tigerfort: the Stripey Captain, with a bat friend perched on her head keeping her ears warm (Default)
[personal profile] tigerfort
I kinda-sorta promised in my last rant that I'd come back to NMS, so here we go. Again, this is a game I've had a lot of fun playing. But in addition to the total failure to live up to what the devs said/thought they could deliver, there are a number of design issues.



Leaving aside once more things like the dreadful writing, the fact that someone thought "the end of the story leaves you back at the beginning again" was not only clever[1] but original, and the sheer unexamined colonialism (which the developers, naturally, see as totally nonpolitical. Le sigh), let us speak of unforced design catastrophes. I'm going to pick on two examples in particular; one from the UI, and one from the game engine, and describe them. I'm not going to try and understand why they are the way they are; that side of things I'll leave at a despairing "Why?".

[1] I have no objection to cyclical stories as such, it's just they're so rarely done well.

So, the UI. The UI overall is wildly inconsistent and uneven, to the extent that even the menus don't all work the same way. On the most basic level, the same button activates "sprint" on foot, and "boost" in a vehicle. But sprinting toggles on and off, whereas boosting needs the key to be held down. In a ground vehicle, that is. In a ship, boost sometimes needs holding down and is sometimes a toggled mode, depending on circumstances. On screens like the inventory, the same key can perform wildly different functions on adjacent pixels. It's a mess.

But there's a reason I kept going on about the conversation UI last time. Even by the standards of NMS, the conversation UI is just incomprehensibly bad. Consider selling some nanites to a nanite trader, which is pretty typical and consists of the following steps:

  1. activate the conversation UI by pressing and holding "E" while facing the target

  2. wait while the galaxy's slowest typewriter prints the initial text to screen one character at a time (you can mouse click to speed this up)

  3. left-click to acknowledge that you've seen the initial text

  4. wait (or click) through the interaction text

  5. left-click to acknowledge that you've seen the interaction text

  6. move the mouse pointer from the centre of the screen to the interaction option you want

  7. click-and-hold the left mouse button for ~1s to select that option

  8. wait (or click) through the response text

  9. left-click to exit the conversation UI



If you wanted to sell more than one set of nanites, or trade with someone and learn a word, tough. You've got to go through the whole lot again, because it apparently didn't occur to the designers that taking you back to the list of interaction options was a thing that might be good. ("End conversation" is always one of the options, btw, despite every other option also ending the conversation.)

Sometimes, you'll get a chatty alien, and the initial text will be broken into two or even three parts, none more than about a dozen words, but each of which requires separate clicking. Better yet, when you get a random introductory bit, with initial-intro text, intro-question text, and intro-question response all before you get to the start of the bits listed above.

It's awful, and totally unnecessary. Far and away the worst conversation UI I've ever seen. Even just assuming that the player can handle having more than one sentence on screen at a time would be a huge improvement.


So much for UI design. What about the engine?

Mostly, I don't feel qualified to comment on the game's understructure. But there's a certain fascination to the awfulness of the rendering engine. Rendering a massive, complex 3D world is a massive, complex task, requiring a pretty large complex piece of code. I'm going to assume (without evidence, but I'm feeling generous) that there was an actual reason the developers couldn't use one of the existing 3D rendering engines that already has a large dedicated team of people who spend all their time optimising making stuff appear in 3D on computer screens.

But even if you're rolling your own rendering engine, why on earth does it apparently take several seconds after getting out of your ship for the game to figure out what's in front of you? Rocks, plants, animals and even major landscape features continue to pop suddenly into view for at least a couple of seconds. They aren't obviously being created in any particular order, and certainly not on the basis of how close they are or whether you can see them.

"Oh, that's a pretty flower on the ridge, I'll [pop] walk round the huge metal deposit that's appeared between me and it, totally blocking it from view. I wonder if this planet has any [pop][pop] Oh, it does have trees [pop][pop][pop][pop][pop] lots of them, in fact[pop]. Don't see any anim[pop] crap, I've just walked into a carnivorous elephant."

OK, game engines that need to render procedurally-generated open worlds have to do quite a bit of heavy lifting, but I've just never seen anything else that gets this so totally wrong.

Given that it already takes a couple of seconds to actually get out of the cockpit, I don't understand why that time can't be used to work out what the player can see and render that first before they hit the ground and start moving. It's bizarre.

Profile

tigerfort: the Stripey Captain, with a bat friend perched on her head keeping her ears warm (Default)
tigerfort

June 2023

S M T W T F S
    123
4 5678910
11121314151617
18192021222324
252627282930 

Most Popular Tags

Style Credit

Expand Cut Tags

No cut tags