Operation Red-Nose Project

About 13 months ago, I started working on a project to replace the software used by the local Operation Red-Nose (ORN) organization. Up until then, they had been using an old DOS-based system that was long past its best-before date. In fact, they were only using part of the system, and even that had major usability issues. The project was suggested by my supervisor at the MCF Practice Firm, a volunteer at ORN. The purpose of the project was for me to pick up some specific programming skills. At the time, I couldn't complete the project since I got a job offer five weeks into the program. But that job didn't last, and I was able to return to the ORN project. (During the summer, federal funding for the practice firms was withdrawn, and so MCF was forced to close.)

For the ORN project, I decided on CakePHP and MySQL. I chose PHP and MySQL since they are pretty much ubiquitous in the I.T. realm. For the framework, I quickly narrowed down the choices to CodeIgniter and CakePHP. I chose the latter since it strongly encouraged the use of an MVC design. The ORM also impressed me, with its ability to get all the data associated with a transaction in one operation (typically). That is, all the SQL joins were taken care of automatically.

In a nutshell, the ORN application manages many aspects of the operation, in particular, the event nights during the annual campaign. The system keeps track of volunteers. During an event night, volunteers are checked in to the system and assigned to teams. Operators take calls from clients and enter their ride requests into the system. The dispatcher assigns the requests to the available teams, and tracks the progress of the requests. And at the end of the evening, the treasurer tracks the donations.

An essential part of the development process was demonstrating the system to the ORN volunteers and getting feedback. After each weekly meeting, I'd normally have several pages of suggested improvements. The main issue was understanding how he volunteers would learn the capabilities of the system, which wasn't always easy for a programmer geek like me. As the programmer, I knew how to get things done. But for others, especially those with little computer experience, it wasn't always clear.

Add to that some language issues. ORN Kingston is managed through La Route du Savoir, an organization that provides training and services to French speaking people in the Kingston area. Although everyone at ORN speaks English, one person had a bit of trouble with some of the more technical language I used. Often, someone had to translate my technical language into a form that that person could understand, and her requirements into language I could understand. But we managed, and the result was a system that everyone was happy with, even if it took a few weeks after the start of the 2012 campaign to get all the required features in place and the defects fixed.

Is there more work to do on the new Kingston ORN system? Sure. The data entry form for the telephone operators can be improved. I took a few phone calls myself, and readily saw one problem with the layout of the entry fields. For the 2013 campaign, I'll rearrange the entry fields to reflect the order that questions are asked of the callers. That is, if the first question is "Where are you?", the pickup location should be the first item on the form. In general, that form should have a list of questions for the operators to ask to ensure that nothing is missed.

How did CakePHP work out? In retrospect, I think it was a perfectly reasonable choice. At no time was there anything that CakePHP couldn't handle. There was one annoying issue with session timeout. However, upgrading to version 2.2 addressed that problem. The upgrade didn't go without some effort, though. When I started the project, I didn't always follow the preferred naming convention. For version 2.2, I had to rename my controller classes.

What about MySQL? The application currently uses 16 tables, which is puny compared to other production systems. Although I prefer more robust DBMS's, I could clearly see how MySQL has matured over the past few decades. The InnoDB engine (now the default) is sufficiently robust for this type of application, supporting foreign keys and transactions, vital features in even the smallest of databases. In contrast, today in the year 2013, there are still production MySQL databases that use the old MyISAM engine. In my opinion, any database administrator who entrusts vital company data to a MyISAM database should be fired immediately.

If I were to start the project today from scratch, would I do things differently? Styles of programming for web applications are always changing. For the ORN project, I designed it as a fairly traditional server-based web application, with just a bit of client-side programming. But the trend is to move more processing to the client, with the view and controller components on the client and the model on the server. But for this type of application, I don't think I'd do things any differently.