[flagged]
replies(8):
On top of that, the hardware requirements (256MB of system RAM, and the PlayStation 2 only had 32MB) made it enough of a challenge to get the game running at all. Throwing in a heavyweight parsing library for either of these three languages was out of the question.
Most of the time, the programmers who do this do not follow the simple rule that Stroustrup said which is to define or initialize a variable where you declare it (i.e. declare it before using it), and which would solve a lot of bugs in C++.
Yeah but we're talking about a 2004 game that was pretty rushed after 2002's Vice City (and I wouldn't be surprised if the bug in the ingestion code didn't exist there as well, just wasn't triggered due to the lack of planes except that darn RC Chopper and RC plane from that bombing run mission). Back then, the tooling to spot UB and code smell didn't even exist or, if at all, it was very rudimentary, or the warnings that did come up were just ignored because everything seemed to work.
They wrote a JSON “parser” using sscanf. sscanf is not bulletproof! Just use an open source library instead of writing something yourself. You will still be a real programmer, but you will finish your game sooner and you won't have embarrassing stories written about you.
If you don't know what “good” looks like, take a look at [Serde](https://serde.rs/). It’s for Rust, but its features and overall design are something you should attempt to approach no matter what language you’re writing in.
To be honest, I just don't like how you disparaged the programmer out-of-context. Talk is cheap.
The only C code that I have recently interacted with uses a home–grown JSON “library” that is actually pretty good. In particular it produces good error messages. If it were extracted out into its own project then I would be able to recommend it as a library.
Apart from that, many of us thought that Java serialization was good if just used correctly, that IE's XML parsing capabilities were good if just used correctly, and so on. We were all very wrong. And a 3rd party library would be just some code taken from the web, or some proprietary solution where you'd once again have to trust the vendor.
But it is even more important for today’s game studios to see and understand the mistakes that yesterday’s studios made. That’s the only way to avoid making them all over again.
Besides, the complaint about not having a heavyweight parser here is weird. This is supposed to be "trusted data", you shouldn't have to treat the file as a threat, so a single line sscanf that's just dumping parsed csv attributes into memory is pretty great IMO.
Definitely initialize variables when it comes to C though.
> And a 3rd party library would be just some code taken from the web, or some proprietary solution where you'd once again have to trust the vendor.
Open source exists for a reason, and had already existed for ~15 years by the time this game was begun. 20 years later there are even fewer excuses to be stuck using some crappy code that you bought from a vendor and cannot fix.
It's only now that single player progress is profitable to sell that video games have taken save game encryption to be default.
It's so stupid.
This isn't that uncommon - look at something like Diablo 2 which has a huge amount of game data defined from text files (I think these are encoded to binary when shipped but it was clearly useful to give the game a mode where it'd load them all from text on startup).
Your average hire for the time might have been self-taught with the occasional C89 tutorial book and two years of Digipen. Today’s graduates going into games have fallen asleep to YouTube lectures of Scott Meyers and memorized all the literature on a fixed timestep.
Nothing has changed appreciably. If they would let you login to a private invite-only lobby that would likely speed things up greatly— but it’ll never happen.
> If they would let you login to a private invite-only lobby that would likely speed things up greatly— but it’ll never happen.
Did they remove this option in the last couple years?
One of the anecdotes from Titan Quest developed by Iron Lore is that their copy protection had multiple checks, crackers removed the early checks to get the game running but later 'tripwires' as you progress through the game remained and the game appeared to crash. So the game earned a reputation for being buggy for something no normal user would hit running the game as intended.
But also keep in mind in 2004 the legality of many open source projects was not really tested very well in court. Pretty sure that was right around the time one of the bigger linux distros was throwing its weight around and suing people. So you want to ship on PS2 and XBOX and PC and GameCube. Can you use that lib from inside windows? Not really. Can you build/vs buy? Buy means you need the code and probably will have to port it to PS2/GameCube yourself. Can you use that opensource lib? Probably, but legal is still dragging its feet, and you get to port it to PS2. Meanwhile your devs need a library 3 weeks ago and have hacked something together from an older codebase that your company owns and it works and means you can hit your gold master date.
Would you do that now? No. You would grab one of the multitudes of decent libs out there and make sure you are following the terms correctly. Back then? Yeah I can totally see it happening. Open source was semi legally grey/opaque to many corporations. They were scared to death of losing control of their secret sauce code and getting sued.
However, it's not even remotely "like crack". Crack is really really really really fun, period, no "just enough of the time" about it. The reason people get hooked on crack is because it's guaranteed to be fun.
If I had to choose a substance that most closely mirrored variable ratio reinforcement conditioning, it'd probably be ketamine.
I worked in gamedev around the time this game was made and this would have been very much an ordinary, everyday kind of bug. The only really exceptional thing about it is that it was discovered after such a long time.
What? No. What even are you suggesting? Hell, games with OFFICIAL MODDING SUPPORT still require you submit bug reports with no mods running.
Editing game files has always been "you are on your own", even editing standard Unreal config files is something you wont get support for, and they are trivial human readable files with well known standards.
>One of the anecdotes from Titan Quest
Any actual support for this anecdote? Lots of games have anti-piracy features that sneakily cause problems, and even could fire accidentally. None of those games get a reputation for being buggy. Games like Earthbound would make the game super hard and even delete your save game at the very end. Batman games would nerf your gliding ability. Game Dev Tycoon would kill your business due to piracy.
None of these affected the broad reputation of the game. Most of them are pretty good marketing in fact.
struct test {
int my_int = 0;
int* my_ptr = std::nullptr;
};
You cannot initialize them with a different value unless you also write a constructor, but it not the issue here (since you are supposed to read them from the file system)
https://nee.lv/2021/02/28/How-I-cut-GTA-Online-loading-times...
And in 2004, didn't have a published specification, or much use outside of webdev (which hadn't eaten the world yet).
> and SAX parsers are 22
And, especially at the time, pretty much exclusive to Java, right?
Put another way, which are the high-quality open-source implementations of those formats that the developers should've considered while working on SA in 2003 and 2004? Or for that matter, in the 2001-2002 timeframe, when the parsing code was probably actually written for use in VC?