Our staff is moving on to the final testing phase for the new accounting software. There still are bugs and gaps, but it keeps getting harder to find them.
Software testing is boring. You have to try all possible paths that might cause problems. That soon gets tedious. It’s not easy to pay close attention to small details for hours on end.
This isn’t just a problem when building software. Construction (and most everything else) also suffers from it. You take something apart, and put it back together very, very carefully, and damned if there isn’t a screw or a washer left over. It’s even worse for abstract stuff like estimating or accounting. Too easy to space out and make mistakes.
That’s why computers are so helpful for such tasks. They don’t make math errors, and they don’t get distracted or forget. Of course, computer software isn’t magically perfect. It’s a long, slow slog to build it carefully, then debug and test, test, test it.
Programmers aren’t good at testing their own code. They’ve already looked at the output too many times during the construction process. For many years, brand new TurtleSoft employees were testers, instead. It was a way for them to learn the app and move up to tech support. Newbies also did random things and made mistakes, just like users. Sadly, they only had a few productive weeks of bug-crunching before boredom kicked in.
Our tech support needs have diminished. We won’t hire testers for TurtleSoft Pro. So what’s the solution?
For two decades, our staff traveled all across North America to give hands-on classes: first for our Excel-based accounting/estimating software, then for Goldenseal. I taught 90% of them. Every class turned up at least one bug. It was embarrassing to have them pop up in front of 20 people. On the other hand, at least we found the problems, and fixed them soon after. Discovery is always the hardest part. The class format was like having fresh eyes: it worked even for bugs in code written by the teacher.
TurtleSoft won’t have hands-on classes for the new accounting software. There’s a better option now: YouTube videos. No need for anyone to travel. Users can access them for free, any time. Hit pause if you need it. Watch twice. All that’s missing is coffee and donuts at the break.
Video production also solves the testing problem. It’s a chance to kill two bugs with one stone.
Next week we will start making short instruction videos, for every possible thing you might want to do in the new accounting software. That means running every part of the app a few times, and figuring how best to present it. Along the way, it’s a chance to find and fix any lurking bugs or missing features.
Most likely there will be a few hundred videos. It will take a week or two to get the routine down. Then we probably can make a few a day, upload them to the TurtleSoft channel on YouTube, and link from the website. Tech support will be up and running even before the software is out.
Accounting and estimating videos will also be good marketing. Magazine ads worked best for us in the ’90s, then SEO in the ’00s. This decade, social media will get the word out. YouTube is a good start. The main down side is that competitors will run ads on our channel. But maybe we can get paid for each click.
Dennis Kolva
Programming Director
TurtleSoft.com