Return to “Dev Logs”

Post

Monday, June 12, 2017

#1
Monday, June 12, 2017

I'm afraid that I'm going to be keeping this log briefer than any of us would like (no, I mean *actually* brief this time :ghost: ). If you're reading this, you no doubt have noticed the new forums. Talvieno and I worked over the past week to upgrade the forum software, which turned out to be hundreds of times more painful than expected. In the end, the greater part of my time was dedicated to forum-related things rather than LT code, so there's kind of a double-whammy situation going on here: on the one hand, I don't have much to talk about, on the other (same?) hand, I really don't want to spend more time not working on LT code :?

---

Josh: Forums Fires and So Forth

Already said it, pretty much. phpBB3.0 -> 3.2. Pain. Database migration. Pain. Old style did not support the new version (or vice-versa). Had to find new style that looked decent and close enough to our old one that it wouldn't shock and terrify everyone too badly (trust me, prosilver, the default theme, would shock and terrify you all; no one would touch the LT forums ever again :ghost: ). Pain, time. Time. Lots of time.

LT-side, minimal stuff going on. ECS wrap-up delayed, but still feel good about it. Worked quite a bit at getting a very simple graphical API for Sean to use for simulations. Not much fun watching console output all day for simulations! Especially if you want to get into the low-level AI like dogfighting...and he does.

Adam: BSP Perfection

Adam has officially perfected the robustness of our BSP trees. Like, to an insane degree. Floating-point precision has been told to shove off. Millions of rays being cast against complex shapes, perfect intersections being returned and never any false positives or negatives regardless of the input, etc. It's awesome. He's moving on to optimization now. This is the part where our narrow-phase physics tests become blindingly-fast!! I'm excited. If and when he's successful, we'll be looking at having more milliseconds than we know what to do with :D

Sean: Continued Colony Work + Low-Level AI

Sean and I spent some time this week talking about how the 'low-level' AI systems work (i.e., the AI that executes specific actions that are directed by the high-level AI). He's very eager to get working on some of this stuff like combat maneuvering, obstacle avoidance, pathfinding, etc. Unfortunately, with things rather in-flux on the Lua side due to the migration to the new ECS, we don't have a great way to do 'simple' visualization tools that aren't in C (like the ones Adam is using for BSPs). That's not at all ideal, since Sean is working in Lua and we want to keep it that way. Thankfully, as of a few hours prior to writing this log, I've finished building a very simple visualization tool that will let him write visual simulations. I'm very excited to see what we can dig into come tomorrow :) I anticipate the coming week being an interesting one for AI!!

---

Apologies for the brevity, but time is in remarkably short supply at the moment while work is in remarkably high supply. Send us all your positive vibes for this week -- with luck we'll be nailing low-level gameplay systems, physics, and AI all at the same time! :squirrel:
“Whether you think you can, or you think you can't--you're right.” ~ Henry Ford
Post

Re: Monday, June 12, 2017

#3
And what a perfect time to have brought others onboard!

Even with Josh busy fighting the forum fires, the team as a whole still made real progress. That's a win.

To specifics: this isn't a question, exactly, or even a request -- just me doing a little wishful thinking out loud.

Namely, one of the things I've been looking forward the most to seeing in action is the AI of Limit Theory at levels above that of ship combat. Now that Sean is starting to poke at that a bit, I'm very curious about where NPC "traits" stand today, how traits will guide the behaviors of NPCs in ways visible to human players, and how all of these add up to produce interesting/plausible/fun NPC decision-making beyond dogfighting. I'm curious about dogfighting AI, too; just not quite as curious about that as about things like pattern perception, project creation, task breakdown, effective delegation of tasks (both projects and atomic actions), interpretation of assigned tasks into actions ("I was just told 'go mine,' now I have to decide where and how to mine"), interpersonal effects (Pat works well with Chris but undermines Alex), and how the attitudes of the NPC leaders of a large faction will affect high-level planning.

If a devlog (or Kickstarter update) were to emerge in the next month or so that touched on this subject of NPC traits and on how NPC AI in Limit Theory is intended to work at higher levels of abstraction, I would be very interested in reading it!

(Which reminds me: any value in a Kickstarter update that lets backers know the forum just got a visual upgrade?)
Post

Re: Monday, June 12, 2017

#5
I think that's as close to a perfect update as I've ever seen Josh 3.0 make; brief is good in so many ways.

As others have rightly said, the value of having more than one person work on this is also starting to show up, and for the first time ever, it actually feels to me like LT is making genuine progress.

I would **love** to get a very vague understanding of when Josh 3.0 thinks LT Release might arrive: something like "I think another year will do it" or something like that.
We haven't had anything like that since the KS campaign, and it would be terribly insightful into Josh's viewpoint of whether he actually has a plan now, or is still just working on stuff that needs it (and hence, he won't really *know* when it's completed)...
--
Mind The Gap
Post

Re: Monday, June 12, 2017

#6
Ringu wrote:
Tue Jun 13, 2017 1:24 am
I think that's as close to a perfect update as I've ever seen Josh 3.0 make; brief is good in so many ways.

As others have rightly said, the value of having more than one person work on this is also starting to show up, and for the first time ever, it actually feels to me like LT is making genuine progress.

I would **love** to get a very vague understanding of when Josh 3.0 thinks LT Release might arrive: something like "I think another year will do it" or something like that.
We haven't had anything like that since the KS campaign, and it would be terribly insightful into Josh's viewpoint of whether he actually has a plan now, or is still just working on stuff that needs it (and hence, he won't really *know* when it's completed)...
Well, there is a little known thread I made back when I joined this forum which does shine light on process and what we have and have not seen. Which reminds me, I do need to alter the features list a tad to better represent the official one. I suggest reading the dev logs for an idea of what Josh can do in a month.
Image
Post

Re: Monday, June 12, 2017

#8
BFett wrote:
Tue Jun 13, 2017 1:49 am
Ringu wrote:
Tue Jun 13, 2017 1:24 am
I would **love** to get a very vague understanding of when Josh 3.0 thinks LT Release might arrive: something like "I think another year will do it" or something like that.
We haven't had anything like that since the KS campaign, and it would be terribly insightful into Josh's viewpoint of whether he actually has a plan now, or is still just working on stuff that needs it (and hence, he won't really *know* when it's completed)...
Well, there is a little known thread I made back when I joined this forum which does shine light on process and what we have and have not seen. Which reminds me, I do need to alter the features list a tad to better represent the official one. I suggest reading the dev logs for an idea of what Josh can do in a month.
I remember that thread, BFett, but I could never reconcile it with the dev logs or any JoshReports because the things he worked on were always components at a lower level, so they contributed to the features, but were not enough information to extrapolate from.

Even taking the current form of dev logs, I can't see what Josh thinks is left before any of the features on that list would be considered "Release Quality", or even "Beta Quality".

If Josh currently has a dev plan of components and higher-level pieces that are needed to get to release or even beta, then we could estimate potential completion dates based on the team's velocity or even just guess: if there is no plan, it just means that there's kind-of sort-of a list in Josh's head, and for me that's the last big structural change LT needs to get the project into a state where it's completable. As it is, if Josh is unable to estimate even to the level of "which decade will LT appear in" because there is no plan, the big risk is that we'll end up with a program that has perfect BSP trees but no actual game written.
--
Mind The Gap
Post

Re: Monday, June 12, 2017

#10
we'll be looking at having more milliseconds than we know what to do with :D
Orbital mechanics :)

Or throw it all to more and better AI
Image
When you're trying to fill an infinite multiverse, if you're not willing to consider the entire creative output of humanity as a starting point, you're wasting your time.
User: JoshParnell is accountable for this user's actions.
Post

Re: Monday, June 12, 2017

#12
I'm happy to see a new devlog, but after reading I get sad knowing I have to wait a long time to read another one :)

It is, as others have said, awesome to see LT moving along even when you are deep in fighting forum fires and punching SQL stuff in the face. I have a gut feeling the next devlog will be amazing :D Personally I'd love to see pictures/screenshots of your simulation interfaces, yes even the text versions, this is stuff I'd never get to see otherwise, code we've all seen and the final product we'll all see, but seeing what you guys have are looking at outside the code itself would be cherry on top, even though I might not understand it :)
procedurally generated comment
Post

Re: Monday, June 12, 2017

#15
Miklos wrote:
Wed Jun 14, 2017 6:06 am
I'm happy to see a new devlog, but after reading I get sad knowing I have to wait a long time to read another one :)

It is, as others have said, awesome to see LT moving along even when you are deep in fighting forum fires and punching SQL stuff in the face. I have a gut feeling the next devlog will be amazing :D Personally I'd love to see pictures/screenshots of your simulation interfaces, yes even the text versions, this is stuff I'd never get to see otherwise, code we've all seen and the final product we'll all see, but seeing what you guys have are looking at outside the code itself would be cherry on top, even though I might not understand it :)
Well, I didn't really think anyone would want to see our ugly simulation testing interfaces but...maybe if so, I can include some pictures next time :D
“Whether you think you can, or you think you can't--you're right.” ~ Henry Ford

Online Now

Users browsing this forum: CSE, ubeezee and 1 guest

cron