Game Discussions Star Citizen Discussion Thread v12

"For Vanduul combat, a section of gameplay was completed where enemy perception is reduced when searching for the player. This led to discussions of how to implement low visibility and other sense reductions across the board. They then prototyped the initial implementation for this specific gameplay use, which will also be used as a springboard for fully developing the feature. Vanduul perception and investigation progressed too."

I don't even.

For combat with aliens nobody has seen outside a janky model view animation were it ineffectively poked with its sci-fi stick...

How to implement low visibility? Seriously? I mean, we joke about backers not having played a game since the 90s, but sometimes you wonder if the devs haven't developed a game since the 90s.


Oh... yeah.

What is a subsumption tool? Are they developing text editors now?

My guess is its a fancy way of assigning goals and motivations to NPCs. But probably with a high fidelity front end so it looks good when they show it to backers, whereas most devs would be happy with, indeed, a text editor.

So, if i may, i'll take an example from NWN2. Each NPC had a number of triggers you could assign to it. On heartbeat was one. On Conversation. On Combat. On Death. And so on. About 9 or so in total. Things that would trigger the NPC to do something based on certain events. The On heartbeat one was the one where you really could go crazy giving them a life.

An amateur might start out giving custom scripts to each NPC. Like a barman who only serves drinks and wanders around a bar doesn't need the same actions as an NPC guard right?

But once the number of NPCs start piling up and you have many different types of them each with their own idle convo lines, unique and shared conversations, walk locations, things of interest, actions they can take, etc. Then what you really want to do is bring it all together with master scripts and instead of attaching unique scripts to each NPC, which can be a mess to maintain and bug fix, instead, assign the NPCs roles and character types (you can still do unique when you need to, just give them a unique tag).

So, in NWN2 you could add tags to an NPC, so we might have tags like: JOB=Barman, Location=City, Character=grumpy, Voiceset=Oldman

Then each NPC for each trigger calls the same base script and then that looks at the tags as necessary to determine what the NPC should do in each situation.

You can then make them as simple or as complicated as you want them to.

A guard can have a patrol route. Say there is a Guard with an ID=CityGuard1 then you have the heartbeat script tell it to follow the waypoints with its name attached (or walk between specified destinations or randomly walk) and only if certain events trigger (combat happens nearby, someone steals something in their perception range, someone talks to them) do you initiate other activities, then when over, it goes back to the base routines.

The NWN2 toolset and scripting set were quite limited, but you could still make it do some fancy stuff with a bit of clever coding.

This is basically what CIG are doing, but probably in a much more complicated and unnecessary way and taking like 100x it would if done the same way other developers have been doing it for decades, with no benefit and probably, because its CIG, may even be worse than other implementations.
 

Viajero

Volunteer Moderator
It feels like a warm up ;)

(Weirdly he still seems to be making EDO vids, despite that grand exit ;))

Yeah. From what I can see there is also very strong qualitative differences on the type of rant topics in each.

One one hand his opening main rant is about a very rare engineering material very difficult to find (certainly just a bug) at the time, that is preventing him to unlock the "gosh!" last of the EDO engineers, out of around "gosh!" 12 or 13 he has already unlocked, and through marathonian gameplay sessions of 16 hours... And then he also rants a lot about FDEV communication post EDO launch and how player concurrency dropped at the time and how he gets less clicks. Almost shed a tear.

On the other hand for SC he rants about a game that simply kills him for no good reason every other minute and imparts upon death all relevant death penalties even when most of the time the player is not at fault for it, and even includes a rage inducing in game bug outtake section.

I mean, I am not going to judge about players deciding what they want to rage or rant about, but it is clear one game is actually played to oblivion, the guy even burns out due to excess. The other seems to be "played" as long as the frequent sudden death and other self admittedly rage inducing bugs allow him to.
 
Last edited:
Development is speeding up!!!

j4vfbfxrsrc81.jpg


Oh...

Tech had a bad 2021...

Source: https://www.reddit.com/r/starcitizen/comments/s8f7na/expectations_and_reality_of_2021_roadmap/


Art, ships, & pratfalls did quite well though ;)
 
Install it on a SSD! I mean, how many gamers don't have a SSD these days?
Already said here = many have a SSD but with SC installed on a HDD anyway!
Because as games of more than 50Go are common now and SSD of less than 250 Go are common also, gamers keep their SSD for their favorite games. So when they try SC during the free fly, they put SC on the huge HDD disk of 2To they added beside their SSD.
As you love bingo, very often when someone come on Reddit with "I don't understand why I have stutters/missing textures with my above average computer", it's solutionned with one of the BINGO card "SC installed on HDD" or "pagefile not set on SSD".
 
From the SC report

The AI team as a whole closed out the year creating feature testmaps for AI behaviors and usables.

Feature testmaps are new autonomous testing processes that run each day and print out reports, which are hugely beneficial to the team as they enable them to identify bugs significantly quicker than before.

“A significant challenge we always had on the team was understanding when an issue was specific to a certain feature or if an external change from another team caused the problem. Patterns identified by the feature tests will solve this problem." AI Team

Currently, the team are testing approximately 50 usables and 20 AI behaviors that can be interacted with by the player. Typically, a usable has between three and six directions characters can approach them from and the same number they exit from, each playing a different animation. With such a high number of permutations, automating the process was essential. The team will soon move to more advanced feature tests for the more complicated behaviors, such as the bartender, engineer, and security guard.

Do you have any thoughts on this tool? Is automating these tests a common practice in the AAA industry?
 
You got the quote from Erin in 2018-2019ish? stating that things wouldn't be added to the roadmap unless they were pretty certain they could deliver? ;)

Apparently they've got a whole new approach to that. Retro-active additions ;)

-I've noticed CIG have been retconning their old patches on the roadmap a little. Several cards have been added/removed from 3.3/.5, 3.6, 3.12, and 3.13 since I last made a record of them. This change apparently occurred sometime during Q4. Edit: Miscalc on one cell
Source: https://imgur.com/a/t8CRPnm


I find it odd that they would retcon all the way back to 3.3/.5 but I dont keep records of what the cards were in the past.
Taking a case example, lets look at the 4 new AI cards for 3.12, these are:
  • AI intercepting torpedos
  • AI Fire Mode Usage and Targeting Priorities
  • AI Accuracy Convergence
  • Capital Ship Combat Behavior Improvements
These I believe are technically correct, they could have said this at the time and people would have gone with it. In general I notice that sometimes CIG will bulk up patches with cards that are very related. Like adding bulk to the meat.

EDIT:

Lol, from the author again. Doom on the sub ;)

Source: https://www.reddit.com/r/starcitizen/comments/s80b3r/316_was_the_smallest_patch_on_record_while_315/htdnz3y/
 
Last edited:
Yeah. From what I can see there is also very strong qualitative differences on the type of rant topics in each.

One one hand his opening main rant is about a very rare engineering material very difficult to find (certainly just a bug) at the time, that is preventing him to unlock the "gosh!" last of the EDO engineers, out of around "gosh!" 12 or 13 he has already unlocked, and through marathonian gameplay sessions of 16 hours... And then he also rants a lot about FDEV communication post EDO launch and how player concurrency dropped at the time and how he gets less clicks. Almost shed a tear.

On the other hand for SC he rants about a game that simply kills him for no good reason every other minute and imparts upon death all relevant death penalties even when most of the time the player is not at fault for it, and even includes a rage inducing in game bug outtake section.

I mean, I am not going to judge about players deciding what they want to rage or rant about, but it is clear one game is actually played to oblivion, the guy even burns out due to excess. The other seems to be "played" as long as the frequent sudden death and other self admittedly rage inducing bugs allow him to.
I can't take it serious. One hand he rants he doesn't even want to play SC anymore - next minute he announces 10 bet things for beginners and how he loves SC. I tend to believe he is just collecting the cash and SC allows him to do so better than EDO maybe.
 
From the SC report



Do you have any thoughts on this tool? Is automating these tests a common practice in the AAA industry?

My main takeaway is that this should subsume them in embarrassment ;)

The team will soon move to more advanced feature tests for the more complicated behaviors, such as the bartender

But as to automating that specific type of process, who knows. It's the type of unremarkable in-house tool that devs don't talk about much, and which gamers generally care about even less. Devs mainly just deliver a game with NPCs that walk up to things and mime interaction. Rather than spending 10 years talking about doing it.

Maybe some day SC will too? That would be more interesting ;)
 
Already said here = many have a SSD but with SC installed on a HDD anyway!
Because as games of more than 50Go are common now and SSD of less than 250 Go are common also, gamers keep their SSD for their favorite games. So when they try SC during the free fly, they put SC on the huge HDD disk of 2To they added beside their SSD.
As you love bingo, very often when someone come on Reddit with "I don't understand why I have stutters/missing textures with my above average computer", it's solutionned with one of the BINGO card "SC installed on HDD" or "pagefile not set on SSD".

If people don't have a SSD drive for modern games these days they are going to have a bad experience. Not just with SC.

Sure, you will still get those people who install to HDD, but that number is shrinking, and as we've seen time and again, the CIG faithful have declared bad performance due to someone not having it installed on a SSD only for the person to turn around and say "it is installed on a SSD", at which point the faithful look for something else to criticize about the person's computer.

But fair enough on Hawkes for pointing it out, you're right in that it should be on a SSD, like a lot of modern games.
 
My god! CIG have invented automated tests!!!!

Time to tell all my QA automation colleagues that they now have a job thanks to CIG!
Fool of me to have not been more precise here...
As someone who uses automated tests in my work on a daily basis, I am not asking about the use of automated test. I am asking for the creation of an automated test with what appears to be a real map with NPCs/PCs interacting with objects. This type of test is normally done manually by QA testers right?

Maybe some day SC will too? That would be more interesting ;)
Why ?
 
From the SC report



Do you have any thoughts on this tool? Is automating these tests a common practice in the AAA industry?
Any competant developer should be creating tools to check the validity and integrity of both game environment and entity data, as well as code outcomes, from the outset and especially if you have a large complex environment and population of entity types (see my post above about the number of bugs found in the product I ran the development on). If CIG are only now doing this after 10 years, then better late than never, but points to larger problems...
 
Last edited:
Development is speeding up!!!

j4vfbfxrsrc81.jpg


Oh...
That's a really interesting chart.
I can only assume from it that CIG have solved all their networking, persistence and entity handling problems and are running across multiple threads/cores with no unintended locks or crashes. Afterall they've barely touched the Core Tech since 3.10.....or have they and the "cards" were just a front? ;)
 
Last edited:
Any competant developer should be creating tools to check the validity and integrity of both game environment and entity data, as well as code outcomes, from the outset and especially if you have a large complex environment and entity types (see my post above about the number of bugs found in the product I ran the development on). If CIG are only now doing this after 10 years, then better late than never, but points to larger problems...
To create complex autotests that emulate human QA for core functionnalities, you need to have the core functionnalities almost done and stable if you don't want to redo your autotest tools at each patch.
Doing it at late stage of pre-alpha does not indicate a problem, it's the right moment to create them indeed.
 
"For Vanduul combat, a section of gameplay was completed where enemy perception is reduced when searching for the player. This led to discussions of how to implement low visibility and other sense reductions across the board. They then prototyped the initial implementation for this specific gameplay use, which will also be used as a springboard for fully developing the feature. Vanduul perception and investigation progressed too."

I don't even.

Hang on guys, after going into the tank for 5 hours, I think I have this solved:

proximity_detect_vanduul = 0.75 * proximity_detect_default;

You're welcome.
 
Back
Top Bottom