Closure Review

Project Summary:

Contributors

Project Review

The SSG002 and SSG003 projects had to migrate the current CAM applications

• UG Prospectus
• PG prospectus
• A-Z
• Maps  

from aging Praline server to new supported IS infrastructure.  

SSG002 was the Migration project, with a core funding of 217 days. SSG003 was the Enhancements project, for UG and PG Prospectus, once they had been migrated, sponsor Funded.

From the outset, it was obvious that the two projects were entangled and a clear separation not so easy. At one point the PM proposed having just one project but this was rejected as Director's Office prefered to keep them apart for easier control of Finance.

One main issue was that due to security issues on Praline, the milestones were brought forward so that A-Z and Maps migration were done much earlier than planned (by several months) so Praline could be decommisioned earlier. However, when we found ourselves with  a budget issue, it meant the lower-priority work A-Z and Maps had been done ahead of the higher priority PG Prospectus Enhancements, so it was not possible to descope the A-Z and Maps which had been the plan should a budget issue arose, and instead more funding had to be agreed. CAM's feedback on this is shown below.

As the project closes, the migrations have taken place, new admin modules created, and the enhancements made. We also have a new LAMP architecture as a template that can be used for future PHP application development.

Objectives (for SSG002 and SSG003)

Deliverables

Scope

Originally as per deliverables above.

However, SSG002 also brought into scope : the setup of a new Linux infrastructure supporting PHP to reduce the coding effort. That is, to keep code as PHP rather than rewrite as CF9. This saved many days effort.

the creation/update of Admin Modules for the four applications to allow CAM to do all the work they currently have to request Support assistance for, or by accessing the live Praline server and databases.

Schedule

The UG and PG Prospectus were migrated and enhanced as per original plan.

The A-Z and Maps were migrated earlier by 2-3 months, but this caused issues with funding as outlined below (Feedback from CAM).

See https://www.projects.ed.ac.uk/project/ssg002/plans/tasklist-project

Budget

For clarification to sponsor and to WIS.

Total estimate for the two projects (SSG002 and SSG003) comes to 350 days.

Of this, 217 days are funded by IS.

Of the remaining 133 days, 25 days are being funded by CAM's Support days as agreed by Jim McGeorge and David Smyth.

Therefore, that leaves 108 days to be paid for.

45 days so far billed to CAM, so CAM to be billed a further 55 days. The other 8 days will be paid by Jim McGeorge as agreed with David Smyth).

CAM do not pay for the new LAMP infrastrtcure.

Analysis of Resource Usage:

Staff Usage Estimate: 317 days

Staff Usage Actual: 350 days

Staff Usage Variance: 10%

Other Resource Estimate: days

Other Resource Actual: days

Other Resource Variance: 0%

Explanation for variance:

The SSG002 project was where there was budget overspend. Some of the build tasks took longer than had planned, BRD and UAT also.

See a comparison of SSG002 tasks estimates and actuals at

https://www.projects.ed.ac.uk/system/files/xls/Current%20vs%20Estimate-2013-03-22-Detailed_0.xlsx

SSG003 came in around 13 days under original brief estimate of 100 days, due in part to some of the build being anticipated in SSG002. The savings came from adopting a 'lite' approach to the UAT testing as CAM had fed back from SSG002.

Key Learning Points:

CAM Feedback

Priorities: CAM was asked to prioritise the project components but IS later advocated that we change the order so that the projects could be combined. CAM were unaware that the impact of this re-ordering could jeopardise our priority work. Had we gone with our own priorities, we would have been in a position of finding alternative solutions for A-Z / Maps, which is preferable to the position of finding funding to complete work on PG. Estimates: CAM noted that many of the estimates for seemingly straight-forward elements of the project work were inaccurate and were are unable to identify why these overran so significantly (eg Maps/A-Z UAT rework). Time recording records don't seem to clarify this.

Process: CAM is happy with build and the documentation for build, but felt that other elements of the process were inefficient and may have contributed to the overrun. e.g. BRD/UAT documentation process is very lengthy and found to be largely unnecessary. The SDS was often delivered before the BRD had been signed off, or in tandem with BRD, possibly due to rescheduling of project and developer resource, meaning that work for the SDS duplicated work that was already being done by CAM for the BRD (eg mockups) and/or did not take updates to the BRD in to account. Also concerns over the amount of time spent on the UAT documentation that could have been scaled back.  New environment: CAM queried where time spent on developing new environments should be solely paid for from our budget if future projects will be using this infrastructure. Concerns over this work, particularly where allocated as 'unexpected/unplanned'. Support budget: CAM was surprised to learn that the c. 35 remaining support days have already been used by SASG and that support budget is shared across the whole group, rather than allocated to each department, as implied by the support update. This was never communicated to us and we had thought these days were available to us to use as required.

PM Feedback

The project has delivered on the original objectives and deliverables.

All four applications were migrated within the timescales.

All four applications were given new or enhanced Admin modules allowing CAM to do almost all their support themselves, in regard to file changes and other actions that previously required either support calls or direct access to the servers and databases by CAM.

A new architecture was set up, LAMP, that saved CAM sufficient moneys to allow the project to deliver all the migration, the Admn modules and all the enhancements.

The new LAMP architecture provides Apps with new development options for PHP and MySQL for future projects.

The CAM applications and databases - high-profile user-facing - are now on secure, supported, resilient and backed-up infrastructure. The old i/f was end-of-life and (demostrably) unsecure.

The UG and PG Prospectus Enhancements were delivered, also to quality and timeliness.

The final effort was 10% over the original Brief estimates.

The developers worked very well with CAM, and there was a good relationship established. Some of the feedback from CAM will help future planning for similar proejcts.

The project had to cope with taking on three new (as in, new to Apps) developers at short notice and simultaneously. This was advantageous but also could have been problematic. As it turned out, the work went well and Resource and Project Manager worked well together to move the project forward over the Q1 period, with quality support from the Senior Developer throughout.

PM draws attention to the questionnaires completed by Apps and by CAM for an overview too.

https://www.survey.bris.ac.uk/?surveyid=154235&op=results

or PDF

https://www.survey.bris.ac.uk/?surveyid=154239&op=results

or PDF

Outstanding issues:

none