The Web as Application Development Platform

These days, you hear a lot of talk about how the Web is going to become the next great application development platform. "It will free us from processor- or brand-specific applications," say most of the advocates. Mostly this seems to mean freeing us from the Windows/Intel ("WinTel") monopoly.

For the most part, the web-based applications we've seen to date have been publicly-accessible databases or query tools published on some of the more technically-aggressive web sites. These systems generally allow you to search for technical materials, or to open a support call using web-based applications. Other more common implementations are infrastructure-related management tools, allowing you to configure your router, hub or printer using firmware-level HTML servers embedded into the hardware directly.

Everybody agrees however that in order for the technology to successfully displace WinTel, it has to be accepted and implemented on the corporate Intranets for use by the rank and file. The theory is that this type of application development platform will open up whole new ways of thinking about internal application development, as well as new markets for "smart" terminals.

Mainframe-based applications that typically used 3270 terminal functions could now be implemented as HTML with CGI calls to DB2 and Oracle databases. Lots of large companies like Chrysler and American Express are already experimenting with these types of applications, hoping that they can leverage the centralized nature of the data and application code, thereby avoiding most of the problems encountered with distributing these elements across the wild network frontiers. Appropriately, IBM and Oracle, among others, are offering CGI interfaces to their database engines, either for free or for very low cost.

But these types of applications are for the most part extremely vertical, and will only be used by call-center and data-entry personnel who are very task-centric. This is not a sizable enough purchasing community to make a significant market. None of the current vendors of terminal hardware (TeleVideo, Wyse, etc.) are going to be missed much when they get replaced by web browsers. In order to make a market, there need to be web-based versions of mainstream applications, like financial management and inventory systems.

Guess what? They're out there already. Companies like Action Technology and Computron Software are both demonstrating web-based implementations of their client-server application suites. Their customers can deploy the client-side of the applications using any web browser they choose, all while centralizing the application code and data on a large-scale, secure server.

Action's Workflow Metro package is available for demonstration and browsing at http://www.actiontech.com, for those of you with deep enough pockets to be interested in buying it. I played with it a little bit, and was fairly impressed with what I saw. It needs more time in the oven, but it looks to be promising technology. You can fill out a requisition form online -- which normally would have to be filled out on paper -- and click a button to submit the form to the appropriate party for approval.

Very slick, but again it too is very vertical in nature, and best kept to the larger organizations that have so many employees that they have to worry about efficiencies of scale. These large-scale applications just aren't mainstream enough to make a market. What's needed are end-user productivity packages.

Guess what? They're out there already. Campbell Services is preparing a web-based version of OnTime, their excellent group scheduling package. By installing a pre-built CGI-based scheduling server onto your Intranet server, you can give personal and group calendar capabilities to all of your users. Check out their online demonstration at http://www.ontime.com. The first version is read-only, but Campbell expects to have full read-write capabilities within the next few months.

Even personal, end-user products are becoming available on a weekly basis. Visual Components, Inc. has just released a beta of their spreadsheet plug-in for Netscape Navigator. The product, called Formula One, provides Excel-compatible spreadsheet capabilities to users, allowing them to perform arithmetic and graphing from within the browser directly. Users can save the spreadsheet to a file, or they can embed it as a link within an HTML page. Other vendors are developing similar components, and the promise of Java as an end-user development environment ensures that we haven't seen the end of these types of personal applets.

So, we see that there are web-based applications available for large-scale, internally-developed vertical systems, usable enough to be deployed by international conglomerates. We've also seen that there are off-the-shelf products available from some of the mainstream client-server development houses, suitable for use by most medium-sized organizations. We've also seen that there are even end-user personal-productivity products, capable of providing some of the functionality that you get in Microsoft Office.

Having applications available across all of these levels seems to make a market to me. I'd even go so far as to say that this is just the beginning, and we're likely to see a lot more of the mainstream, typically Windows-based client packages becoming available in a web-based version.

The question for you then is "So what?" Will you use the web and web-based applications for your internal development or deployment? Or are you happy with what you've got in place now, and don't need anything else? I don't think you'll have to worry about making any major decisions anytime soon, but you should begin preparing yourself for these and other questions. There's definitely a market being born.

Written by Eric A. Hall.
Copyright © 1996 EHS Company.