Question

From Estimation to Delivery – thoughout the software development life cycle,

1. Which all documents are involved and
2. What is the order?

I am not sure whether methodology have much impacts on documents, anyway let us consider Waterfall.

Answer 1

The answer is – as has been stated – it depends. I’m sure lots of people will answer for Agile methodologies (which are a far more movable feast) so for completeness I’ll go with what you’d have for a fairly standard waterfall methodology:

• A scope document – very high level outlining what is and more importantly what is not in scope of the project and what assumptions are being made. The main purpose of this document is to set expectations of what will ultimately be delivered – you’re not saying how things will work but you’re trying to answer questions such as will there be reporting? Will it pass data to other systems? Will you have to write your own user management functionality or pull from AD? Where you can’t get definite answers to these things then include an assumptions section and list what you’re assuming will be the case so people can correct you if you’re wrong. It should also include things like target implementation dates (not as a commitment but so people know what is anticipated and manage expectations accordingly).
• A functional specification – What the application should do on a business level. This can be divided out into business requirements (the business processes it’s automating and how they work) and functional requirements (what the system does and how it does it – screen navigation, how calculations are made and so on) but more commonly they’re combined except for the largest systems. It should also include “non-functional” requirements such as performance, load, security and so on.
• A technical specification – The most likely to be missed out. A detailed technical design including things such as object models, schema diagrams and information on how detailed technical problems are being addressed.
• Test plans and test scripts – How the application is being tested with detailed test cases, data and expected results, covering all elements of the system.
• User guide and Release Notes – How to install, configure and use the application.

The one I’d add to this is a support document – a short (less than 10 pages) crash course in what the app does and how it does it. Developers will often not read the full specifications (either because they don’t have time or don’t want to) so this document should be enough to allow them to understand what it does, how it works, the areas of the application which are most likely to be problematic and so on. It would be written a few weeks after go live by the team who built and implemented the system.

Of course depending on your methodology you may have none of these documents but if you’re running a standard project in an old school structured, waterfall way, this would be pretty normal.