Jump to content

Thoughts on why the bugs in the latest update weren't caught


mleii

Recommended Posts

This post was going to go into the patch notes thread, but decided I didn't want to derail that thread any more than it currently is and this at least gets some of the this in a better forum and will have less likely of a chance to get deleted.

This is all purely my opinion and I have no insight into the actual QA and development that goes on at WildCard.  So stop here if you don't want to read me ramble on about this and that.  LOL

It is based on a lot of experience in the software industry, as a developer, tester and jack of all trades.  A great many of those 25 years were spent at Microsoft, mostly as a software tester there on various things one of which included a year working on a product related to Xbox Live services.  And to be clear I don't work there anymore.  I won't pretend that I know all, as I still learn a lot every day, and I won't pretend that I know exactly what happened here, all purely my outside view knowing how things normally work.  I also realize I won't convince most people, and honestly I don't really care, which makes me wonder why I'm really even writing this.  So hopefully someone finds this interesting.  And perhaps it might get some sort of insight from someone at WildCard if they decide to respond and correct all my poor assumptions. :D

I am not defending them in any way shape or form, they really need to figure out a better system here as it impacts a lot of people when mistakes like this are made and gives them an even more black eye each time the same mistake is made.  I highly recommend watching and listening to the video at GDC.  I have a bit more understanding of what they do and why, but there were several items in the talk that I felt were very much off the mark or they missed key parts as the current reality didn't seem to match up with what they were saying was or is happening.  But I'll leave that topic for another thread and day.  LOL

RE Why did the bug Ark crash due to controller shutdown even see the light of day?
The likely scenario is that the QA team is testing on dev kits and not on real consoles and unfortunately those don't always have the same network and device behavior that real consoles do.  I know that when I was playing around with writing up games on the Xbox 360 that there was a method to play those indie games on a real console, as I did it, but I'm not sure that they can do the same with a full game.  I would venture to guess that there is a way to test full games on a real console, but given some of the bugs I've seen, like this one, I would venture to guess they either cannot or do not.  The other potential reason they missed it might be that they're using wired controllers on a real console to save on having to replace batteries constantly and didn't see this.  But given some of the other bugs this release I'm not so sure that was the case.  There may be parts of this we may not fully find out as parts of this may fall into an NDA (Non-Disclosure Agreement) and they cannot talk about it without breaking this contract.

RE Why has this bug never been fixed and has been there for N months?
As for other bugs existing for a long time and not being fixed, there are multiple reasons this can happen, all without someone being lazy.  If you are well and truly interested in the reasons why it's not that hard to find the information you seek, lots of dev talks by many developers besides WildCard have written on this topic, both for games, and not games.  

A bunch of potential reasons for why a bug might or might not be fixed off the top of my head.  Bug doesn't meet a certain bar, it has a workaround, only impacts a few users, cannot find a fix that doesn't break other things, have a fix that need a lot more testing, cannot find a solid scenario to reproduce the issue consistently so that they can fix it or be sure they've fixed all related scenarios, fixing it needs to come later when all elements that it could impact are in so they don't have to fix it again, the bug was actually fixed but a last minute change broke it and users were complaining about getting an update or other bugs were more important to get out a fix sooner and then fix it later so they shipped it anyway, someone trusted someone else that the bug was fixed but it was not, the bug was fixed in one branch of the code and they forgot to get it into the build that was released and I'm sure there are more that I can think of if I get more coffee'd up.

RE Why did this even pass Certification at Microsoft?
Great question, I don't have full insight into this, and we may not get full insight as now we're likely to fall into an area where developers and testers cannot disclose things due to an NDA but from what I have learned from various sources, the certification process does not include checks for these sorts of things.  As has been pointed out elsewhere before, they

RE  Why didn't you answer my question about software testing or development?
Likely I either do not know or didn't think to answer that question.  If I get a legit question where it seems you really want to know and aren't trolling I'll try my best at answering it if I can and if I have some time.

Also, I have thought about writing up a post on how to write up issues.  Specifically how to write them in a way that makes them much more likely to get addressed.  I've rambled about it in various threads, but have not written up something more fully fleshed out.  If there is interest in that let me know.

Finally if anyone has any corrections or additions please let me know, there is a lot about game development I don't know as I mainly have dealt with testing/development of Applications with UI, APIs, Application Frameworks, WebServices, Websites which are entirely different from testing games.  I know enough about game testing to know I wouldn't want to do it. :)  There's a lot more to it than most people think there is.  Writing test automation for the things I've tested are relatively easy, writing test automation for games is possible for some things, but most things require actual game play to test, which is why having alpha testers like us helps to find a lot more bugs than a large test team possibly could.  Test automation is relatively easy to scale out, scaling it out on game systems is much harder and in some cases not possible.  Getting real world testing by paid testers doesn't or cannot scale out much without having it in an alpha/beta test phase by a lot of users.

Link to comment
Share on other sites

pure hearsay at this point but this is what i've heard they're trying very hard to release asap so they're trying to jam in all the content they want so they could release as soon as april, if they can get to beta and start on optimization and bug fixes that gets them closer to release. and they might wipe servers upon release so anything that has happened till then is irrelevant

Link to comment
Share on other sites

14 minutes ago, tibs said:

pure hearsay at this point but this is what i've heard they're trying very hard to release asap so they're trying to jam in all the content they want so they could release as soon as april, if they can get to beta and start on optimization and bug fixes that gets them closer to release. and they might wipe servers upon release so anything that has happened till then is irrelevant

Yeah, they're going to have to jam in a ton to get done by April for sure.  My "earliest" best guess based upon past and things going right, which this patch didn't help, would be June.  And that's pushing it pretty hard, giving them some bake time, but they'd need to get more hands on deck to make that happen for sure.

Link to comment
Share on other sites

Just now, Facebunz said:

http://www.gdcvault.com/play/1024229/-ARK-Survival-Evolved-Lessons 

Here's a link to their GDC video, very informative. If you're short on time the last 20 minutes explains how content gets scored by Wildcard on whether or not to implement it. 

Yes, thank you!  I should have linked that.

Link to comment
Share on other sites

It feels like this is the same sort of way I would patch\update a database server if short on time. I'd do some testing but eventually just roll it out and fix the problems I can't anticipate after implementing it.  As far as totally breaking private servers, you'd think this would have been tested on an actual console and quickly realize password connection doesn't work anymore, and controller going off crashes it, feels like that is obvious, but as usual the main focus is on official servers and PvP related balancing and patching. (which has nothing to do with my happy world of private server PvE which many of us love very much)

 

Also why drop a 24gb patch on friday late in the day? so much for friday night games after a long week at work, let alone the whole weekend no Ark and nobody can patch anything or talk to microsoft about whatever certification issues they need resolved until the next week. (god forbid you don't have a proper very high speed internet connection, days to DL that kind of patch)

It would be nice to be in actual final release production so their is a bit more accountability then, yes we dropped the patch as the last guy turned off the lights at WildCard for the weekend but its in DEVELOPMENT, so we can do whatever we want! (knowing us suckers are hooked on this wonderful game!)

Link to comment
Share on other sites

Here are my two cents. First I want to say that this game is in Alpha Status. That means they are focused on content not bugs and issues. Once the game reaches full release content wise it will enter Beta Status. Of bug fix. Most games do not allow an outright release of a game in Alpha Status. Normally it is a closed circuit that few get to see. Normally people who have a long gaming history. The reason things are done in this way is to prevent double work loads. This is meaning that as new content comes out the coding and how the game is played changes, and some bugs will go away others my come back and some may show up. If it is not a  game ending bug, it will be noted and left for Beta. This is because the next content patch could repair the bug or mask it or many other things. So to prevent repairing a bug 20 times prior to the whole game coming out, it is best to way till all major content changes are released. Once the last sets of content come out ( the next 2-3 months) the focus will shift from content to bug repair. This is where you will see a shift in complaints from bug problems to lack of content release. (Note: all of this is a natural cycle in game creation). So once all the content is out we will see much more support on bugs and things. This is because the focus in game production will change. Then we will see a game release and the circle begin again. Content release, bugs, then final production. 

As for this specific patch, when they test these games they are very limited in the scope of the playing. They have people playing on a closed network all tied in together, versus on release where everyone from everywhere pushes the game. It was unfortunate that one of the times this patch error occurred was on a Friday release. That made everything have to wait until Monday. It's more of a poor situation rather than the fault of one specific person or company.

Overall, I played the game right after release and I play it now. The content and bugginess of the game overall have drastically improved. This is the first MMO Sandbox with multiple elements to hit Xbox. And thus there will be a bumper road more than the second and third games. While Wildcard has made mistake, so have we as consumers. Wildcard however has continued to produce a great game with great support, for an Alpha release game. I hope this helps with the question being laid out. 

-Lemur

 

Link to comment
Share on other sites

26 minutes ago, Kingsoup said:

It feels like this is the same sort of way I would patch\update a database server if short on time. I'd do some testing but eventually just roll it out and fix the problems I can't anticipate after implementing it.  As far as totally breaking private servers, you'd think this would have been tested on an actual console and quickly realize password connection doesn't work anymore, and controller going off crashes it, feels like that is obvious, but as usual the main focus is on official servers and PvP related balancing and patching. (which has nothing to do with my happy world of private server PvE which many of us love very much)

 

Also why drop a 24gb patch on friday late in the day? so much for friday night games after a long week at work, let alone the whole weekend no Ark and nobody can patch anything or talk to microsoft about whatever certification issues they need resolved until the next week. (god forbid you don't have a proper very high speed internet connection, days to DL that kind of patch)

It would be nice to be in actual final release production so their is a bit more accountability then, yes we dropped the patch as the last guy turned off the lights at WildCard for the weekend but its in DEVELOPMENT, so we can do whatever we want! (knowing us suckers are hooked on this wonderful game!)

Well, keep in mind, WC didn't drop the patch on Friday so to speak. They handed it over to MS on Monday last week and it took until Friday to get out of certification. WC wanted it out sooner (according to Jat's Twitter)  but they were not in control of it at that point. MS finished certification around 5 pm PST on Friday which is why the late weekend release occured.

Link to comment
Share on other sites

6 minutes ago, Bigskybear said:

Well, keep in mind, WC didn't drop the patch on Friday so to speak. They handed it over to MS on Monday last week and it took until Friday to get out of certification. WC wanted it out sooner (according to Jat's Twitter)  but they were not in control of it at that point. MS finished certification around 5 pm PST on Friday which is why the late weekend release occured.

To be clear, from what they've said it appears that WC doesn't control the time the cert finish, they do however, at least this is what they've said before, appear to control the time when patches are copied up to the servers where you download from.  Though once WC gives the OK it's again out of their control, again at least that is my understanding of the process.

Also, given the build numbers at a .2 instead of .0 likely there were at least 2 builds that restarted the process.  At least I'm assuming that based on the server version number being .0 and the client version numbers being .2, and soon to be at least .3.

While we don't have nearly the same number of customers for the product I work on and we're not under a time crunch or most times deadlines for releases, we don't normally release on Fridays as we normally go with what was released on a Thursday.  Mainly because Fridays are a good day for a dev to get in "just one more" fix, and we don't normally want folks to have to work late on a Friday or over the weekend.  We do however do deployments on Friday's for things that meet a certain threshold as far as if it resolves an issue that customers are impacted by or our internal customers are blocked in supporting our external customers.

Likely the call was made to release late night on a Friday as they thought the improvements made outweighed the problems that might occur.  Given the changes to use the newer version of Unreal I probably would have pushed back myself and made sure more testing was done, but again I'm most very much not aware of all the things involved in releasing games on the Xbox.  In hindsight, it was a mistake, though only a mistake if it could have been caught with their testing methodology/infrastructure.  Again, this part I'm unsure of and likely we won't get an answer due to NDA.  I've been watching a few other game developers as they stream and talk about this a bit more so I learn a bit more each time.

Link to comment
Share on other sites

  • Administrator
21 hours ago, tibs said:

pure hearsay at this point but this is what i've heard they're trying very hard to release asap so they're trying to jam in all the content they want so they could release as soon as april, if they can get to beta and start on optimization and bug fixes that gets them closer to release. and they might wipe servers upon release so anything that has happened till then is irrelevant

I can confirm that this is completely untrue.

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...