A Different Development Approach for Version 2
Poker Copilot version 2 is well underway. I’m experimenting with an embedded SQL database, to get rid of the 150,000 hand limit. I haven’t yet made it fast enough for my likings, so I’m going through the standard list of database optimisations. Including using a database warehouse-like star schema.
Now that I’ve made the eyes of 95% of my readers glaze over, here’s how I’m changing my development approach. I developed version 1 via many small iterations. This worked because I started with zero users and I had no firm idea of what features people wanted in a mac poker tracker.
The first cut of Poker Copilot only worked with Full Tilt Poker ring games. Even before I released version 1.0 I had people asking me for Poker Stars support. So I added that, and re-released. Then came the demand for tournament support. Add, re-release. People were clamouring for a heads up display. Add, re-release. Bankroll chart. Add, re-release. And so it went.
As the number of Poker Copilot users grew, this approach hit problems. The quick turn-around time meant I wasn’t doing adequate testing. Adding new features sometimes broke things for existing users. Every time this happened, I felt deeply embarrassed, and braced myself for a backlash. To the credit of my users, the backlash never came. Instead I encountered patience, support, and a never-ending stream of feature requests!
For version 2 I’m using a more rigorous and traditional approach. I’ve made a list of intended improvements and prioritised them. Now I’m doing the development. Next comes testing and debugging. Eventually I’ll have a month-long beta release, available for the eager and brave. While this is underway I’ll get the documentation written and the website updated.
It will take longer and will be somewhat less transparent. But it will spare users from being an unofficial test team.
There will be an early access program with which you can try out the new features before they are fully tested. This still allows me to get some feedback, but without disrupting the great number of users who just want Poker Copilot to work. And keep working.