Akasun
   |
Posted - 2009.12.04 23:58:00 -
[1]
 Originally by: Sandar Orontur
That being sad, I absolutely don't see why CCP rolls out an expansion like this when there are still such problems. Of course you will always find problems after a release which you didn't before, that's just natural. But this particular bug could well have shown before IF(!) proper testing was done.
fyi, these problems were known from sisi testing despite the limited time people had with the client. the client was initially unuseable due to the severe graphics slowdowns it had. when that got addressed, i know i encountered crashes that ended up corrupting the preferences folder. i did report these occurrences, and i know from discussions on the osx channel that others were also seeing similar problems and reporting them.
i fundamentally don't think there was enough time to analyze the known problems and address them given the tight release schedule. because of that, i believe there were requests in this channel asking that dominion be delayed due to the known instability of the mac client and the lack of adequate investigation and fixes. as previously stated, we were all hoping that ccp/transgaming had already identified and fixed the problems in a non-deployed release tree. that turned out not to be the case ...
what i would really like is a devblog on how ccp is going to improve quality going forward. for all this vaunted talk about agile development and scrums, i have seen two major releases with severe problems after deployment. for some of these problems, hints of these problems were seen on the test server. mac users always end up having extremely limited testing time, and the hiccups that show up indicate that internal testing isn't catching these problems. while this is a mac channel, i know that other areas also continue to have quality problems. for example, we continue to see periodic database crashes and bizarre localization-dependent bugs.
ccp, as you've seen from the many posts, it's a common expectation that a new content patch will be buggy and unstable. i can't believe that you're happy with this state of affairs. so, how are you planning on on changing this perception? for example, we see all these calls for stress and feature testing on sisi. those are great. however, these tests were done mostly without mac users due to the lateness of client availability. in addition, they rarely get you the 20-40k active users that are on tranquility. my expectation would be that you're adding your own load/automated activity generators to help give you a better idea of how the transition from sisi to tranq will end up being. is that the case?
for mac testing, an agile model would suggest always having a mac client available in some form (with missing features if necessary) vs. the current late availability model you have now (that ends up having missing features anyway). in general, what kind of manual/automated/stress testing are you implementing to improve test coverage? are there gaps that would benefit from focused user testing? for example, getting an idea of infrastructure changes (as opposed to just the new feature stuff) would allow users to help in those areas as well. would a more formal beta testing program with checklists and feedback help? as it is now, i have this suspicion that people think their bug reports disappear into a black hole. that doesn't inspire confidence.
i'd like to see a time when people don't anticipate a new patch by creating a thread on what long-term skill to queue up. please let us know what you're doing to get to that goal.
|