<!--quoteo(post=218756:date=Oct 24 2007, 11:27 AM:name=Pieter Boelen)--><div class='quotetop'>QUOTE(Pieter Boelen @ Oct 24 2007, 11:27 AM) [snapback]218756[/snapback]</div><div class='quotemain'><!--quotec-->Exactly. We need to actually establish an "official" team of people who want to decicate their time and efforts to squashing the current mayhem. Not the random testing and bug reports that we have currently, but an actual dedicated team and a focused bug-catching effort. In theory all we need is one modder and a couple of testers. It'll take a while to get it all ironed out, considering it's many mods and many people don't have much time to spare, but I think it needs to be done. Otherwise we keep messing around. <img src="style_emoticons/<#EMO_DIR#>/modding.gif" style="vertical-align:middle" emoid="

" border="0" alt="modding.gif" /><!--QuoteEnd--></div><!--QuoteEEnd-->
OK, I'm game and I have "real world" experience in a software testing environment, both from a development perspective (code) and from a Alpha/Beta/User Acceptance testing perspective. In addition to testing, I have some limited knowledge of the code and can help with bug fixes.
It will need to be a systematic effort, including documenting all the efforts. This means that we'll need to start with a stable "working" version, add one mod/change at a time, generate test conditions for that change, assign those test conditions to testers, test the change, report any bugs, fix the bugs, and then go on to the next mod/change.
This will require that we have individual mod changes, and not wholesale changes in one test round. Do we have this kind of info available, i.e., mods separated into discreet code units OR do we simply have the various alpha versions with many mods/changes in them?
Here's a generic plan:
- Choose the testing team
- Choose the stable version to start testing
- Identify discreet mods to be tested (and not simply "this is stuff from Build 14 Alpha 1", etc.)
- Create a ranked list of all the mods to be tested (according to complexity and/or some other measurement)
- Generate a tentative schedule/project plan for testing based upon the ranking(s). Should be posted somewhere that only the test team has access. This should be updated and kept as the "master" project plan.
- Update the bug tracker and assign a team leader
Here's a generic process for the actual testing:
- Generate test conditions using a standard template, then adapt (choose appropriate conditions) for each mod (e.g., Interface, Land, Sea, Battle, Land Fight, New Quest, New Characters, New Ships, New Items, etc.). This will probably be the most time consuming part of the process, as we will have to identify lots of conditions (toggled mods off/on and various combos)
- Assign test conditions to testers. It would probably be a good idea to have certain testers always use certain mod toggles to make things go smoother. E.g., someone always tests with "expert" toggles (AMMOMOD on, WEAPOSNMOD on, FOOD on, etc), some others test with "beginner" settings and some others test with "mixed" settings.
- Team leader creates and uploads new files for testing, identifying the exact change made
- Perform tests and report bugs from test conditions
- Fix bugs and re-test conditions until bugs are fixed
- Repeat this process for the next test cycle (next mod/change) with the "new" stable version
Let me know when we are ready to start, as I'm at a slow time for my consulting business until the end of the year and have some free time over the next couple of months.
kevin