Its the start of a new week up here in the thin air, and as I sit here watching the sun come up over Legault Mountain with my cup of tea, I plot and scheme to try and find a way to spend more time here and less time in Denver.
Out my window Kalira's filly yawns and stretches, very cat-like for a thousand pound critter, and patiently waits for her breakfast while Marshal naps on the back of my chair as I type.
It is such a peaceful place with the rabbits, foxes, hawks, deer, rocks and pine trees – I just wish I had more time to enjoy it all.
Today at work I get to attempt the impossible yet again. this time I'm building (out of some tin cans and string) a database that is faster than the one the State had built (over four years at something a bit north of what I will make take-home over the next ten years) for their 'electronic poll book'.
See, I have to test this thing, which normally would be very easy… But the government doesn't do anything 'easy', and this is a stellar example.
As we all know, each person gets one vote, and the poll book is designed to ensure this happens by containing a database of every registered voter in Colorado – which is quite a few – and checking them off as they vote. Where the problem comes in is that to test it, I can only use each first/last name one time and someone in their infinite wisdom decided that having search functionality was bad, so it has to be an exact first/last name.
So, I have to build a separate database that holds enough of the voter data so my test engine can get the data it needs over a 4 hour test, but this database also has to be able to support better speeds than the database I'm testing against so that I can find where it fails.
Yep. Just another day in my life.
As I sit here and think about it, I realize that every year the things I've had to test get harder and harder… Five years ago when I started doing this at work it was all simple web site testing and I got weeks to do the test, then the web sites started being corporate and requiring things like SSL encryption (a pain in the ass), SAP (a major pain in the ass), and lots of general hoop-jumping in 5 days or less.
Then came the application tests which require the reverse-engineering of client/server communications protocols and data and the creation of completely hand-coded test harnesses – and of course they want this done in 2-3 days.
Now I'm doing government stuff which has all of the pain in the ass of corporate testing *and* all of the custom test harness crap, but also throws in a scale that is an order of magnitude bigger and contractors…
I'm giving contractors their own special section in this rant because of all the things I have to deal with, they are by far the worst. Everything is highly secretive so they won't ever tell me anything about what I'm testing, they spend all of their time figuring out ways to slip something past me so that the test is a guaranteed success (milestone payments are on the line usually), add days or weeks to things that should take hours, and are just generally a large pain in my ass.
So, yeah, this is sort of a primer on why I get cranky when I talk about my job.
Speaking of job, I have to head down the mountain to be the miracle worker. Have a great day out there folks.