4 ms·

> I'm an online dev for multiplayer games, and I disagree with your take. In my experience the sheer number of bugs we've seen from "I tested it in the local on

> I'm an online dev for multiplayer games, and I disagree with your take. In my experience the sheer number of bugs we've seen from "I tested it in the local only build mode so it's fine" and the impact that has on the project makes it worth _not_ supporting an offline mode. I'm not an exec, but practicing good habits during development (there's a phase where it's not worth it sure, but by the time you're on a live project...) makes actually building a game easier. Both should be available cause yes, you need to test like you said. But if I'm setting up skinning on a character or cloth simulation on a cape, I want my turn around time as fast as possible from DCC to engine. Maybe I just want to validate my PBR materials with a specific level's lighting rig and baked GI/reflection probes. Those specific examples are client side only, so I absolutely want that to work on an offline build. The longer a roundtrip to the engine takes for content (including the time it takes to start the game up), the harder it is to iterate and thus quality also falls. Forcing online always for devs at the very least, makes booting the game slower (or just fail) and you need to reboot the game constantly (even with hot reload). That cost adds up quick across many devs and many years. You need both lest you drive your artists and TAs insane. My last company hemorrhaged artists to literally go work next door on very similar styled art content from one live service game to another just because our tools were so exhausting to them to roundtrip, and online wasn't even forced! >https://x.com/Pilestedt/status/1760077808146014340 https://x.com/Pilestedt/status/1760077808146014340 - helldivers backend team was 4 people. _That's_ the point of this newfangled tech. WoW took 4-5 years to develop with "The original World of Warcraft was created by a team of 40 people, which eventually doubled in size as the launch drew close." https://www.polygon.com/2020/1/20/21070494/world-of-warcraft-history-development-chaos-wow https://www.polygon.com/2020/1/20/21070494/world-of-warcraft... Where as as Helldivers 2 has about 100 devs at the end of development taking 8 years according to their CEO. So not seeing the wins there. In terms of concurrent players around launch, WoW had around 0.5 million within months and Helldivers 2 seemed to have peaked at 0.75 million across PC and PS5. 19 years apart and not really seeing that 5x in practice... None of this suggests to me that this "tech" is a useful as many think. > Handwaving away this problem doesn't make it just og away. I mean, we have tons of precedent of these middleware licenses working just fine when redistributed as binaries to users' devices, so how is this handwavy? Like yes, there's an issue with already shipped games that make use of such licenses, but after such a law passes, the licenses will have to adapt to licenses similar to what already exist. And prior to this recent problem in games, middleware licenses worked just fine too and often shipped with server binaries. These licenses are not the norm as explained by all the middleware that does get packaged with almost every game today...

maccard 2y ago
Providing testbed maps for rapid iteration and is not the same thing at all as allowing you to play offline. > WoW took 4-5 years to develop with "The original World of Warcraft was created by a team of 40 people, which eventually doubled in size as the launch drew close." You’re moving the goalposts here. You asked about cloud tech initially. The topic of “have game budgets and teams gotten out of control” is a totally different one. > so how is this handwavy? These aren’t just game middleware technologies that we’re talking about - things like Java, SQL Server, have non distributable licenses. Systems like LaunchDarkly require active subscriptions and don’t allow for redistribution and are often deeply embedded in applications. > And prior to this recent problem in games, middleware licenses worked just fine too and often shipped with server binaries Some games still do ship with server binaries. Valheim is a good example. But, to your point earlier about WoW vs Helldiverd - games aren’t developed the same way as they were “prior to this recent problem”, player behaviour and expectations have changed. You can’t just go back to the way things were here any more than you can with social media, banking, email, movie rentals, etc.
- sighinggoldfish 2y ago
  > Providing testbed maps for rapid iteration and is not the same thing at all as allowing you to play offline. Sure, but I mean, at this point, I'll even take a guarantee that I can just run around on a map without spawns just so the world art is preserved. The initiative is pretty broad about what it means to be playable for this reason. Actually, that's the whole point of this initiative is to establish such a bare minimum. And if the game is truly a service, and not a good, than this wouldn't even apply! Companies would just have to be explicit about the service nature of the product which they are not unless you pay a subscription like in the WoW case, which this doesn't affect. So you could still do all this tight coupling that is apparently good for development and ship games just fine if players are properly informed on the service nature (which means some kind of expectation for how long the service will last). I agree that a testbed map isn't the same as playing offline, but I don't agree that it's preferred, or even natural, that development require that offline play be so arduous to accomplish. Especially when something as complex as WoW or PSN matchmaking + a game's specific netcode (Demons Souls PS3 emulation case) can be reverse engineered without any support from devs. Most arguments along the lines of "it's just too technically difficult" do not past the smell test to me. > You’re moving the goalposts here. You asked about cloud tech initially. The topic of “have game budgets and teams gotten out of control” is a totally different one. I feel like I pretty succinctly explained that this tech did not provide a huge boost in productivity in development nor did it allow for substantially more scalability of players. Not sure how that's moving the goalpost. You posted a tweet highlighting 4 people allowed for 5x the player capacity in a week. I don't know how many multiplayer engineers worked on WoW pre initial launch but considering the team was smaller than Arrowhead, I doubt it's significantly more than 4 and their concurrent player numbers around launch weren't significantly different (certainly not 5x). I get that Blizzard probably predicted bigger launch numbers than Arrowhead so there's at least some merit to being able to scale up fast. But overall, I think, like most modern software "advances", we've been duped on their actual performance and efficiency gains, while most gains have actually come from better hardware and internet connections (hardware again). And end users really aren't feeling much of an improvement. I think more has been lost to users and players actually. Less mods, less weird community servers with new game modes, etc. And as mentioned earlier, this actually helped incumbent devs find new revenue streams! The new supposed features of this tech is more to serve all the meta stuff like progression tracking and matchmaking which requires greater centralization of online services. Ye in many cases, like Helldivers 2, the actual gameplay is P2P. So the only thing players get out of the this new "tech" is the thing designed to keep them addicted which is stuff like lootboxes and Skinner box progression mechanics. (I acknowledge my last points here on meta mechanics as goalpost moving, I do view this as part of a much larger issue) > Systems like LaunchDarkly require active subscriptions and don’t allow for redistribution and are often deeply embedded in applications. I mean, this is just a bad state. I view this initiative as fighting shit like this. It's bad for devs too. What happens if LaunchDarkly or some other platform holder goes down or cuts off access to a dev for some perceived breach of conduct or license agreement? That's so damaging to a developer, and if they aren't established, existentially threatening. We shouldn't encourage this kind of coupling. As for the non distributable licenses of common distributable like Java and SQL, I'm not terribly worried the economics won't catch up to make it feasible to just use distributable license. We've already seen lots of Java technologies relicensed to GPL 2 over a decade ago. It's all possible without collapsing the economics of it all. > games aren’t developed the same way as they were “prior to this recent problem”, player behaviour and expectations have changed. You can’t just go back to the way things were here any more than you can with social media, banking, email, movie rentals, etc. We certainly can go back, or better yet, do better than we used to! I really don't think there's a good technological arguments for why we dev this way now. It's mostly political and I think we should resist this type of excusing of rent seeking style software as "tech advancement". All those areas you mentioned are the way they are not because of technological necessity but because of political and cultural forces. We can style them in a way that's more compatible with a more free and expressive user base, and ultimately, a freer market. The more, but still limited, openness of the pass certainly contributed greatly to the grown of games!

maccard 2y ago

Providing testbed maps for rapid iteration and is not the same thing at all as allowing you to play offline. > WoW took 4-5 years to develop with "The original World of Warcraft was created by a team of 40 people, which eventually doubled in size as the launch drew close." You’re moving the goalposts here. You asked about cloud tech initially. The topic of “have game budgets and teams gotten out of control” is a totally different one. > so how is this handwavy? These aren’t just game middleware technologies that we’re talking about - things like Java, SQL Server, have non distributable licenses. Systems like LaunchDarkly require active subscriptions and don’t allow for redistribution and are often deeply embedded in applications. > And prior to this recent problem in games, middleware licenses worked just fine too and often shipped with server binaries Some games still do ship with server binaries. Valheim is a good example. But, to your point earlier about WoW vs Helldiverd - games aren’t developed the same way as they were “prior to this recent problem”, player behaviour and expectations have changed. You can’t just go back to the way things were here any more than you can with social media, banking, email, movie rentals, etc.

sighinggoldfish 2y ago
> Providing testbed maps for rapid iteration and is not the same thing at all as allowing you to play offline. Sure, but I mean, at this point, I'll even take a guarantee that I can just run around on a map without spawns just so the world art is preserved. The initiative is pretty broad about what it means to be playable for this reason. Actually, that's the whole point of this initiative is to establish such a bare minimum. And if the game is truly a service, and not a good, than this wouldn't even apply! Companies would just have to be explicit about the service nature of the product which they are not unless you pay a subscription like in the WoW case, which this doesn't affect. So you could still do all this tight coupling that is apparently good for development and ship games just fine if players are properly informed on the service nature (which means some kind of expectation for how long the service will last). I agree that a testbed map isn't the same as playing offline, but I don't agree that it's preferred, or even natural, that development require that offline play be so arduous to accomplish. Especially when something as complex as WoW or PSN matchmaking + a game's specific netcode (Demons Souls PS3 emulation case) can be reverse engineered without any support from devs. Most arguments along the lines of "it's just too technically difficult" do not past the smell test to me. > You’re moving the goalposts here. You asked about cloud tech initially. The topic of “have game budgets and teams gotten out of control” is a totally different one. I feel like I pretty succinctly explained that this tech did not provide a huge boost in productivity in development nor did it allow for substantially more scalability of players. Not sure how that's moving the goalpost. You posted a tweet highlighting 4 people allowed for 5x the player capacity in a week. I don't know how many multiplayer engineers worked on WoW pre initial launch but considering the team was smaller than Arrowhead, I doubt it's significantly more than 4 and their concurrent player numbers around launch weren't significantly different (certainly not 5x). I get that Blizzard probably predicted bigger launch numbers than Arrowhead so there's at least some merit to being able to scale up fast. But overall, I think, like most modern software "advances", we've been duped on their actual performance and efficiency gains, while most gains have actually come from better hardware and internet connections (hardware again). And end users really aren't feeling much of an improvement. I think more has been lost to users and players actually. Less mods, less weird community servers with new game modes, etc. And as mentioned earlier, this actually helped incumbent devs find new revenue streams! The new supposed features of this tech is more to serve all the meta stuff like progression tracking and matchmaking which requires greater centralization of online services. Ye in many cases, like Helldivers 2, the actual gameplay is P2P. So the only thing players get out of the this new "tech" is the thing designed to keep them addicted which is stuff like lootboxes and Skinner box progression mechanics. (I acknowledge my last points here on meta mechanics as goalpost moving, I do view this as part of a much larger issue) > Systems like LaunchDarkly require active subscriptions and don’t allow for redistribution and are often deeply embedded in applications. I mean, this is just a bad state. I view this initiative as fighting shit like this. It's bad for devs too. What happens if LaunchDarkly or some other platform holder goes down or cuts off access to a dev for some perceived breach of conduct or license agreement? That's so damaging to a developer, and if they aren't established, existentially threatening. We shouldn't encourage this kind of coupling. As for the non distributable licenses of common distributable like Java and SQL, I'm not terribly worried the economics won't catch up to make it feasible to just use distributable license. We've already seen lots of Java technologies relicensed to GPL 2 over a decade ago. It's all possible without collapsing the economics of it all. > games aren’t developed the same way as they were “prior to this recent problem”, player behaviour and expectations have changed. You can’t just go back to the way things were here any more than you can with social media, banking, email, movie rentals, etc. We certainly can go back, or better yet, do better than we used to! I really don't think there's a good technological arguments for why we dev this way now. It's mostly political and I think we should resist this type of excusing of rent seeking style software as "tech advancement". All those areas you mentioned are the way they are not because of technological necessity but because of political and cultural forces. We can style them in a way that's more compatible with a more free and expressive user base, and ultimately, a freer market. The more, but still limited, openness of the pass certainly contributed greatly to the grown of games!