If You Want Your Project to Succeed
You must perform these essential activities if you want your Framework project to succeed:
1. Build a prototype and get it reviewed and formally agreed and signed off by both end-user representatives and developers. Users will know what they are getting and developers will know what they need to deliver.
This is a basic form of expectation management and also goes some way towards managing classic IT project problems such as lack of user engagement and scope creep.
2. Publish a Minimum Supported Configuration Document – where you formally state define the minimum configuration your solution will viably support, including what servers, client platforms and web browsers your application will support:
- Minimum hardware requirements (see Computer System Requirements)
- Minimum software requirements
- Supported screen resolutions
- Minimum networking capabilities
- Maximum data volumes.
A formal MSC will:
- Inform decisions about the overall solution cost
- Establish the environment required to test the deployment of the solution or any patch/hotfix made to it.
- Raise management's awareness of the risk in implementing a "sub-MSC" solution.
For more information refer to Application Performance.
Regularly performance test to the minimum platform while doing development. Maintain and republish this document during the project.
3. Publish a Business Value Proposition – where you formally define the business value of your application, especially where an existing IBM i 5250 application is being modernized or replaced.
Formally state how and why your application will make doing business better and/or faster and/or smarter. If you can’t define the application’s value proposition in words and pictures it is extremely unlikely you will be able to implement it in software.
The introduction of visual components by themselves (ie: things like radio buttons, drop downs, menu bars and colour gradients) rarely represent significant business value. Recidivist end-user behaviour is a common sign of low, absent or poorly explained business value in IT projects.
4. Make time in your project plan to for deployment and testing. You would be amazed how often this area is discounted by developers as something that "will only take a few days". This is classic area where IT projects experience cost overruns.
You need to factor in time to design a deployment strategy, implement it and then test all the supported platforms – in addition to your normal application level unit and suite testing. Optimally you should have human and hardware resources solely dedicated to application testing.