Completion Report

Project Summary:

This project's key purpose was to build on the basic functionality delivered by the first RETAIN project so that the application could bring benefits to a wider research user base across the University.  The key areas needing enhancement were: -

• Functionality to record legislative data for bio projects and substances, both GM and non-GM. These new features involved enhanced security options so that this sensitive data could be restricted to authorised users
• Enhanced radioactivity functionality to allow researchers the option to work in volume or radioactivity when recording containers, aliquots and disposals
• Enhanced BI reporting to assist with ad-hoc queries from Schools, Government bodies and Police Scotland
• Enhanced standard reporting for School admin staff and researchers

The project also took the opportunity to migrate the application from Coldfusion 9 to Coldfusion 10 for future supportability

The enhanced radioactivity work was extremely complex and had impacts on many other areas of the system, so this element of the work took much longer than anticipated and meant that the reporting enhancemements had to be de-scoped during the later part of the project.

Analysis of Resource Usage:

Staff Usage Estimate: 190 days

Staff Usage Actual: 191.7 days

Staff Usage Variance: 1%

Other Resource Estimate: 0 days

Other Resource Actual: 0 days

Other Resource Variance: 0%

Explanation for variance:

The project came in on budget due to clear prioritisation of User Stories by the Product Owner in Health & Safety in line with agile principles.

Key Learning Points:

What went well?

• The project team (both IS Apps and Health & Safety) worked well together and were able to deliver a large amount of key functionality for both bio & radioactive users
• The Business Reps from Health & Safety were very effective at prioritising (& de-prioritising) User Stories within the available days which allowed us to make maximum usage of the time available
• Having an Agile coach on the team helped us to stay on track, particularly in the earlier development iteratations
• Having a Business Analyst on the team (rather than the role being undertaken by the Project Manager) meant that backlog grooming and QA were undertaken quickly and effectively
• Our Dev Tech rep was much more involved than on previous agile projects and attended daily stand-ups at key points in the project.  This helped them to understand the project and promoted team culture
• The gathering of conditions of satisfaction by the Business Analyst & business reps worked particularly well which meant that User Stories were clearly specified prior to development starting
• The migration from Coldfusion 9 to Coldfusion 10 went very smoothly

What didn't go so well?

• Automated testing & documentation were sometimes deferred in order to deliver more functionality within an iteration.  This meant that User Stories weren't properly 'done' within the iteration and new stories had to be created to complete these key activities in later iterations
• Resource conflicts from other projects (STU236 & timetabling in particular) not only resulted in lost days but also interupted the smooth flow of work as many of the User Stories were too complicated for one developer to work on in isolation, particuarly as one of the developers was new to RETAIN
• Towards the end of the project the agile coach was less available due to other priority work which allowed the team to drift away from strict agile processes (e.g. knowing when a story is 'done')
• The developers sometimes directed queries via the Business Analyst rather than directly to the business contacts which delayed responses and diluted the responses
• The maintenance iteration became extended due to the complexities of the radioactivity stories.  This meant that we weren't able to address the BI/Reporting requirements

What would we do differently next time (lessons learned)?

• More up-front design - we would benefit from more up-front design for major system changes so that we can walk through the different implications and scenarios.  This wouldn't have prevent the large amount of development that was required for 'working in activity'  but would have highlighted the extent of the work much earlier and allowed it to be addressed in a more planned manner
• Independant agile coach - it would be much easier for the agile coach to guide the project if they had no emotional investment in the project (i.e. had not developed the application) as this would allow them to be more impartial in their decisions
• Real user testing - we would involve real end users in the testing to ensure that we had captured any gotchas from real-life scenarios
• Data refresh prior to project start - a data refresh from Live into Beta & Test would help to ensure that we were testing against real data examples. This would both prevent time being wasted on issues relating to poor test data and would ensure we were testing against real-life complex examples
• Standard test scenarios - due to the complex radioactivity calculations in the system it would be beneficial to have documented test scenarios with expected outcomes to assist with regression testing, both for automated testing and QA
• Set-up of proper Test EASE accounts for UAT - users undertaking testing should be set up with proper EASE accounts to ensure that data is kept secure, especially if a data refresh from Live has taken place

Outstanding issues:

• Two outstanding Jiras to be tested & released to live - go-live completed 30/07/15 & confirmed ok
• Speak to Anne re low Apps Management involvement with project - was this an issue? - action for Dawn Nicholls post- project