Tuesday, November 24, 2009

November 24

Tuesday, November 24, 2009
Last month, we rolled out two very large projects at work. I'm not sure what I was thinking setting the same go-live date for both projects, but that's what we did. Of course I wanted the roll-outs to be smooth and for everyone to love the changes and for the projects to be successful. Overall, I do think we had successful roll-outs of both projects, but they weren't perfect, that's for sure.

One of the projects we rolled out was waitlisting - allowing a student to get on a waitlist if a class is full, and then moving them into the class as seats become available (either because a student in the class drops or the department increases the capacity of the class). Pretty cool if you ask me... much better than continually checking back to see if the class has opened up. Throughout the last month since we went live with waitlisting, over 1650 students have been moved off a waitlist and into a class, and there's still almost two months until school begins. I know we'll see a lot more movement.

Anyway, I've been stressing out over the last couple weeks because we discovered a flaw in the process that we somehow missed in our testing. That's the hard thing about software testing - trying to think of each possible scenario. We spent hours coming up with a manual work-around for this semester, and I begged Jim (our development manager) to get a project scheduled in January so we could address the issue prior to Summer registration (which begins in February). We had a meeting yesterday with Jim and one of our very favorite developers ever, Greg. Greg used to be a dedicated Student Records developer prior to a reorg that happened amongst the developers in January, so we were excited to have him assigned to help with our waitlisting problem. He had done some research and offered several suggestions. We agreed to try one of the solutions, which both solved our problem and wasn't a very invasive code change. I think Greg had the change made in a test environment in like 5 minutes, and it looks like it will work for us! We will do some more testing over the next couple days, but I'm hoping the fix will be in our production environment as early as next week. Today I am thankful for Greg - for his knowledge and his expertise and for being a super star developer. He made it so I will sleep better tonight, and tomorrow night, and the next night... :-)

0 comments: