Original thread here:
http://wiki.pelicanparts.com/showthread.php?t=658230&highlight=Salt
(For as long as this link works.)
Going back to November, ADL asked me if I could get a huge development effort done to be released in January. Considering that it would take me 400 hours (or roughly 4 months of doing almost nothing but this), I told him no, the timeline was not feasible. We agreed to instead work towards a June release, which was feasible.
I got all of the development work and testing done ahead of schedule, including some major rework. Our testing team was right on schedule. During this time, ADL's protege was my contact and responsible for ensuring I delivered the correct product. Starting two months ago, I began asking the protege to begin his user acceptance testing, as it is due May 4th. Every time I met with him (about twice a week), I asked him to provide me with his test cases and what data he needed to test with, so I could start getting things ready for his testing.
The protege is a huge procrastinator, by his own admission. He has always gotten away with starting his testing very late because he has either been testing small changes or the release has been postponed. No such luck this time. He elected to start drafting his test cases on Tuesday, and I worked with him yesterday to identify his data needs. We ran the first test case this morning. Low and behold, we discovered that he had some additional requirements that he never communicated. At this point, his only two choices are to accept the product as-is, or to pull it from the release. There is no time for rework and retesting.
Upon further research, I discovered that releasing it as-is is not an option. When I left work, the protege was meeting with ADL to explain to him that because of his procrastination (which I have documented, so I have no fear if he tries to blame me), this product will have to be pulled.
This product was a six-month stop-gap until a more permanent solution could be developed and deployed at the beginning of next year (an effort I am not involved in). The thing I haven't told the protege yet is that the next release this product is eligible for goes live in the middle of November. It would only be live for a month and a half. It no longer has a positive ROI.
The bottom line is the protege's procrastination wasted four months of my time (and less of other people's time to varying degrees) and lost us the opportunity to save millions of dollars.
It looks like I will have to start the process of backing-out the change on Monday. Pulling the code is easy, but undoing the database changes will take some time to get ready, and I'm sure I'll have to attend dozens of meetings where I explain why this had to be pulled to various Change Managers and Release Managers.