Yopaz said:
Alexnader said:
Yopaz said:
snip
You don't need a lot of knowledge to explain why this guy is wrong. You just need a tiny bit of common sense.
Ah, common sense, it's not like that's ever steered anyone wrong before. You make reference to the sometimes ridiculous sizes that save files can get to, however that's not really all that much of an issue. I'm sure it used to be in the days where memory was at a premium but we're dealing with machines that handle gigabytes of information. Writing the save file to a new location as opposed to writing it over the old location is not going to hurt performance too much as you're doing essentially the exact same operation. Only now with a significant reduction to the risk of data loss. This is of course assuming that there aren't any complicated optimisations that rely on overwriting one tiny bit of the save file each time.
One thing I learned from dabbling in programming is not to trust the end user to be smart. Keep the workings of your program as isolated as possible. Why give them a warning about corrupting their save file when you can just take away that risk entirely? Sure, if they interrupt Blow's save process they could lose progress however with frequent autosaves you could limit that loss to at most one level's worth of progress. It's far better than having them lose their entire save file. The lecturer in my computing class was used to dealing with self-taught coders who knew intricate, hacky ways to shave a millisecond off some operation and he made pains to drill into them that simplicity and security should almost always take priority over small optimisations.
Anyway I don't get why people are chewing Blow out over this, it's a sensible proposal for a different save system and above all it's a technical one that should be of little interest to most gamers, especially console gamers many of whom joyfully profess their ignorance to the workings of the magic box that sits under the TV and "just works". This is something a few programmers would quietly discuss while sitting around a table and gesturing at a system diagram or something. I'm glad that the Escapist ran this article and gave us an insight into the technical workings of game development but it's not really news and is certainly not "RABBLE RABBLE RABBLE" type news.
I'm not saying common sense is never wrong. I am saying it's common sense that if a space shuttle got problems a person with extreme knowledge on cars (chosen as a less advanced vehicle) is probably unable to walk in and scoff and fix the problem for the rocket scientists. If there is a massive problem in something advanced there have probably been attempts at fixing them once.
Yes, the world should be like Pokemon Trading Card Game where your GameBoy could run out of batteries and when you replaced the batteries you would continue at the exact same spot as where you left off. a game made by Nintendo, one of the biggest video game companies have made a foolproof save system once. This is not something just seen in iOS games. If a company that has already done this once haven't done it again despite how they managed to do so perfectly why haven't they done it with any recent games? Either too hard, too expensive or borderline impossible seems to be the best reasons. Blow said they should remove the warning to save money and work. I am guessing he does not know how much effort it takes to make a system that does not corrupt save files for an AAA title based on the fact that he's only made one game. I am willing to guess you have never made one either. Rule of the thumb though, if something is a problem now and has been a problem for a long time, there probably isn't a quick easy fix.
Rules of thumb are just as useless as common sense. I could say another rule of thumb is that if the problem has been there for a long time odds are no one's been bothered to actually look at it. One of the graduates taught by my old lecturer came into a company and saw that the numerical model they were using took days to run once, she went in and studied the system for a bit, implemented a simple optimisation and the model only needed hours to do what once required more than a day. This doesn't prove your rule wrong and the only thing that ever could is the aggregate experience of a statistically significant number of human beings on earth. Consequently that's the only thing that could prove it right.
Blow has more experience than either of us when it comes to developing games. He successfully implemented his save system in a game that was used by thousands of people. I daresay neither of us can dismiss his opinion out of hand. There is another developer in this thread who has agreed with Blow's assessment too.
However to give up and profess ignorance is no fun, so lets try some of your common sense. Why does common sense dictate that a save file for Call of Duty, arguably the most AAA of the AAAs, should be any different to a save file for Braid? What data need be stored that's specific to your save file? Time played, location, difficulty setting, your weapons and ammo, possibly health. All of these are stored as variables, initially defined by a high level computing language, interpreted into symbols and then eventually compiled into machine code. Ultimately 1s and 0s. Not necessarily any different to what Blow used. Aside from procedurally generated games, sandboxes or those with narrative choices like Mass Effect most games would have no more complex a save system than Braid would have.
Why are we even talking about AAA titles? If you read John's original article [http://the-witness.net/news/2012/07/thoughts-on-consoles-and-certification-processes/] we see that he makes most reference to other 2D games similar to Braid. He doesn't try to tell other developers what to do, he says that Microsoft should streamline its creditation system and build his solution into its own existing save API, that way developers don't even have to think about the problem at all.