Monday, October 18, 2010

How many bugs have you squashed?

Shaughn78 (go play his Risen Hero if you haven't already) completed his playtest of Path of Evil. A surprising amount of things went wrong somewhere near the end. These were things I'd already tested. I guess I need better version control. As a result I've started another playthrough of my own. I've already racked up 30 or so bugs and things in the first module, from extra spaces in conversations, actions not be properly identified with colors during conversation,  to an infinite gold exploit. Fixing things is never done (Bliizard is still patching Starcraft 1 and Diablo 2 aren't they?), and the larger the mod the more testing required. In fact PoE has been mostly testing for the past 6-8 months. How many bugs do you think you've squashed in your own work?


  1. Hi Kamal,

    I think you may have found the reason for some of my own delay in module creation. I tend to sort out bugs as I go along - almost testing every aspect of the part I have coded and how it may affect other areas as I go.

    Furthermore, sometimes I notice something else as I do a play test, which in turn can lead to discovering another unrelated bug.

    So, the bottom line is that it is quite hard for me to answer your question specifically, because I could be said to still be coding, even if it is still correcting the initial code. However, if I was to define a bug as something I thought was definitely working and had to go out of my way to repair, then I would said that I can recall only one specific bug to date: the "skin" problem bug. I thought this was working for a long time (and thought I'd fixed it once), but in the end I had to rewrite what I had already done.

    There have been lots of "working bugs" such as transition area problems and plot items in a multi-player environment, but I resolved those as time went by.

    Now, of course, when this is finally released, then I suppose my real bug list will begin ... but I hope it will be relatively small, or most definitely minor issues only.


  2. I have found the most frusting bugs are the ones you create fixing or just cleaning something else up. Some minor thing is now fixed, players probably would have never noticed it, and now 1 or 2 major bugs pop up.

    What I found too with play testing is the creator isn't the best tester. I tested Risen Hero throughout creation but I knew the "right way" of doing things. It is hard to come up with all the what ifs that other people will do and try.

  3. What Shaughn says. External playtests are incredibly useful for those reasons. For instance he realized that I needed to have unlimited ammo from my merchants since he was constantly running out. I never noticed that because I almost always melee.

    Some things though you do have to check yourself. Sometimes you'll want to trigger something, a playtester will never know if it doesn't fire, so they can't report it not firing as a bug.

    Also need to do an internal playtest with an actual party and not an overpowered testing party. You catch balance bugs that way. My endgame was far out of whack balance wise because I had only tested it with a testing party, not an actual party.

    Consistency checking:
    Most of the things I'm reporting to myself as bugs now as "consistency" things. For instance I indicate an action during conversation via color and brackets. I'm finding the places I didn't, or didn't switch the color back to the regular conversation color. Another thing is I try to put the "yes" options first in conversation and then "maybe" and finally "no". And I give xp for unlocking locks and disarming traps. So anything that's locked or trapped has to be checked.