Mosaic as Corporate Data Collector and Dispenser


Chris Prah, TASC - The Analytic Sciences Corporation

Diane DiGiovanni, TASC - The Analytic Sciences Corporation


Abstract

While the power of Mosaic as an internet data retriever is apparent, it is also a convenient and intuitive mechanism for local data distribution and collection within a corporate environment. Private corporate information can be protected and distributed to users possessing the appropriate rights, whereas nonprivate information can be distributed to everyone. Corporate security constraints force many organizations to construct firewalls that prohibit system access to outsiders - those people whose IP addresses fall outside the corporate umbrella. Firewalls block outsider access but do not limit access to the WWW for those inside the firewall. As such, Mosaic-client/server systems that provide and collect internal corporate data as well as providing WWW access can be easily built and securely deployed. This paper focuses on the development of enterprise wide client/server systems under this type of configuration. Our paper, while discussing the data presentation capabilities of Mosaic, will focus on the abilities of Mosaic to receive data via input screens utilizing Forms.

Introduction

The deployment of client/server applications is increasing for many reasons, one of which being the the utilization of the clients cpu and native display device. Introduction of the client software onto the client machine is very labor intensive to install and maintain as many client machines have different physical configurations (CPU, memory) and even more disparate software environments. Very generally, support for client/server environments requires about one support staff for every 20 clients in those facilities having more than a few client/server applications.

Mosaic offers a very practical and viable alternative as it is a client that can be used to run many client/server applications. As such, deployment of client/server applications only requires development of the Mosaic-centric server software. This allows development time and startup costs to be minimized - the client, Mosaic, is already coded and deployed on the three major client environments, MS Windows, XWindows, and Macintosh. Almost all efforts can be devoted to server side coding and since many developers are already versed in larger platform development, coding and testing time is kept low. Basically, Mosaic allows the deployment of client/server applications without the need to develop, code, and test the client software.

One of Mosaic's greatest strengths is its intuitive user interface that allows self proclaimed computer illiterates to navigate the internet with minimal or no instruction at all. We've used this readily accepted interface not just to dispense corporate information but also to gather corporate information.

The Marketing/Proposal Research System

Our recently developed Marketing/Proposal Resource System (MPRS) is an information retrieval system that gives our proposal writers access to relevant information necessary to develop competitive government and commercial proposals. It is based on "information bases" of various kinds of significant data. An information base is a collection of related texts (like TASC staff resumes) that have been processed to allow full-text search and retrieval of its contents. This method is much more powerful for such a system than are traditional relational databases because the data in an information base can be arbitrary in size and structure while still allowing powerful (and easily formed) queries against its contents.

This system uses the Wide Area Information Server (WAIS) as its text retrieval engine and local hypertext browsers (Mosaic and Lynx) as tools to access the WAIS information bases. The hypertext browsers allow users to navigate to various locations in the MPRS and to issue queries against the WAIS information bases.

The MPRS also allows users to download files in their native format to their local computers for editing and inclusion in other texts.

MPRS Data Sets

The MPRS consists of seven data sets generated from disparate corporate sources. Some of the data has been deemed secure and requires userids and passwords to access (we use the http security mechanism to implement this requirement). Here is a list of the data found in the MPRS and a brief description of the process flow and data formats of each data set.

  • Non-Secure Data
  • Secure Data

    MPRS Data Update Process

    A data update cycle occurs weekly which gathers data from various corporate machines for processing and incorporation into WAIS databases. Employee submitted data is merged with corporate data where appropriate, files are processed into suitable format for incorporation to the WAIS databases, reports indicating data problems are produced, and the WAIS databases are rebuilt with the fresh data.


    The Approach

    The MPRS is a collection of distinct subsystems containing somewhat related information. Providing a consistent and uniform data interface throughout these subsystems was a main objective in the design process. Data is presented and gathered from via Mosaic Forms, which are built from data received from Common Gateway Interface (CGI) Perl scripts accessing WAIS databases. The idea was to create a comfortable interface in forms, allowing users easy access to data and the ability to automate many functions. We will discuss one of the subsystems, the CBD system, in more detail.

    CBD's

    The CBD subsystem is updated every business day with new CBD's received from a vendor via the internet. We receive one file containing all CBD's and proceed to parse it into files that fall into three data categories: procurements, awards, and special notices. There are three groups of data: daily, weekly, and total. To avoid duplicate entries in the WAIS database, these groups are disjoint. As such there are nine WAIS databases that house the CBD information (three categories by three groups).

    Users can query any or all of these WAIS databases simultaneously via forms built by Perl routines. These programs have error checking routines that validate whether the form has been properly filled out. If not, a kindly worded error message is displayed on the form where the error occurred. As an example, here is the form presented to users wishing to enter the CBD subsystem:


    CBD System


    Enter User ID:

    Enter your password:


    Choose whether to:

    Build/Launch/Save a query.
    Scan CBD's.
    Launch a saved query.
    Delete a saved query.

    Press this button to continue: .

    To refresh the screen, press this button: .


    The CBD's are not secured information, but a userid and password are required to enter this subsystem; it is used to uniquely store any queries that the user may haved saved. Clients create their own user accounts via another form so no administrative burden is incurred. After completing the form, selecting Build/Launch/Save a Query will cause the following form to be displayed:


    Build Query Facility


    Category and Group Selection: Help

    Select at least one: Procurements Awards Special Notices

    Select at least one: Daily Weekly Total


    Enter Search: Help


    Save this query so I can launch it whenever I choose? Yes. No.
    Mail Results

    Enter comma delimited e-mail addresses to have the query results automatically mailed. (for example: john@tasc.com, mary@abc.com). Keep it blank if you don't want to mail the results.


    Automatically run this query every morning and mail the results? (if yes, you must enter the mail-to address(es) above)

    Yes. No.


    To process your request, press this button: .


    CBD data is updated daily and while Mosaic Forms provide an easy way to display nice looking screens, users became burdened with the daily chore of re-keying their queries. This required the ability to save a query so that it can be invoked seemlessly. Choosing the "Launch a Query" form results in those users' queries to be displayed for selection. In addition to saving queries, one can also choose to have it automatically executed every morning immediately after the new days data is added to the system. The results of the query are then sent to list of e-mail addresses entered in the "Mail Results" window. If no hits are found a message indicating that is sent to the list of recipients.

    Conclusion

    The use of Mosaic as an enterprise-wide client application package is very powerful: users do not need to learn a new paradigm for each application deployed as the interface maintains the same look and feel; data can be displayed and manipulated in its native format; security features can be readily implemented; data can be sent or gathered; prototypes can be rapidly developed . All this and the decrease in the administration costs of installing and maintaining client/server applications makes this inward use of Mosaic practical and efficient.

    Chris Prah

    Chris Prah is a Technical Staff Analyst at TASC whose current focus is the introduction of WWW features into the information architecture. He has BS and MS degrees in Computer Science. Contact him at

    cprah@tasc.com


    Diane Di Giovanni

    Diane Di Giovanni is a network analyst for the Information Systems department at the Analytical Sciences Corporation. She is responsible for the installation, management and support of Novell Netware, Digital Pathworks and NFS-TCP/IP networks. In addition, Diane is responsible for the implementation of office automation technology. Diane has a B.S. degree in computer science and is currently pursuing a MS degree in Computer Information Systems. Please direct any comments/questions to her at
    ddigiovanni@tasc.com