May 11, 2008

Output, Part 5

This is actually going to be the last part of this. I know I said six parts, but the sixth is really just about software development in general and not about my current situation in particular.

I find that there are certain things beyond my control that still manage to frustrate me. Most of the time I'm able to adopt a certain stance about such things that lets me live with them, but some I just can't.

The biggest one of these is that, at work, we're expected to shepard projects through the entire software development life cycle. Sure sometimes it gets broken up because the person who got the requirements, designed, and estimated the thing has higher priority tasks, but generally speaking we have "teams" with very few team projects - even those team projects we do have rarely require high levels of interaction between more than two people. Ideally, we'd have true business analysts who'd get requirements, designers who do the design, and programmers who work with the guts of the applications. Honestly, I think I could deal with being the last two, possibly even any two of the three, but being all three at once puts a bit more of a strain on me than I'm entirely comfortable with.

Also: we have no standards for automated testing. Currently, most testing is done by hand. Admittedly, this is probably largely due to barely having set foot into the modern era of programming. The company is still a mostly COBOL shop with some ASP (not .NET) and Javascript. Only in the last 3 or 4 years did they start dabbling in VB.NET, and any sort of entrenched business position is going to take quite a while to ramp up to more modern practices. It will take some of our people's retirements before it can ever fully happen.

Anyway, I hope I can find it in myself to accept my current position for what it is, whatever that may wind up being to me.


Brian Sullivan said...

The reluctance to fully embrace modern programming practices (even while attempting to adopt a modern platform) was one of the few things that irked me about DTC. More often than not, you'd see people using an object-oriented language to write procedural code. I feel your pain on that one, man.