Fog Creek Software
Discussion Board




Video games dont crash, why?

I still play video games and I have noticed that out of all the games that I have played for the last 10 years, I have had only a few games that lock up or crash.  And I am assuming that video games are pieces of software(console games only), what techniques do software game developers use to ensure that games 'dont' blow up.  Actually the only console games that I have played that have locked up a couple of times are from Bethesda Software(Morrowwind), which was still a great game.  Are graphics engines so refined that anybody can write a game and not screw it up?

I dont think it is that game developers are just better programmers, I have talked to young game developers(20-25-30) who work in teams of only 10 or 20 who work normal 40 hour weeks, so it is not that.  Maybe games are easier to create?

Any ideas?

Berlin Brown
Saturday, July 10, 2004

Perhaps because they are coded so close to the hardware that any 'crashes' are by definition 'not compilable'

.
Saturday, July 10, 2004

And of course some common sense observations, the flow of games are different than that of regular software.  Some people will spend 8 hours a day with a piece of software(let us say MS-Word), they are changing the state of that application in terms of configurations, menu options, opening windows, opening files.  Regular software is a different beast.  When working with a video game, since the user is only given a controller with 5-6 buttons on it, the state of the game cant change too much.

Berlin Brown
Saturday, July 10, 2004

And actually, video games do crash. If you get pass the "Video Game Crash" of the 80s from this Google search:

http://www.google.com/search?q=video+game+crashes&btnG=Search&hl=en&lr=&ie=UTF-8

You can see that games do indeed crash. In, here are some specifics:

X-Box
-------
http://www.google.com/search?hl=en&lr=&ie=UTF-8&q=xbox+crashes&btnG=Search
http://www.devhood.com/messages/message_view-2.aspx?thread_id=69574
http://news.teamxbox.com/xbox/2087/Xbox-Kiosk-Crashes-Addressed
http://www.avforums.com/forums/archive/index.php/t-107537.html

Playstation 2 (harder to find)
--------------------------------
http://www.google.com/search?q=%22playstation+2%22+%2Bcrashes+-racing&btnG=Search&hl=en&lr=&ie=UTF-8
http://www.theregister.co.uk/2000/03/08/sony_to_recall_all_playstation/
(Actually couldn't find any others)

Did find this review:

http://www.gaming-age.com/cgi-bin/reviews/review.pl?sys=ps2&game=ffxi

Which talks about the benefit of not being on a PC is that the programmers only had to focus on one platform, one graphics card, one player, etc, which enables a more solid game.

CF
Saturday, July 10, 2004

> Perhaps because they are coded so close to the hardware that any 'crashes' are by definition 'not compilable'

That makes no technical sense at all.  In fact it may make crashes more likely.

The reason that games don't crash is that they are expensive to produce and have a short shelf life.  The only way for a game to be successful is for it to be stable when it comes to market. 

christopher (baus.net)
Saturday, July 10, 2004

Games crash plenty of times in development, it's just that for console especially you cannot get approved by the manufacturer if you have any crashes at all. So you don't even submit it until you've absolutely tested everything in house or even external QA.

But it's usually bad or unexpected data that causes crashes, and like somebody said above, there are usually only limited inputs and no user-created data in console games.

Game Programmer Lou
Saturday, July 10, 2004

What games are you playing? 

Games crash all the time.

Billy Joel on Software
Saturday, July 10, 2004

Since the XBox is merely a PC (Pentium III running Windows) it shouldn't be surprising that it crashes a lot.

Joel Rifkin on Software
Saturday, July 10, 2004

"it's usually bad or unexpected data that causes crashes"

Really?  Over the years I've seen many programs where you click on a button or select a certain menu item and *CRASH*.

How is that "bad or unexpected data"?

Joel Kleinholtz on Software
Saturday, July 10, 2004


Few games crash, actually. Even on the XBOX.

The reason: rigorous QA - testers play the game all day, trying all kinds of things to break the game. This goes on all the way through development. And then you have the publisher QA teams, having a go at it again before it's approved.

For XBOX specifically: Microsoft does not allow game patches (bugfixes) through XBOX Live, so your game better be perfect from the first day on.

PC is a whole different story, but that's because it's got several thousands different configurations.

YMMV.

A Pterodactylus Ate My Baby
Saturday, July 10, 2004

I'm not a big games person but when  do play it seems like I'll play for 24 hours straight in one game, which is a lot of up time for one program.

I've never seen a game crash.

A lot of these games use the same game engine, so that game engine is very well tested across many games for stability and the game designers just use level editors to put the game together with no programming needed.

Games are obviously doing a lot of things at once and doing them well, interface with a vide variety of specialized sound and video hardware, and so forth, so its absurd to say that games are less complex than business apps.

Game developers do take far more pride in their work that developers of other sorts of software.

Dennis Atkins
Saturday, July 10, 2004

Yep. Games don't crash because you can't patch them & the game must also have the manufacturer's stamp of approval before you can release it.

Just think of the console game market being the same as the pre-internet software market. I suspect that in 10-20 years when every console in the world is hooked up to the internet, crashes & patches will become more common.

Take, for example, a game like The Sims. The PS2 version wasn't upgradable at all with the hundreds of downloadable skins, wallpapers & furniture that helped make the PC version the best selling game of all time.

But once you can do that, what's to prevent you from downloading & installing a patch as well?

Oh, and I have had the occasional console game crash on me. Like Dennis, I don't play all the time, but when I get a new game (rarely) I tend to play it a lot for the first few days.

I remember Baldurs Gate Dark Alliance on PS2 crashing on me. My girlfriend & I played through the game twice about a year apart, and I played a bit on my own too. Sometime after the 2nd time we'd beaten it, I was playing & hadn't saved in a long time, and the game just froze on me.

www.MarkTAW.com
Saturday, July 10, 2004

Deja vu

http://discuss.fogcreek.com/joelonsoftware/default.asp?cmd=show&ixPost=152267

Janonymous
Saturday, July 10, 2004

Games rarely crash (in particular console games) because it is one type of software in which consumers will not tolerate crashes.

For regular PC software, they have been trained to accept crashes as "normal".  And the ability to send patches after the original ship date makes buggy software more likely to be released.

I disagree that it is because of the console being a fixed and known platform.  The bugs that crash most PC software are mostly the result of plain bad programming or inadequate QA, not variations in hardware configurations.

T. Norman
Saturday, July 10, 2004

"I disagree that it is because of the console being a fixed and known platform.  The bugs that crash most PC software are mostly the result of plain bad programming or inadequate QA, not variations in hardware configurations."

Nope, wrong.  I work in the game industry, and I can vouch for the fact that QA on consoles as opposed to PC's is greatly simplified by uniform hardware configurations. 

A 3D intensive game often will run fine on one PC with a particular type of video card, but will crash on another PC.   

Immature programmer
Saturday, July 10, 2004

PC games crash all the time. In fact, it's a running joke now that no game can be released to retail until the first patch is ready.

I've never had an Xbox game crash. I've never had a GameCube game crash. I've only ever had one PS2 game crash (GTA: Vice City). PC games crash all the time. Like a previous poster said, the uniformity of hardware is a HUGE win for solid console games.

Brad Wilson (dotnetguy.techieswithcats.com)
Saturday, July 10, 2004

Pretty simple. The developer's plataform is _exatly the same_ of the user's plataform.

Nix
Saturday, July 10, 2004

Console games in part don't crash because of the testing that they go through.  Yes, it is true that hardware targets being stable is a big factor, but they are tested much more than any othe piece of software I've written in my 20 year career.

I've spent many nights on the office futon soak testing and babysitting tests of the game being played by AI bots and 15 year old boys with too much time on their hands.  It is rigourous.  The amount of certification that goes on is quite encouraging.  (One time we turned the automatic bot into a feature that allowed the game to be recorded and then saved and allowed you to race against yourself.)

Stil, even with all that, almost every game I've shipped has had a crash bug in it somewhere, just very rare to find.  Champions of Norrath has crashed on me in several places (I didn't write that one, just played it), as has Metroid Prime.

Game Developer
Sunday, July 11, 2004

Damn, Champions of Norrath is the next one my girlfriend & I want to play. I hear it's better than the Baldurs Gate II (zzzz).

www.MarkTAW.com
Sunday, July 11, 2004

Console games might crash rarely, since the dev box is the same (usually) as the customer box.

PC games crash like, well, like windows. If you think games do not crash, you have not played evercrack (everquest). This is a game that is played online by about 400,000 customers with wildly variant PC platforms. It gets an update pretty much every month. If you ask any eq player about the directx9 patch, you will get an earful. As games go, it is rather notorious for some updates requiring PC upgrades (err, the last 2 computers I purchased were to be able to keep playing my evercrack). I've managed to get 3 BSOD on a brand new (dell business PC) XP machine in the first month I've had it from that game.

I have gotten DirectX 9 bugs from 2 different games so far that lock up the graphics card in a 640x480x4bit mode that can only be cured with a power cycle.

Peter
Sunday, July 11, 2004

"Nope, wrong.  I work in the game industry, and I can vouch for the fact that QA on consoles as opposed to PC's is greatly simplified by uniform hardware configurations."

My statement was about PC software in general, not games.  When your financial or word processor or photo editing software crashes, it is seldom because of a particular video card or hardware variation.

T. Norman
Sunday, July 11, 2004

"My statement was about PC software in general, not games."

This thread is about games that crash, not PC software in general.

Game Programmer Lou
Sunday, July 11, 2004

In discussing why games don't crash, it is still useful to discuss why non-games do crash.

T. Norman
Sunday, July 11, 2004

I also suspect it is much easier to raise an army of teenage testers to enthusiastically thrash a console game to death than it is to get people to do the same to your component that performs advanced SQL searches on a car insurance database?

Ian H.
Sunday, July 11, 2004

"In discussing why games don't crash, it is still useful to discuss why non-games do crash. "

Not when you are trying to argue that the reason console games don't crash is not because of uniformity of hardware.  It has a great deal to do with rigorous QA which is made possible by uniformity of hardware.

Granted, there are other factors other people have brought up.
1) Game developers take more pride in their work. 
Yeah, I suppose that's a factor.  The good ones take great pride and ownership in their work and are embarrassed by bugs; they'll spend the extra time to make sure everything's just right.  But it's not quite accurate to say other types of developers generally take little or no pride in their work. 

2) It's easier to get people to rigorously test games than it is other types of apps.
Yeah, that's true.  But it's not that much harder to get people to rigorously test a PC game than a console game.

3) Crashes in games, like embedded software or mission critical apps, are catastrophic and intolerable.  This is not necessarily so with other  types of PC applications. 
Ok, I agree with you on that point T. Norman. 

4) Sony and Microsoft have rigorous standards for approval of commercial titles for PS2 & X-Box respectively.  There are  no certification/approval standards for PC applications in general.
I've heard of it happening, but I have yet to work on a console title that got approval on the first submittal.

I'd have to say uniformity of hardware and official approval standards are the biggest reasons why console games have high QA standards and rarely crash when compared to other types of software (including PC games).

Immature programmer
Sunday, July 11, 2004

I would also add that most game programmers are significantly better than the average non-game programmer.  Game programming requires understanding of so many different things that are beyond what the average programmer is willing or able to learn.

T. Norman
Sunday, July 11, 2004

I'll echo the comments thus far: console games, and games for game-oriented platforms that existed in days of yore (like the Amiga) may generally be pretty reliable, but PC games are buggy pieces of shit.

The latter appears to be something that's getting worse, not better.  Console consumers are intolerant of buggy software; PC games just seem to accept that 1.0 won't run right, and that they may have to wait until they've applied many patches.  Online gaming appears to have exacerbated this.

Rodger Donaldson
Sunday, July 11, 2004

The testing appears on an exact copy of the machine you have in your living room, and this is the only setup that exists in the wild. Unlike the PC, where there are billions of possible setups.

Just me (Sir to you)
Monday, July 12, 2004

I'd never seen a console game crash till I played Metroid Prime on my brother's GameCube.

In the first ten minutes or so it froze on a screen and I had to power down and restart from the beginning.

Fifteen minutes into the second try, the graphics suddenly went all weird and spiky like the polygons' anchor points were "off"; it was unplayable.

After that I decided to find something else to do.

Oddly, my brother seems to have finished the game at home with such problems.  Perhaps this was an application of the ten-foot rule someone mentioned in another thread.

Kyralessa
Monday, July 12, 2004

Driver 3.

(I rest my case)

Mike
Monday, July 12, 2004

Let's amend that to:Nintendo games hardly ever crash.  I would like to see a long list of genuine Nintendo crashes, where there wasn't a GameShark or some hack tool involved.  I have played many of Nintendo's consoles since the first one in '85, and I've NEVER seen any Nintendo game actually just stop dead.  I have, however, made other systems do that, usually because the console got too hot and just quit.  Keep in mind that in order to "crash", and appear as such, gameplay would have to stop -- that's a pretty major problem.  There are lots of little glitches in games, but usually they go by too fast to see or to impact gameplay that drastically.  I.E. with 3d games, being able to "see through walls/floors/etc" when in certain spots is pretty common.

sir_flexalot
Monday, July 12, 2004

My opinion on why game development is somewhat better than non-game development has to do with the introduction of "big money" into the field.

Academic research in computing is somewhere between 5 and 30 years ahead of the general industry. That gap is far smaller in game development than in the average IT shop.

If you want to see where theory hits the pavement, it would be looking at that industry. Post Mortems are something never done in any IT company I have ever worked for, but very common in the game industry.

Peter
Monday, July 12, 2004

Well, Metroid Prime did crash on some cubes, as was pointed out. But in general I agree, and that's just a special case of the rule: the stricter the control a publisher exercises over releases, the fewer the crashes.

Nintendo games are usually extremely stable and they also have very few third-party games. The other consoles have a more diverse selection, and therefore more junk. The PC has no single publisher in control at all, and you get games that crash like there's no tomorrow. Unless the particular game's developer/publisher takes care, as Blizzard or Bioware do.

Chris Nahr
Monday, July 12, 2004

"Academic research in computing is somewhere between 5 and 30 years ahead of the general industry."

Where do you base these numbers on?

Just me (Sir to you)
Tuesday, July 13, 2004

simple case of economics...

software is as bug free as it needs to be...

Ken
Wednesday, July 14, 2004

*  Recent Topics

*  Fog Creek Home