Porting Pinnacle Station forward 11 years | Hacker News Reader

Porting Pinnacle Station forward 11 years

me3tweaks.com

79 pointsDLay26 comments

Video games are a weird type of application. They seem to be one of the only applications that isn't in the process of becoming a web application (outside of just rending it on a remote server and even that isn't growing very fast). This is despite the fact that every platform/OS essentially has had it's own graphics library/interface (DirectX, OpenGL, Metal). At least Vulkan seems to be making a lot o progress in making more things cross platform although IIRC it leaves a lot of area to be implementation specific.

I think the dominance of Windows & console for video games has meant that developers are much more used to shipping immutable dependencies with their game. Relying on Windows SxS to avoid DLL hell. While they might provide updates for the game for a few years there are mostly game related updates rather than library updates (at least from what I have seen).

The big downside to this is that some things which shouldn't be dependent on the game developer providing updates like controller support, game engine fixes, etc.

I'm not sure if the problem is just the game developer/publisher not having any financial motivation to reexport/update their games or if the game engines themselves make it difficult to provide these kinds of updates.

Arguably they have an incentive not to support games for years. Not when they can re-release the same game with a little refresh 10 years later to "fix" issues.

Video games aren’t becoming web applications precisely because there are so many GPU libraries. I think this is also why they have so many immutable dependencies: they rely on low-level optimizations and various hacks which break on newer libraries.

Video games need to do expensive computations in real time. Most other apps don’t, which is why they can be ported to the web.

The other side is that video games have a lot less need to be correct than most applications. Even a blatant user-visible bug is probably fine (and might even add to the fun). Crashing is a problem, but running with massive state errors often isn't.

Running with massive state errors is a huge problem. There are untold amounts of video games that can become uncompleteable if you do a certain thing. Also called "hard lock". This is much worse then a crash.

There's a pretty small class of games where that happens in a relevant way. And tbh players and reviewers seem to punish that a lot less than a game that outright crashes.

I just played through the Legendary Edition and am going through Andromeda now. While I think ME1 is definitely the best, they all have really interesting mechanics that spice up each subsequent game in the series. I was presently surprised that Andromeda took a lot more risks that, imo, have been paying off so far.

Weapons ammo retcon was done very clumsily and broke the lore established in the first game. If they wanted to change the gameplay and make players break cover more often, they should have just hand vaiwe this issue without explanation.

I cannot upvote this enough. The reasoning was solid: for gameplay reasons, ammo worked better.

There's no way to really explain that in lore - but also no reason too: just make the change, and pretend it always was. Players are willing to just go with a solid change like that for solid reasons - you ruin it once you try to explain it in-universe.

I despised it so much that I even installed a mod that made weapons behave like in ME1.

I think there was a way to do it better: thermal clips should've been optional. Thermal clips were a fine idea in principle, but being able to vent when you've hit the clip's capacity or have run out of them should've been an option. That way you introduce ammunition without requiring such a massive retcon.

The ammo retcon was part of such a bigger shift. My main complaint is that adding the heat clips instead of cooldown gimped sniper rifles disproportionately. But shifting from the open outdoor environments with the Mako in ME1 to more contained levels in ME2 also disincentivized snipers.

I had most of my fun playing ME1 trying to take ridiculous long shots on targets outside of the radar range. With some configurations of mods, which ME2 also slashed, this was really effective and kind of hilarious. Some of the physics damage mods on sniper rifles would blow enemies right into the air.

Indeed. Roaming planets was my favourite part of the game, which gave the game a sense of a huge galaxy and a sense of venturing into the unknown. Also, those open environment were the only setting in all three games where using a sniper rifle was justified. 2&3 just felt so confined to linear corridor levels.

That's why I'm really enjoying Andromeda. Took a while to get going but the open world-esque exploration is great.

Agree! I loved the lore and universe built up in the codex

ME1 has more RPG elements but ME2 is the best of the serie.

me2 had some nice character moments, but its main plot was very stupid with ridiculous reasoning that threw away all appeal of me1. Best elements of me2 (and me3!!) were stories that expanded the stories, already established in the first game (genophage and quarians/geth conflict).

To anyone who thinks ME2 is better, I can only recommend this mighty analysis of the series: https://www.shamusyoung.com/twentysidedtale/?p=27792

Lots of parts, but it highlights all the problems of the second and third games really well.

It's weird to see so much praise for ME2 everywhere. I hated the game after ME1. There's basically no main plot. 90% of it is side quests to flesh out side characters. Which, fine, I see why people enjoyed so much, but it felt so pointless to me because the main plot boils down to "throw all these people at a final base and see who makes it out alive". I wanted to see the ME story moved forward, but it took until ME3 before that actually happened.

I agree ME2's story was pretty weak. ME3 definitely picked things up and I didn't hate the ending(s) as much as everyone else.

'lets build this McGuffin thing that we conveniently found at our doorstep at the very right moment and which we aren't even sure how it would work, but we will use resources of the whole galaxy to build it' story was hardly any better.

The movie Contact had a very similar premise but yeah, it's not the best plot device given where the writers could have taken it.

In the movie Contact the Earth wasn't under a genocidal attack by a overwhelming forces of a highly advanced alien race, they just received some blueprints that they thought should do something. And It was at least known how to activate the thing. With Crucible, they didn't even know until the very end.

Very different case, I'd say.

Haha you're right, definitely different. It's hard trying to defend it.

Also, ME2 committed character assassination of Liara, who, for some reason, unexplainably turned from a nerdy aecheologist into a deadly spy assassin. And the game railroaded the player into working for Cerberus. Ugh.

Easily the worst DLC in the game. I replayed it around 2017, a few years before ME:LE came out and it's basically some horde survival modes and timed killing sprees. ME1 didn't age well in general (it's kinda like Half-Life 1 - it had some good parts, but "was innovative when it came out but just doesn't play well to modern sensibilities") and this DLC was the worst part of it.

Mass Effect 1 is the best game in the series, and these guys deserve wholehearted praise for their efforts.

Reply on news.ycombinator.com