Local time: Friday, 05-Dec-2008 12:08:04 EST
Last update at /special/conference/apwww95,Friday, 21-May-2004 09:47:13 EST

Web Applications
for Internal Corporate Information:

a case study at the federal Department of Primary Industries and Energy Resources & Energy Group

John Larocque

information systems consultant
Janeece Pty Ltd


larocque@acslink.net.au
Abstract
Most of the attention today is focussed on the establishment of an Internet and World Wide Web presence as a means for government agencies to publish externally to the community at large. This paper looks, by contrast, at the application of these same technologies to electronic publication needs that are internal to the organisation. The main argument of the paper is that any emphasis on external applications of electronic networked publishing that is not balanced with an appropriate emphasis on internal applications of compatible and complementary applications of electronic networked publishing technology may well lead to the emergence of two worlds of corporate information management. It is argued that this could compromise the overall performance of agency programs. The specific application described in the paper is a system that advertises training courses and provides an interactive means by which prospective attendees can register for attendance. The application design is described in terms of client-server system integration using existing desktop applications combined with http, CGI and email applications. Observations are made on selection criteria for internal web applications.
Keywords
government electronic publishing internal external corporate information management client server system integration

Introduction

Earlier this year I was involved in two World Wide Web projects at the Resources and Energy Group, federal Department of Primary Industries and Energy. One of these projects, Aus Fish Facts, involved the publication of policy and research information on the Web in the conventional sense of making this information available to the world at large. In contrast with this, the other project, the IT Training web pages project, involved the use of these same technologies for internal corporate information systems.

The Aus Fish Facts pages were launched by the Minister for Resources on May 12th, quite appropriately from the National Aquarium in Canberra. I want to recommend to you the Aus Fish Facts web URL as an example of useful and accessible information about one of our most important national commercial and environmental assets. Keep it in mind especially if you have kids in year 10 and above looking for a school project. You can find it on the Web at http://www.dpie.gov.au/dpie/fisheries.html - but I will not be speaking to you today about the Aus Fish Facts web.

Apart from recommending it to you, the reason for mentioning the Aus Fish Facts web is that the contrast between it and the IT Training web gave me the idea for this paper.

Most of the interest in the Internet today is focussed on the World Wide Web as a way to publish information to the world outside the organisation. In the classic sense of inter-networking this involves providing access to information across network domain boundaries by serving this information out from a web server that is visible to the rest of Internet.

But the IT Training web site serves clients entirely within a single network domain. It is hosted on the inside of the DPIE firewall. This realisation led me to ask about the motivation for the project. The answers are quite interesting, and I will return to that after presenting to you the facts of the IT Training web site.

My approach here will be self-consciously bottom-up. I will first take you through the application itself. Then I will move up a level of generality and look at the architecture of clients and servers involved. Finally, I will share with you some observations from an organisational and human point of view to draw attention to the policies and principles related to the selection of this kind of web project.

The IT Training Web

Along with other services, the Resources and Energy Group IT Unit provides training sessions on a regular basis to about 250 people. Several one and two hour sessions are given each week in aspects of all of the corporate desktop applications like Word and Excel. To support this, the training coordinator needs to advertise up-to-date session details to the department and to correspond with attendees.

This is what it looks like to the people involved:

The administrator's web page also provides the training coordinator with various reports and detailed instructions relating to the management of the web pages (examples not shown).

Architecture

From the point of view of the integration of specific client and server components, employing web-based technologies to support the IT Training web has resulted in the following approach:

Training Web architecture

Why Excel?

Web enthusiasts will have noted already that the data capture function was not built using a data input form on a web page and are probably, for this reason, already thinking that the design is seriously flawed.

The reason that an Excel spreadsheet is used for data capture is simply that this is how the training coordinator was already doing it. What we have done is to treat the Excel spreadsheet as a building block to be integrated with the web application.

Our primary aim was to make the interface with session attendees interactive without increasing the workload of the training coordinator. There would be no productivity benefit for the training coordinator in merely automating data capture in a different way. By leaving the Excel-based data capture alone the new system did not require the training coordinator to change her tool nor did we incur the additional development costs of building a web page data capture form.

What has been automated is the update of the web page advertising the current state of training session details. To do this the training coordinator has to two things:

The http server then passes the name of the text file to a CGI script which reads the input file and re-writes the training sessions web page.

Why AppleScript?

It felt intuitively right to use AppleScript in an all-Mac environment. MacPerl was considered, and MacPerl is a good port of Perl 4, but anyone who has had the pleasure of using AppleScript finds Perl to be a beast of a scripting language.

I would say, parenthetically, that if Apple does what it says it is going to do and makes AppleScript available on Unix with no run time performance penalties, then AppleScript will become a serious candidate as a scripting language for CGI developers.

But our experience with AppleScript is that it does slow things down. Out of curiosity, I rewrote the acknowledgment CGI in MacPerl. Response times fell markedly.

Why Eudora?

The Resources & Energy Group has a large network of Macintosh desktop machines on an Ethernet backbone. The Ethernet carries TCP/IP and AppleTalk traffic. The local email system is QuickMail and the network has an SMTP/QuickMail gateway server to redirect internal (Quickmail) and external (SMTP) mail packets.

A design objective for the training sessions web was to deliver expressions of interest from prospective attendees to the training coordinator as email that showed up in her QuickMail mailbox. Prospective attendees make these expressions of interest by filling in check boxes on the training sessions web page and submitting their choices as a form (see illustration 1).

We have a classic client-server system integration problem here. The Netscape web client and the http server communicate over the TCP/IP services. The QuickMail server and clients communicate over the AppleTalk services. The question is how and where to bridge these services.

The direct approach would have the CGI script that parses the data stream from the expression of interest form to pass that information on to a QuickMail client running on the same machine as the http server. Unfortunately, the QuickMail client is not scriptable.

We could, theoretically, have explored QuickMail internals and inserted files in the QuickMail client environment and set the QuickMail client to look for new mail to send. But this would have cost time to investigate and implement.

Instead, we again took the building blocks approach and made use of the existing SMTP/QuickMail gateway. To do this, we installed an SMTP client (Eudora) on the http server machine and used that to send mail to the gateway.

We chose Eudora because it is scriptable. The CGI that parses the data stream from the expression of interest form, passes that information on to the Eudora client.

Proxy Server Interference

The IT Training web was up and running within a few weeks. However, one day, quite out of the blue, the expression of interest form stopped working.

Based on the MacHTTP access log we were able to diagnose this fairly quickly to a fault with the DPIE http proxy server. All of the Netscape clients had been configured to address http requests to the proxy server so that requests destined for servers outside could traverse the DPIE firewall.

The proxy server in use at DPIE is the CERN httpd. Although it has now been fixed in later releases, there was a bug in the CERN httpd server that prevented POST method arguments from being sent through the server.

Getting a one line fix to the httpd server has proved to be more difficult than one might reasonably expect. I have to assume that only internal communications are affected by this bug. I do not want to make a big fuss about it here, but the attitude seems to be that the firewall and the web are all about external communications. I will ask you to recall this observation a little later.

Implications for corporate information management policies and principles

I want to shift gears again now and look at some of the more general management implications of the phenomenon represented by this practical example.

What we are seeing here is not just the emergence of the Internet and the World Wide Web. What we are seeing is the penetration of globally networked electronic publishing into the mainstream of corporate life.

The significance of this phenomenon is that the application of these emerging technologies can have an impact on recurrent and capital expenditure. It also raises questions relating to the organisational reforms that can accompany the introduction of significant new technologies.

Risk of two worlds of information management

I want to draw your attention to the danger that the current emphasis on an external web presence might result in a separation of the internal and external corporate information worlds.

Organisations are now starting to feel the full effect of decisions they made in the early 1990's regarding the moves they made then, or failed to make, to position themselves for networked computing in the mid-1990's.

Those who adopted strategies to build infrastructure for distributed open systems based on multi-protocol networks are the ones who are finding, now, that they can take on board the opportunities for improved organisation performance that are now being presented by the World Wide Web and related internetwork applications.

Those who failed to position themselves in the early 1990's are looking at significant costs and delay.

The R&E Group of DPIE is one of the lucky ones who had made the requisite technology investment decisions. So they were already in a position to start the process of introducing the web and the Internet to people in the Department when this became a serious option this year.

However, if you examine our case of the IT Training web project, you find that you can create an interesting hypothesis concerning a potential exposure to increased human resource costs. This is due to a potential loss of consistency in the user interface across the tools provided to people on the desktop when you introduce access to external online services. Analysis of the conditions surrounding the formation of the IT Training web project illustrates this quite well.

The R&E Group, like most organisations, had just spent the last decade trying to achieve some kind of integration of corporate applications and standardisation of user interface. This effort of the past decade was based on the principle that a consistent user interface at the desktop reduces training costs and improves productivity.

Anticipating the need to introduce emerging online networked technologies, R&E Group put the latest version of Netscape on every desktop and can now, on a selective basis, make it possible for any or all of these web clients to request services outside the department's firewall.

The look and feel and functionality of web-based applications is different from those we now have as our corporate standards.

There is thus a danger that as use of the new online external information services increases, the people on our desktops are going to start seeing 'two worlds' - the internal world of integrated corporate applications with its standard user interface on the one hand, and the external world of inter-networked information services with a different user interface on the other.

There is another way to look at this, too. Bifurcation of corporate information management into the internal and external worlds may also be driven by the way that web projects are initiated and managed.

Earlier this year I conducted an informal survey of people involved in web building across the public sector. The people I contacted represent a cross section of functions from traditional IS through Library services to policy and corporate communications areas. I did that survey because I wanted to understand what was driving the push for World Wide Web presences for government agencies.

I covered this in some detail in another paper that I gave earlier this year in Canberra so I will not dig over that ground again here. However, I need to recall that one of the main results of that survey was that on the whole, the push for a Web presence for government agencies is coming from the librarians, the information resources and the public relations sections and in some cases from specific policy areas within agencies.

Since doing that survey in April this year I can now add another feature on the basis of first hand observation. It appears that the creation of web presences for government agencies is happening in a bottom-up fashion in organisational terms. The people involved in the creation of agency web presences are 'electronic publishing-aware' middle managers who are quite self conscious of the fact that their executive management has yet to realise the significance of what they are doing in this regard.

This is not to say that executive management is unaware that web presences are being created. I should also add that this in an observation of several departments not just one.

The reason for mentioning this is that I believe that in the next twelve months, executive levels will begin to take on more active ownership of the electronic online aspects of the programs within their care. Further, I believe that the external aspect of electronic online communication of information will tend to eclipse the internal aspect of online communication for corporate information management.

This is a natural result to expect. The people who are now doing the ground-breaking work within agencies are focussed on an external web presence. They will telegraph their understanding of the justification and enthusiasm for an external Web presence to executive levels.

As it happens, the R&E Group IT Training web is an exception in this respect. But even there, in an agency that is looking at both internal and external applications of this technology, the internal application gets second billing when it comes to something like the proxy server interference noted earlier.

If you accept that there is a possibility that two worlds of information management might arise because of the online electronic publishing phenomenon, then we also have to ask whether there is anything essentially wrong with that. I think the common sense answer here is that it does matter a great deal that agencies achieve a high level of integration across the internal and external information boundary.

Lack of such integration will inevitably lead to:

Such things easily translate to reduced program performance.

Risk abatement and benefits management

There are several antidotes to the exposures identified above. The practical approach advocated here is to identify opportunities to undertake projects similar in nature to the IT Training web project.

When selecting such projects:

It has to be emphasised here that an agency must already have positioned itself, technologically, to pursue such projects. I know that some agencies like DPIE on the whole are in a position to do this because they consciously set about creating the necessary strategic infrastructure several years ago. Other agencies could not see that far ahead or were incapable of taking advice and as a consequence now have to be followers in this respect.

But just having the technological capability is not enough to cause the impact of these technologies to actually show up in terms of program performance. To do that you have to also be in a position to change the way the work is organised. This is what, I think, people mean when they use terminology like business process re-engineering.

In that informal study mentioned earlier, I found two perceptions that were shared fairly evenly across the board. These were:

It is really the second point that is important here. Our pilot projects are demonstrating the need to employ a wide range of skills and services in order to make use of highly integrated computing and communication systems. These skills and services actually already exist in different parts of most organisations. The management challenge is to nurture their effective combination and utilisation.

Policies, guidelines and staff development programs have to be put in place to assist those policy or program areas that wish to apply the sort of technology represented today by the Web. This is most unlikely to be valid at the department level, but is more likely to become valid at the major program or division level in the organisation.

The example of this that I know is the Australian Geological Survey Organisation (AGSO), which exists in its own right within DPIE in the same way that the Resources & Energy Group does. AGSO has now begun to implement an information management strategy based on the consolidation of the traditional IT area with its Library services, publications and graphic services.

There may be other cases. I only know of this one because of my involvement as a consultant at the time that AGSO executives were forming that strategy. I mention it here because it is important to realise that just getting to a point where it could become part of a restructure took several years. The first time that I became aware of the idea was back in 1990 when we were formulating a broad set of strategies to position AGSO on an open systems technological footing. The idea of consolidating various areas was advanced at that time but failed to find expression in the plans of the day. It was not until I was asked to return to revisit the information management strategies three years later in 1993 that I found that the idea was still alive and that it could now be expressed in planning terms.

I tell you this as an illustration of my belief that there is no magic in shrink-wrapped technologies. There are no easy, quick fixes taking advantage of the technologies that have brought us to this conference today. If your organisation has not positioned itself on infrastructure, then you are going to be disappointed if you are looking to the Web and electronic online publication as a way to improve program performance. Still, it is never too late to start. In fact, the advent of online networked electronic publishing presents another opportunity to begin the processes of change.

[Return to Table of Contents]
COPYRIGHT © 1995 by AUUG95 and APWWW95 Charles Sturt University. ALL RIGHTS RESERVED.