testing stuff

[Anduin]

I remember when stuff used to always be tested using pfiles on a dif port. I think doing this again would be a good idea. I know it takes a bit longer but re-implementing stuff (like when rift went crazy) takes long too. The AC problems etc. are another example.

[Anonymous Mammal]

As far as I can remember, Tynian has always used a test port to test new code, but only when testers were available to do so.

As for the rift thing, Tynian said "a bug slipped in." Or something to that effect.

[Solaron]

I remember when stuff used to always be tested using pfiles on a dif port. I think doing this again would be a good idea. I know it takes a bit longer but re-implementing stuff (like when rift went crazy) takes long too. The AC problems etc. are another example.


Tynian tested it to make sure that it was working as he intended it. He might not have envisioned the issues you're encountering, but don't take that as a "I never tested this code.".

Maybe future testing should involve a large test base? Some from each class to see how the different playstyles handle the new code? Not sure.

[Katrana]

I test code for a living. Creating test plans, etc all that fun junk, I won't get into details. It's what I do day in and day out. My point? Bugs, defects, always, always get through. You can't envision every scenario, and in this case if you don't play the characters daily or think like they think, you'll miss something it's inevitable.

[Cordir, Lady of Fate]

Maybe if players put (Tester!) in their title, Tyn would know who was available at any particular time to test? I know I've been online when Tyn has asked for volunteers to test, and he didn't get a single taker - though that is admittedly rare...

Most of my mortals aren't of a range that's any use to him for testing purposes, sadly. :( Dang. I should quit playin my imm and get leveling!! ::whistle::

[Tynian]

Eh, respectfully, I think you're making too big a deal out of this.

It's pretty rare that we suffer a serious problem. They are almost always caught and fixed in testing.

In the end, I have to balance the need and level of testing against the time it takes and the potential risk of something going terribly wrong. All in all, I think I've struck a good balance...