The world is full of commercial software developers who breathe a huge sigh of relief when they ship a new release of their project. They start thinking about vacations. I used to be able to relate to that.

Before starting WGM, I used to be among the purist crowd who considered the ability to just “fire and forget” in development a major job perk (if not a requirement). I naively felt that projects started at their conceptual best, when you built from a blank slate, and I didn’t enjoy watching them degrade and get “boxed-in” through the endless concessions made during maintenance and marketing. Ideas were pure and reality was evil. Users was somehow a dirty word.

My reality at WGM, thankfully, has been much more practical; and my collaborative, iterative development eyes are wide open. There’s a powerful, unavoidably-magnetic sense of pride and ownership here that makes it impossible for us to just “fire and forget” anything.

I often find myself entranced by our hosted helpdesk servers. They’re running hundreds of copies of Cerb4 for customers who exist in this — as my past self would say — “messy real-world”. The same past self would gawk at the liability of hosting all this data, and whimper, “I just want to write code!”. After all, every design flaw and inefficiency is embarrassingly magnified on this scale… and yet, I love it!

It’s way too easy to forget how many inefficiencies you’re forgiven when you’re working on a scale of one copy on one machine. I’ve really found my way as a developer by honestly looking at what hasn’t worked, and still trying new things knowing they could very well dead-end, because it’s always a much stronger lesson than simply googling “the best way to _____” and blindly assuming someone else’s opinions or conclusions. Obviously it’s practical to do research, but that’s a far cry from letting something you “read somewhere” become your immutable opinion.

So, do yourself a favor and find a way to be responsible for hosting at least 100 copies of your own software for real users. Naturally, you can charge whatever you want, as long as you still end up with a population of at least 100. Take care to not oversell your hardware, because I’m assuming you do actually want to provide a useful service to people; and your brand is going to be in the gutter if things are completely unusable.

You’ll be getting paid to learn how to write better software. You’ll get the mindshare of being the application people choose because it doesn’t bog down or die when you have a mere 500,000 records in the database.

(If there’s any interest on this topic, I’d love to do a couple more posts about the things I do to benchmark and monitor my applications in our high-use hosting environments.)