Jan Burse, created Oct 04. 2009
Do you know, what I did this summer? I will confess everything with relentless openness. I lent and/or bought books concerning the programming of user interfaces not related to AWT, Swing or SWT. have bought computers on which either no Windows was installed or where I have installed something else than Windows. I programmed in programming languages different from Java, JSP, HTML or SQL. I admit doing this daylong, weeklong and months-long. What drove me to this act of horror?
Well, it is planed that my sales system delivers different virtual products to a multiplicity of platforms. Unfortunately one is a little bit lost when counting on Java. Admittedly Java helps in the creation of the sales system itself, since it is possible to develop across multiple platforms. The same holds for the development of executable products. But how does the product and the Java arrive at the client side?
There is a big gap. Although a browser makes it possible to download a single file, there is little help when the product should run offline, when it is structured into multiple files or when third party components are needed. An installation program has to be created which automates the necessary steps for the customer. Unfortunately the freely available installation program generators were not convincing: Weak adaptability paired with high labour costs. We therefore decided to write our own installation programs, if nothing else to become able to understand the topic better.
The last 12 years of XLOG were mainly dominated by Java development. To create autonomous installation programs however we had to keep our fingers off Java. It began a time journey into the past. Old already dead believed acquaintances like the native platforms of Windows, Mac and Unix showed again their faces. I believed myself carried back to my study time. The affair was very much time-consuming since quite a bit had to be established, but everything run surprisingly smooth. The various platforms had improved over the years and needed specialities such as self-extracting, concurrent user interfaces and script execution were relatively simple to realize.
We needed half a day to elaborate the use cases for the installation programs. The realization now goes on for 3 months part time and is finished up to 80%. The result is quite presentable, nevertheless some question marks remain. How well will not only be the product received but also the installation programs? Will it be necessary to implement measures to increase the acceptance?
Before we continue agonizing about acceptance, we will proceed to the next step and start with production. This means adapt the product development to the insights of the development of the installation programs and to complete both. To finally gain empirical data on the acceptance of our procedure by a product start.