207 réponses à ce sujet

[DarkSpooky]

Some people just do not get how bug-tastic coding can be, especially if things are interconnected. Change one decimal or two and the whole system could either cascade in an unpredictable fashion or they could inadvertently create sky-net. There was a story written by devs who did Stalker where they talk about how the AI was CONSTANTLY doing things they did not expect, same goes with Black&White, where the pets did things the devs did not intend.

Some of you guys think coding is easy? No harder than a car? HAH! A car is a physical object that can be diagrammed and any problem can be IDed without too much hassle. Bug hunting in code could be likened to looking for the shortest needle in a needle stack. I don't know too much, I do know enough how problematic it can be, about code but I do deal with people who do. Coding can be one of the MOST infuriating and time consuming part of game design.

Modifié par DarkSpooky, 27 juin 2012 - 12:08 .

[Mgamerz]

Patches typically come out around MP DLC time now, don't they? Sans Patch 1, since it was just a general fix-bugs-post-launch.
Edit: And SP DLC that adds new content. Excluding EC since it doesn't add new content that needs major changes.

Modifié par Mgamerz, 27 juin 2012 - 01:25 .

[xtorma]

Bryan Johnson wrote...

I think Mgamerz and Ninja Stan have a good understand of most of the challenges faced. Of course there is more to picture that can't be discussed for various reasons.

I can say we do have a list of issues, we do watch the forums. Oh and we do play our own game, have I personally seen issues called out, yes. There has been attempts to try and catalog everything that people have noted. The more places we as employees have information written the more places information is to get lost. That is why we have our own internal bug tracking methods and we include forum information in them. Every fan made attempt (that I have seen) to maintain a list has not lasted more than 2 weeks from the difficulty it is to verify things etc.

Another note, especially about multiplayer is user style testing needs 4 people and then as Ninja Stan mentioned there is a ton of questions. Is this something that just came up in the last patch, has it been there since ship, does it occur only after extended play, does it only happen on a certain platform etc.

Oh and yes I am QA

So once again , how long do we have to wait?

[KiraTsukasa]

Mgamerz wrote...

That picture is a good representation of the coding cycle. You hit that 'requirements have changed' block quite a bit.
I typically use eclipse to program, and when you miss a single , and 2000 lines of code stop working and it underlines the entire code as being wrong and you have no idea where that goes...
Drives you freaking nuts.

I had that same problem when learning COBOL. One single period out of place and it causes something ten lines down the road to stop working, but nothing inbetween and only highlights that one line that isn't working. At least, that's how it worked (or didn't work, depending on how you want to view it) in the version I was using.

I hate COBOL.

[Edalborez]

xtorma wrote...

So once again , how long do we have to wait?

As long as it takes.

[Stinja]

I'm a .NET enterprise application developer, so fully "get" what Mgamerz, Ninja Stan, Byran Johnson and other devs say.
As an user it's still annoying when you run into bugs you thought should of been caught

Every user thinks their issue is the most important, and don't realise the bigger picture devs are under:
- limits from management
- manpower shortages
- time pressures
- poor specs
- changing specs
- not enough testing time allowed
- etc etc

At some point a product has to ship, known and unknown bugs, and basically is it good enough?
I more like to moan about ME3's design decisions, but again that is unlikely to be the devs fault: they're just crunching in their cubicles, trying to normalise on caffeine, headphones on, coding idiotic specs...

Lastly if we're sharing coding comics, this - right here - is the story of my life:

[ryanshowseason3]

I'd also like to harp on the difficulties of working inside of large projects, especially ones that use third party code or worse dlls. ( Meaning another company or individual made it ).

Coding in college and getting a degree is one thing. Working in a foreign codebase that you didn't design or write nearly any of the code is entirely different. You don't know challenging until you have worked in a codebase with over a gig in source code. To even hire a new developer to help the team out it WILL take them on the order of months to years to fully understand the system and freely work in it without help. The fastest I've seen is around 3-4 months.

Some bugs are easy to find despite so much code if you have a good instinct of where it might be. Without any instinct or clue... It becomes a hay straw in a needle stack figuring out where it came from. Not to mention if you find out it is coming from third party code. Or a dreaded third party dll, if its coming from something like that where you don't have any control over the source code it is a very dark day, stuff like that MAY NEVER be fixed. Because then even if they fix the bug you found their next version may have OTHER bugs that break MORE things in YOUR codebase. THEN WHAT DO YOU DO? You can't upgrade because it breaks things but so does the current version. Cry and pray is all you can do.

So yeah even if you have a a CS degree... If you've never worked on a project of large magnitude, you still don't have a clue.

Modifié par ryanshowseason3, 27 juin 2012 - 01:55 .

[Mgamerz]

Typically in a CS course (Taking Algorithms next semester, then AI), you build everything from scratch, and projects don't get that big. Obviously a one man team can't make mass effect unless that one person invented more days of the week.
Another thing is that when you are on a team, not everyone thinks one way. They might have diagrammed the code out in UML (Hate UML...), so that interfaces between files works. (Not sure what language it's written in, going to imagine C++ or C# for the most part)
If your team doesn't work together with a single mindset, when people finish parts they have to work a bit more to get all of them to work together.
Like the US government.... except you know, they never work anything out.

[BattleSoldier]

I am a Java application developer and I fully understand & support what Mgamerz, the Bioware devs & QA guys are talking about. It's literally a pain in the arse to debug through thousands of lines of code. Seeing this huge game product, I would say Bioware has really done a good job in delivering and maintaining this product. They are also the active listeners of the community!

[rgeshevv]

Nice read! And I agree with you, that patching is not an easy job. So, let me see if I get this right.

The developer have schedules, plans, projects, etc. and everything is linked with money given from the publisher. So, they don't have money or they don't want to lose more money and they say: "No! We won't give you money, so that you can make a patch. The game is working. Nothing game braking. Just a few minor bugs. The players will live with them, if not, will stop play it."

Well, I have to agree with the publisher. Why waste money for a game that sold millions of copies? It doesn't make any sense to me. Better to give money to the developers for a new game - new revenue.

That's my opinion. And yes, I'm not a programmer and I don't know how complicated things are. I'm just some dude that bought a game with some minor bugs in it...

[Mgamerz]

That's about right. There is typically a 'support' frame for the game too. I mean, you most times a company won't just release a single patch to game and be like 'done'. Most times the game will be supported for 6 months to a year. ME2 saw support for over a year through DLC (though not enough huge bugs made it so they had to patch it often). It depends on if it's worth patching and spending the money to improve the game depending on the lifespan of the game itself.

If you don't support your games post-launch fans will see it as just pushing out a product to make money. Which, obviously they are, but I'm sure most of the devs also like their job. Except the debugging part. Nobody likes debugging [smilie]http://social.bioware.com/images/forum/emoticons/cool.png[/smilie]
Have to remember that BioWare is a company, and a companies job is to make money.

Modifié par Mgamerz, 27 juin 2012 - 03:19 .

[Ensiform]

So assuming they actually made it through certification and QA testing, how did they only fix ~3 of the bugs listed on the bugfix list? If you couldn't actually have tested them enough to see fully fixed, why did they bother to post that? That's my only issue with their patch process. If it were me I would start prioritizing fixing some of the glaring game breaking issues before I continue with the weekend events and any DLC.

To be honest, the previous games didn't really need as much patching but this has multiplayer where in the lifespan is meant to be longer.

[Mgamerz]

Bug fixing requires work from both SP and MP teams I am sure since most game mechanics will affect both.
Again, not all bugs are easy to find. Perhaps they were attempting to fix Vanguard glitch, but the fix didn't work and they had a deadline to meet. It was either 'don't release a patch and have fans all angry that we don't listen' or 'fix some of the bugs, and have to let the others go'. Though I'm not sure why they were listed as fixed. The devs are only human. They can only do so much.
Also, I don't think weekend events have too much to do with QA. I mean, I'm sure there IS QA on the events, but most change's aren't radical (and the weekend events themselves don't seem to change the game, only the post-event).

[rgeshevv]

Mgamerz wrote...
Have to remember that BioWare is a company, and a companies job is to make money.

Yes! But not for expence of the quality. One and only Moto nowadays: "Quantity over quality." 

More money, crappy content... Publishers moto....

Though, I doubt BW as a developer, don't want their products to be top of the line, it's just EA's crappy policy....

[ryanshowseason3]

Ensiform wrote...

So assuming they actually made it through certification and QA testing, how did they only fix ~3 of the bugs listed on the bugfix list? If you couldn't actually have tested them enough to see fully fixed, why did they bother to post that? That's my only issue with their patch process. If it were me I would start prioritizing fixing some of the glaring game breaking issues before I continue with the weekend events and any DLC.

To be honest, the previous games didn't really need as much patching but this has multiplayer where in the lifespan is meant to be longer.

     You're making the mistake of assuming stopping everything else will make fixing a bug go any faster. And I'm sure they "thought" they fixed some of it but failed to cover all cases(vanguard) or didn't test deep enough( ULM ).

     In practice putting more than 2 resources.( Programmers are known as resources ). On a single bug will not get it fixed any faster. Not to mention if reproducing the issue is a problem in itself you can be doing something else while QA or testers make it happen first then you can sit down and debug it.

     Also testing dlc, weekend events and bug fixes can be done at the same time if performed properly doubling your throughput of content. For instance the new map is already made it just needs tested, also the weekend event putting banshees into other factions is coming along as well. So why not slap ULM on an smg play as vanguard against cerberus on the new map? So you're testing ULM, vanguard glitch, weekend event content, and the new dlc map all at once.

     This is the kind of sentiment that most people who don't have a CS degree and further who haven't worked on massive software projects just don't get. You don't just halt production, if they did the developers would be twiddling thumbs half the day. How does that help?!

[Antagony]

"According to wikipedia"

Yeah, all your credibility is now gone.

[tslewis2]

What is causing my PS3 to freeze regularly (SP/MP)?

Modifié par tslewis2, 27 juin 2012 - 04:09 .

[Mgamerz]

Antagony wrote...

"According to wikipedia"

Yeah, all your credibility is now gone.

You should learn how references work:
References

^ a b c d e f "How Many Lines of Code in Windows?" (Scholar search). Knowing.NET. December 6, 2005. Retrieved 2010-08-30.

This in turn cites Vincent Maraia's The Build Master as the source of the information.

[xtorma]

Basically what it is , is that we can not expect any bug to be fixed. We can get lucky, and they can get the time/money resources alocated for a particular bug, but we most likely will never see our justicars work at thier potential, because it's not that big a deal to them.

I wish it were not true, but there it is. It is a business afterall.

This will change my view of them as a company though , and the way they address bugs in thier games will become a deciding factor in whether I purchase another title. I only have past performance to base these sorts of decisions on.

Modifié par xtorma, 27 juin 2012 - 04:19 .

[Barneyk]

Even though everything written in the OP should be common knowledge, it sadly is not and I am glad he posted it as eloquently as he did!
Good job!

But I would also like to say, the game needed more polishing before shipping.
Both in gameplay, in multi player, in some finesse in the story, etc etc etc.
Whoever decided to push it out this fast was wrong in doing so imo.

[ryanshowseason3]

xtorma wrote...

Basically what it is , is that we can not expect any bug to be fixed. We can get lucky, and they can get the time/money resources alocated for a particular bug, but we most likely will never see our justicars work at thier potential, because it's not that big a deal to them.

I wish it were not true, but there it is. It is a business afterall.

This will change my view of them as a company though , and the way they address bugs in thier games will become a deciding factor in whether I purchase another title. I only have past performance to base these sorts of decisions on.

Not necesarrily. Theres hope, not all doom and gloom. While you're waiting for a large but hard to reproduce bug to come around it makes nice busy work to tackle smaller easier to find bugs. I usually fill my long build times with writing new features.

[ToaOrka]

Maybe I'd be more sympathetic towards Bioware if they didn't say some of these glitches were fixed, only to be proved wrong. Otherwise though, I completely agree with you.

[Xaijin]

I applaud BioWare QA that we did not see any SEV 1, 2 or 3 issues

Actually there's a SEV 1 issue with sound on creative cards that causes spontaneous reboots. Not lockup, not blue screen, a literal reboot. It's been present since launch and hasn't been fixed.

[xtorma]

ToaOrka wrote...

Maybe I'd be more sympathetic towards Bioware if they didn't say some of these glitches were fixed, only to be proved wrong. Otherwise though, I completely agree with you.

From what i understand about it, they are able to test, but in a limited capacity. once a fix goes live, the devs hold thier breath and hope they didnt miss anything. sometimes they do, and it looks like they didn't test. In actuality, they tested as much as they were able, time permitting.

I don't think anyone wants to go to thier boss and say....."the fix didn't work"

[Mgamerz]

Forget the boss, remember most of the community...