The Flip Side of Web Applications

Robert Hess
Developer Relations Group
Microsoft Corporation

December 14, 1998

The following article, which was originally published in the Site Builder Magazine (now known as MSDN Online Voices) "More or Hess" column, continues a series in which MSDN Online columnist-at-large Robert Hess investigates application development in the Internet Age.

No matter where you turn, it seems everybody is talking about writing "Web-based applications." In a series of articles, I've been trying to examine various aspects of this, hopefully providing you with things to think about and areas to examine as you work on your own solutions. Last month, I talked through the general aspects of what we refer to as a "reach" application, or one that focuses on trying to be accessible to anybody, on any operating system, with any reasonably modern Web browser. As you can expect, there are obvious benefits to such an approach, but there are disadvantages, as well. This month, let's take a look at the flip side.

With all of this talk about the Web, it can be difficult to realize that the infrastructure opened up by the Internet can also work great for applications and solutions that don't need to be hobbled by trying to be "totally compatible" with every single Web browser on the market. I receive a lot of e-mail from people who are trying to solve various problems with Web-based applications; sometimes their approaches sound as though they are trying to shove a square peg into a round hole.

It all boils down to that age-old saying "The right tool for the right job." While a tool belt with a hammer, saw, pliers, and screwdriver (okay, and duct tape) might represent the simplest, fully compatible solution, there really are times when the best way to get the job done is with a tunneling electron microscope.

Layered Solutions

One of the deployment concepts behind Windows Distributed interNet Applications Architecture (DNA) is the notion of a design architecture that allows a solution to be layered for maximum effectiveness and deployment flexibility. A primary focus of Windows DNA is to create a conceptual separation between the "Presentation", "Business Logic", and "Data Storage" functionality of your solution, but it also allows you to take maximum advantage of a modular and object-oriented approach to your solution by leveraging COM in a language-agnostic model for interacting with a wide variety of system services. Using a Windows DNA approach, you can expose your application processing logic via broad-reaching Web pages, but you can also leverage this same processing into a time-critical, event-driven solution by allowing customized client applications to call directly into your server processes without having to use a browser interface as middleman. This means that you, the application programmer, have both the freedom and the responsibility to determine the solution that works best for you. It also means that as new and different deployment needs arise, you can easily leverage your existing code base to address scenarios that you hadn't originally considered.

There are perhaps two specific scenarios in which a custom application is appropriate. One is for an application that needs to allow rich interaction from the user (such as with an image-processing application); the other is to provide the functionality of a dynamic Web site, but in a offline state.

Let's examine these two scenarios individually.

User Interaction

Consider the case of an image-processing application. Let's say you work at a digital graphic design company. A Web site and associated database have been put together to warehouse and catalog all of the images being worked on throughout the company. Using this, you can quickly and easily search the gallery of images, view collections of thumbnails, select individual images, check them out, and then after you've finished whatever processing you needed to do to them, check them back in again.

Obviously, the weak link in the chain here is that between checking the image out and then checking it back in again, you need to rely on the user to properly manage the process of bringing the file into a graphical editor, then making sure they check it back in when they are done. Even the check-out/check-in process can be a little awkward without relying on ActiveX® controls or Java applets. One way to resolve this would be to develop a small application for viewing and checking the images in and out. This app could even use the IWebBrowser control so that existing Web pages could be viewed within the context of this application to locate the images. When the proper image was located, the application would interact with the server to mark the image as "checked out," would copy the file locally, and would automatically launch the appropriate editing application. When the checkout application noticed the image file had been closed by the user, it could then ask the user if he wished to check the image back in or not.

Such "source control" support could be integrated into an existing (or in development) application in various ways -- the point here being that this level of interaction really is best performed outside of the scope of an HTML-based Web page, but it can still interact with a back-end database, as well as with any special purpose "search" logic that might have already been developed for a Web-based solution.

Working Offline

In the case of an application that might be a normal Web site, aside from the fact that it needs to operate in a "disconnected" state ...

Let's say your company has put together a company-wide contact-management database in which you can look up, modify, and store information about all of the various companies and individuals that you deal with throughout the day. This can be done as a pretty straightforward "forms"-based Web site. Aside from what might be a rather complex database schema, it is just fairly run-of-the-mill search forms, information pages, and form pages to allow new/modified information to be submitted back to the database. But what about a situation in which a user takes her laptop with her on the plane, and wants to be able to not only view, but to submit, modifications to the contact database?

One approach would be to develop a custom application that could work in both the connected or disconnected state (although a Web-based approach would still work fine in a connected state). In the connected state, the application would simply call straight through to the back-end business logic components to provide appropriate access to the contact information needed, and would submit any modifications in real time. However, when the user wanted to go on the road, she would be allowed to populate a temporary "snapshot" with a small subset of the database, and the custom application would switch to "offline" processing, in which it would allow the user to interact with only the locally stored entries, and perhaps with limited functionality. When the user went back online, the application would automatically synchronize changes with the master database.

Remember the Client Side

The above are only a couple of scenarios in which breaking out of the traditional HTTP/HTML metaphor is perhaps the right thing to do, even though you are still taking advantage of the same infrastructure that might be used in a Web-based environment. As you design your "Web-based" solutions, keep this in mind, and be able to recognize when it is appropriate, as well as perhaps necessary, to create a richer client-side component to solve the problem you face.

Robert Hess is an evangelist in Microsoft's Developer Relations Group. Fortunately for all of us, his opinions are his own.