Braid Dev Says Auto-Save Warnings Waste Time and Money

[Sarah LeBoeuf]

Braid Dev Says Auto-Save Warnings Waste Time and Money

"Decades of developer time" are lost to needless console certification requirements.

If you've spent any amount of time playing videogames, you've probably seen a message telling you not to turn off your system while a game automatically saves. The warning is there to prevent gamers from accidentally corrupting their save files, and it's a standard occurrence in current-gen console gaming. Braid creator Jonathan Blow feels that gamers would be better served with "a more robust save system," calling the standard warning part of the "bizarre over-complexification" of console gaming.

The quick message might not seem like a big deal from a gamer's standpoint, but as Blow points out, "this is something that developers have to implement and that has to be tested, which costs significant time and money." The solution, he says, is to use a save system that won't corrupt files if the console is switched off. "You save the new file next to the old file, flush the I/O, and only delete the old file once the integrity of the new one has been verified (or just don't remove it, and keep two copies of the save at all times). On load you just make sure you load the intact save file if one is corrupted." This was the save system used in Braid, and according to Blow, it's not very difficult to employ.

It probably doesn't seem like it would take much effort to develop and test a warning screen, but Blow says that it could be at least three days of extra work for developers. On its own, that's not much when game development can take years, but combined with the other steps developers have to go through to get their console games certified, it can add up to "decades of developer time."

Blow details some of the other certification challenges indie developers face on his blog, and says that he doubts publishers will be able to fix these issues for the next console generation. He also points out that there's "almost no certification process" for iOS apps, and they seem to be functioning just fine. Blow is certainly not the first developer to complain about the certification process (not even the first this month [http://www.escapistmagazine.com/news/view/118578-Update-Fez-Studio-Rereleases-Save-Corrupting-Patch]), but we're still probably a long way off from seeing Microsoft, Sony, and Nintendo loosen up their game requirements.

Source: Kotaku [http://the-witness.net/news/2012/07/thoughts-on-consoles-and-certification-processes/]

Permalink

 

[Fasckira]

I guess I'd never really thought about it before. I like the warnings that say "Don't turn off when you see this symbol" purely because it explains what the symbol is and why its on my screen. I'd never just randomly turn off a game anyway, when it comes to it I rarely rely on autosaves tbh.

On the subject of Mr Blow however, and I mean this in a neutral manner as much as possible, is he actually doing any new games that we know of? I guess I don't understand why people focus on his opinions so much, they rarely tend to be particularly profound or insightful.

 

[cidbahamut]

All I'm getting from this is that we should remove features because some Indie developer is crying about having to do the work required to develop a game.

 

[9thRequiem]

Here's the thing : You do the extra work to follow his method. All that means is that the end-user doesn't know when progress is saved. This system will, if you turn off, fail to save.

Sure, you could add this on top of a warning, but you really need to let people know what's saved and what isn't. Perhaps this may work acceptably on Braid, where you don't need to save much at all (and so the saving is extremely fast), but for a large game, the chance of switching off during save is much larger.

Give me a warning any day.

 

[WhiteTigerShiro]

Man this guy has got some really petty for reasons to hate on console development. The warnings are there for a reason and are rather useful. It also seems like it would be a hell of a lot easier and cheaper to implement then his suggestion. Speaking of which I love his extremely vague suggestion of how long it takes to implement and test. What the hell is three days supposed to mean. Thats a rather useless term when you are talking about a development team.

It lets me know as soon as I start up a game wether or not I need to worry about saving it.

It helps prevent people from turning off their system mid save and loosing a lot of their progress.

It informs people what the symbol that appears in the corner of their screen is.

You should probably read the actual article instead of just the headline. It explains why most of what you just said is wrong.

 

[Yopaz]

Seriously why does anyone even care what this douche has to say anyway? He has made one mediocre games yet it seems like everything he points out as glaring flaws in the industry are petty complaints portrayed as news worthy. I haven't seen one example of this guy actually saying something I agree with. Making games that don't get their save files corrupted work for a lot of iOS games. Well how many iOS games got the same weight as a modern AAA title? I have had save data from games take up almost as much space as some of the lightest iOS games out there. It's like saying that laptops got something to learn from calculators since you "never" need to change the batteries on a calculator while a laptop only lasts a few hours.

You don't need a lot of knowledge to explain why this guy is wrong. You just need a tiny bit of common sense.

 

[Yuri Albuquerque]

I'm a developer myself (not a game developer, but I already done some things that needed CRC-32, MD5 or SHA-1 validation).

No, his suggestion is not difficult nor expensive. It's actually really good. It's possible to create a method that validates automatically for the programmer responsible for the auto-save (say, when I call save_file, instead of rewriting the file, it runs a code that do what he suggested).

The only drawback is very little performance lost (which is irrelevant, since auto-save is implemented into a separated "process", so it won't affect game's performance) and a little more space on disk needed (also irrelevant nowadays, considering how much disk space we have). And if the saving was validated we could delete the old saving anyway.

 

[MrMunchies]

Oh. You made a popular game?
Tell me how much you contributed to the videogame industry.

And before you ask; yes, I think that he's pretentious.

 

[Yuri Albuquerque]

None of you really understood his point, and you're complaining about something you simply don't get. What he is saying is that you should have a backup of your saving, and your new saving should be validated when loaded. This is not difficult to understand, nor to implement. What the hell are you complaining?

 

[Vivi22]

The warnings are there for a reason and are rather useful. It also seems like it would be a hell of a lot easier and cheaper to implement then his suggestion.

So you think it's easier and cheaper to implement an auto-save system, then implement a warning screen about the auto-save system and test both separately than to just make some minor changes to the auto-save system so save file corruption isn't an issue to begin with? How is making more work easier and cheaper? Make one system which doesn't suffer from the issue then test it. Requiring the warning screen is just adding unnecessary complexity to the game code and testing process.

It lets me know as soon as I start up a game wether or not I need to worry about saving it.

It helps prevent people from turning off their system mid save and loosing a lot of their progress.

It informs people what the symbol that appears in the corner of their screen is.

Except his whole point is that with a better designed saving system you don't need to worry about save corruption if the game is turned off while saving making the warning, and even a little icon saying it's auto-saving, completely unnecessary.

This is what WhiteTigerShiro was talking about when he accused you of not reading the article. Yeah, you're making points in response to what Blow said, but you're missing the fact that the simple idea he puts forward makes your points a complete non-issue.

 

[Mumorpuger]

Jonathan Blow strikes me as the biggest hipster, Apple fanboy.

Ehh. Mr. Blow, kindly shut up and make video games.

 

[WhiteTigerShiro]

You should probably read the actual article instead of just the headline. It explains why most of what you just said is wrong.

Speaking of reading why dont you try actually reading my Post before commenting. If you had bothered you would see that all of my fucking points are directed at his comments IN THE ARTICLE.


Normally I would make some snarky comments here insulting you and your reading comprehension but I think you already did that yourself.

Alright, I guess if you wanna play that game, I'll requote your post with replies purely taken from the article:

Man this guy has got some really petty for reasons to hate on console development.

"which costs significant time and money".

The warnings are there for a reason and are rather useful.

The solution, he says, is to use a save system that won't corrupt files if the console is switched off.

It also seems like it would be a hell of a lot easier and cheaper to implement then his suggestion.

which costs significant time and money

It lets me know as soon as I start up a game wether or not I need to worry about saving it.

It helps prevent people from turning off their system mid save and loosing a lot of their progress.

It informs people what the symbol that appears in the corner of their screen is.

The solution, he says, is to use a save system that won't corrupt files if the console is switched off.

Wow, and I only had to use two actual quotes. So who's supposed to be making fun of who's reading comprehension? Because it seems pretty apparent to me why I would assume you hadn't actually read the article. Though there is one tidbit of your reply that can't really be answered with quotes from the article:

What the hell is three days supposed to mean. Thats a rather useless term when you are talking about a development team.

Odds are he's referring to man-hours, so in this case it takes 72 hours worth of developer time (be it a single person taking a literal 3 days, or a team with that time divvied-up among them) to implement this feature. This is also how he would come the conclusion that the other certification hoops a developer has to jump through can cost up to decades worth of time, because when you have a development team the size that they are today working over the course of a couple years on a game, you easily get probably over a hundred years worth of man-hours that needs to be accounted for.

 

[jehk]

I completely agree about how consoles implement saves. I've programmed similar things using the Windows API and its trivial. However, it does require some lower level API functions (I used a combination of memory mapping and IO flushing functions) that are probably not available to console developers. It boggles my mind that console simply save over the top of old files.

 

[MetalMagpie]

Man this guy has got some really petty for reasons to hate on console development. The warnings are there for a reason and are rather useful. It also seems like it would be a hell of a lot easier and cheaper to implement then his suggestion. Speaking of which I love his extremely vague suggestion of how long it takes to implement and test. What the hell is three days supposed to mean. Thats a rather useless term when you are talking about a development team.

It lets me know as soon as I start up a game wether or not I need to worry about saving it.

It helps prevent people from turning off their system mid save and loosing a lot of their progress.

It informs people what the symbol that appears in the corner of their screen is.

I'm going to be greedy and say it would be nice to have both. It seems like there are two points here.

1. I like to know exactly when the game has saved. This is especially important in games that don't offer any sort of manual save (such as Assassin's Creed). When I want to stop playing, I wait until the next time I see the saving icon, then quit once it's done.

2. However, it seems bad (from an engineering perspective if nothing else) to rely so heavily on nothing "going wrong" whilst the game is saving, whether "going wrong" is the player turning off the game, a pet pulling out the cord, or indeed a power cut. In an ideal world, you should never lose more than one "chunk" (the space of time between two save points) of game progress.

So how about game developers try (where they can and where it makes sense) to implement more robust saving mechanisms, but still display a "saving" icon. If the save mechanism is more robust, it lessens the need to be quite as dramatic with the "DON'T SWITCH OFF WHEN YOU SEE THIS ICON" warnings. Simply stating that the symbol indicates the game is saving should be enough to inform most people that they'll lose progress if they pull the plug before it has finished. But even if they do they'll only lose progress made since the previous save point.

I realise that doesn't deal with the "problem" he's complaining about, but I'm not convinced that's a big issue anyway.

</rambling-thoughts>

 

[WhiteTigerShiro]

None of you really understood his point, and you're complaining about something you simply don't get. What he is saying is that you should have a backup of your saving, and your new saving should be validated when loaded. This is not difficult to understand, nor to implement. What the hell are you complaining?

It's been said that the person with the most to say on a matter knows the least about it, and gamers are always quick to knee-jerk a response based on the source. In this case, an indie developer is pointing out a flaw with the 360's certification process, so of course the reaction is going to be blind rage, because how DARE he suggest anything bad about the great and holy X-Box.

 

[Scrythe]

I can't remember the title, but I remember having a PC game back in the day that all of it's autosaves and manual saves weren't overwritten, so a full playthrough would leave me with about 20-odd saves.

Wait, I just remembered what game it was. Vampire: The Masquerade - Bloodlines.

I guess I forgot all about that because I haven't played it unpatched/unmodded since high school, and I rarely (if ever) save my game. Play it enough times and you can blow through a single playthrough in one sitting.

I'm not sure about it now, but thinking back I do believe that the first Dungeon Siege also did this, but I haven't played that in way too long.