
dramatically reduce the turnaround time for author feedback.
Being a global and open environment, the Web holds the promise of
providing a collaborative platform for rich interactions between
members of a profession or field. Given the distributed nature of this
platform, it is imperative that information on a Collaboratory be
organized according to a shared understanding of the field of
interest. Along similar lines, a Collaboratory must also support
well-managed interactions for the exchange and creation of new ideas
and concepts. We suggested that the features of a collaborative
system should be chosen to maximize user benefits and to minimize
their opportunity cost of using the system. Using complementarity
theory, we rationalized how the coordinated choice of certain design
features would accomplish this objective. With this complementarity
theoretic perspective, we argued that an isolationist view of design
focusing only on one aspect such as media richness or geographical
scope is unlikely to be of much value for collaboration. Our survey
data indicate that users of our prototype MIS Collaboratory worldwide
have found it to be a useful means for productive interactions. We
plan to enhance the prototype with complementary features which were
considered as important by the respondents, and also along the lines
of normative scripting for workflow automation.
References
- 1
-
M. Hamalainen, S. Hashim, C. Holsapple, Y. Suh, and A. Whinston, ``Structured
discourse for scientific collaboration: A framework for scientific
collaboration based on structured discourse analysis,'' Journal of
Organizational Computing, vol. 2, no. 1, 1992.
- 2
-
F. Edgeworth, Mathematical Psychics.
London: Kegan Paul, 1881.
- 3
-
P. Milgrom and J. Roberts, ``The economics of modern manufacturing: technology,
strategy and organization,'' American Economic Review, pp. 511-528,
June 1990.
- 4
-
J. H. Bair, ``The need for collaboration tools in offices,'' Proceedings
of the 1985 Office Automation Conference (AFIPS), pp. 59-68, 1985.
- 5
-
H. Chen, K. J. Lynch, A. K. Himler, and S. E. Goodman, ``Information management
in research collaboration,'' International Journal of Man-Machine
Studies, vol. 36, pp. 419-445, March 1992.
- 6
-
G. Huber, ``Issues in the design of group decision support systems,'' MIS
Quarterly, vol.&n