Completion Report
Project Summary:
Objectives
The objectives in the following table were identified at the planning stage of the project. Some were moved to other timetabling projects as the project developed
| Objective | Achieved under TTU003 | Achieved under TTU005 | Achieved under TTU007 |
| 1. Deployment of a 2014/15 Timetabling database and Enterprise Applications | Yes | ||
| 2. Populate 2014/15 SDB with new course data from EUCLID following EUCLID course roll forward | Yes | ||
| 3. Deployment of 2014/15 Web Applications (OneLan replacement, Web Timetables, Web Room Booking and PADS XML [if required]) in Dev, Test and Live env | Yes | ||
| 4. Roll forward outgoing interfaces which use Reporting Database as datasource | Yes1 | ||
| 5. In August 2014, undertake switch over to 2014/15 Academic year SDB for web apps and interfaces | Yes | ||
| 6. Decommission Web Applications | Yes |
Note 1 - Not all of the planned outgoing interfaces were rolled forward as work on the XML Data Feed for Schools and the Timetabling Link to Office365 was suspended.
Deliverables
The following deliverables were identified at the planning stage of the project:
| Deliverable | Delivered under TTU003 | Delivered under TTU005 | Delivered under TTU007 |
| 1. A Live (King's Buildings) 2014/15 Scientia Database (SDB) with rolled forward course activity information and Authorisation Manager users and roles for 2014/15 course planning | Yes | ||
| 2. A Live 2014/15 SQL Server Reporting Database | Yes | ||
| 3. Backfill of 2014/15 SDBs into Live (Appleton Tower), Test, Dev and TRN environments | Yes | ||
| 4. Mechanism to extract course data from EUCLID and populate 2014/15 SDB with this data | Yes | ||
| 5. Population of 2014/15 SDB with course data to support curriculum planning | Yes | ||
| 6. Deployment of OneLan for 2014/15 | Yes | ||
| 7. Deployment of Web Timetables for 2014/15 | Yes | ||
| 8. Deployment of Web Room Booking for 2014/15 | Yes | ||
| 9. Deployment of PADS XML for 2014/15 | Yes | ||
| 10. Define, document and test switchover approach for feed to Learn VLE | Yes | ||
| 11. Define, document and test switchover approach for Timetabling SOA Service | No1 | ||
| 12. Define, document and test switchover approach for Bookable Rooms web service | Yes | ||
| 13. Define, document and test switchover approach for Office365 synchronisation tool | No1 | ||
| 14. Interfaces repointed to 2014/15 database instances | Yes | ||
| 15. Web Apps repointed to 2014/15 images | Yes | ||
| 16. Web Applications for previous year are decommissioned | Yes |
Note 1 - Numbers 11 and 13 became redundant after the respective projects were suspended.
Analysis of Resource Usage:
Staff Usage Estimate: 68 days
Staff Usage Actual: 81 days
Staff Usage Variance: 19%
Other Resource Estimate: 0 days
Other Resource Actual: 0 days
Other Resource Variance: 0%
Explanation for variance:
Comment on effort used
The final figure for effort used shows an increase from the original budget of an additional 11 days. Two issues advising of the increased effort were raised in the latter stages of the project - the first being an increase of 3 days in July to cover the extended duration of the project, and the second being an extra 6 days in August to cover the final testing needed for the LEARN interface. The additional four days have been used in the final iterations of the testing of the LEARN feed and its deployment into LIVE. This time is split between Development Services; Development Technology, Project Services and Application Management in their work to get the feed fully tested.
It should also be pointed out that the original estimate was not formally adjusted after some of the deliverables were moved to TTU005 and TTU007 (a figure of approximately 5 days) and that the tasks that remained in TTU003 still used this effort.
Revised milestones
The milestones for this project were revised around the time of the Web Apps Roll Forward work being delayed in TEST from late January to early March. This revision was attributable to a combination of unfamiliarity with the packages being deployed; unknown issues arising during the installation; dependency on one colleague for assistance at particular times, and resource conflicts with STU235 & 236. This hiatus was similarly reflected in the outstanding milestones, with the LIVE Web Apps Roll Forward also being about six weeks later than originally planned.
The final milestone - covering the documentation and testing of outgoing interfaces - was delayed by several more weeks because, initially, it was difficult to get what is a small piece of work fitted in around other work. This was finally completed in July, but the successful testing of it proved to be difficult, largely because of insufficient test data, and 'add-ons' making the testing more complex than it should have been.
The milestones for the switchover tasks, originally earmarked for August under this project, were removed and transferred to the follow-on project due to start in August.
Key Learning Points:
It may be argued that this project has had too many strands to it, and has suffered from starting part way through the year, with its original finish due to take place early in the next year 14-15. It has also suffered from crossing over with the Upgrade project (TTU002) and the completion report for that project makes the suggestion that the two should be kept distinct so that there is not 'competition' for the different environments, and also staff resources, especially when there is a need for the involvement of more experienced colleagues. Another strand of work that added to the problems experienced during work on the LEARN feed was the upgrade to the LEARN application. This took place between development & testing and the eventual deployment on this project, and was another factor that had to be borne in minf during the rework thrown up after retesting.
The final delivery - the roll forward of the LEARN feed - certainly became over-complicated with the script to roll-forward to 14-15 having additional functionality (for groups, logging and caching) being included in the eventual testing and deployment. Lumping these together is understandable, as the constituent elements were all 'small' parts of the LEARN feed but putting them together into one deployment increases the risk of the package failing as a whole, and thus putting parts that might work successfully in their own right at risk of not going into LIVE when they are needed by business users. Project work should also not include deploying changes that have been agreed and put together outside the project.
One important lesson learned was that with the implementation plan being updated and revised rapidly, one vital element of the original plan was omitted from a later version. This is detailed in issue 16. This was noticed in time but almost led to a significant exclusion from the deployment. This incident highlights the importance of version control for such documentation, and the need to monitor amendments when more than one member of staff is modifying them.
With regards to the timing of the Roll Forward project, discussions between stakeholders on this project have established that future projects of this type should run from August onwards, so that that year's project picks up the relevant tasks for the current year. In this way, it might also be possible to develop a set of procedures for Production Management to incorporate annual tasks into support work instead of project tasks.
It has also been suggested that, in order to help support this approach, some time may be spent on pulling together all the different elements of Timetabling Roll Forward and Switchover tasks into a coherent set of procedural documentation that is dictated by a calendar of when each set of tasks should be carried out.
Outstanding issues:
There are no outstanding issues, but it should be remembered that the switchover tasks for 14-15 need to be carried out as early as possible under the new project.