I have a rule I thought I should share about implementing new features in Super Space Galaxy. It’s useful for any creative project and is an excellent guard against any last-minute mistakes. The rule is this:
Your first ‘final’ attempt at something creative is almost never truly final… even if you take this rule into account!
In my own file-keeping, I call this first ‘final’ draft ‘RC1’. ‘RC’ stands for Release Candidate, a term I picked up from working as a Tester in the Games industry. In those days, builds of our games we were hoping to release were labelled as RC1, RC2, etc. as new builds that were possible Release Candidates were tested. I was told that RC1 almost never got released and it tended to be RC2 onwards that had a real hope of being the final build.
This rule seems to apply to parts of games as well as whole ones. Normally when I’m implementing a new feature in Super Space Galaxy, I know that chances are it’s not really working, even when I can’t find any fault with it yet. Even if I don’t know what the bugs are, knowing this rule, I can be confident that they will appear.
So why is it so difficult to make an RC1 that works? Surely, knowing this rule, you can just compensate for it and check your first ‘final’ draft more carefully, right?
I think part of the reason this keeps happening is because you have a different perspective while you’re working on something. As you edit a video or program a feature, you’re likely to be looking at small details, trying to make the big picture you have in mind happen. As soon as the feature seems ‘done’, you must suddenly look at that big picture. Changing perspective like this takes time. In my book, publishing an RC1, even one you genuinely can’t find any problem with, would be a reckless move. You’re better off sleeping on it, showing it to others, or just playing around with it for a while to let the problems present themselves.
Secretly, many of the things I show plausible-looking GIFs of in blog posts don’t fully work until they’ve been given another session of work. Often this second session can be something like a second ninety percent on top of the work I already did. I’m not deliberately trying to pass off unfinished work as finished, but it is something I know is likely to happen anyway. Recently I overhauled the TikTok camera even though I already published the blog post about its completion. Next time you see a GIF of a feature I’ve newly-implemented, take a closer look; you may be able to find a small bug even there.
I think this ‘RC1’ rule is so hard to avoid because creative projects tend to involve doing something new. In cases like this you must expect the unexpected. If you’ve just finished a seemingly-final round of work on something new, let it sit as an ‘RC1’ for a while. You’ll probably be glad you did.
Thanks for reading,
Kenneth Dunlop
Kenneth Dunlop's Game Dev Blog 