Return to “General”

Post

Re: THE May, 2014 Devlog Discussion

#2
I am wondering about what sort of populations we will be talking about. If he is aiming for 100v100 as a major space battle, with around 120 in a populous system (or was that just in a single zone in a system which has numerous zones?) If it's per system, I would expect the average to be about 30. To me that seems small for an average system in an advanced space faring society. If it is per zone, then we may be talking good numbers.

At least in my own vision, major systems, hubs, should have a similar appearance to major cities and airports. The world's busiest airports handle about 1200 aircraft a day, if a system has about 10 zones, then the current 120 per zone sounds like a nice bit of traffic to give the system that sense of being a major hub.

So, if it is an average of 30 per system, then in what I suspect is an average neighborhood (all systems within 3 jumps) of about 20 systems, then you have a population of roughly 600.. again that just seems small to me, perhaps that's plenty... But if it is by zone, than it would be a neighborhood of thousands.

thoughts?
Image
Challenging your assumptions is good for your health, good for your business, and good for your future. Stay skeptical but never undervalue the importance of a new and unfamiliar perspective.
Imagination Fertilizer
Beauty may not save the world, but it's the only thing that can
Post

Re: THE May, 2014 Devlog Discussion

#3
ThymineC wrote:This thread seemed logical.
Indeed indeed :clap:
Hyperion wrote:I am wondering about what sort of populations we will be talking about. If he is aiming for 100v100 as a major space battle, with around 120 in a populous system (or was that just in a single zone in a system which has numerous zones?) If it's per system, I would expect the average to be about 30. To me that seems small for an average system in an advanced space faring society. If it is per zone, then we may be talking good numbers.

At least in my own vision, major systems, hubs, should have a similar appearance to major cities and airports. The world's busiest airports handle about 1200 aircraft a day, if a system has about 10 zones, then the current 120 per zone sounds like a nice bit of traffic to give the system that sense of being a major hub.

So, if it is an average of 30 per system, then in what I suspect is an average neighborhood (all systems within 3 jumps) of about 20 systems, then you have a population of roughly 600.. again that just seems small to me, perhaps that's plenty... But if it is by zone, than it would be a neighborhood of thousands.

thoughts?
To be determined, but is there really any point in discussing it? As I said, I will push the numbers as high as they can possibly go...that's really all I can do... :ghost: I am hoping for 100+ in busy zones :D
“Whether you think you can, or you think you can't--you're right.” ~ Henry Ford
Post

Re: THE May, 2014 Devlog Discussion

#5
Hyperion wrote:I am wondering about what sort of populations we will be talking about. If he is aiming for 100v100 as a major space battle, with around 120 in a populous system (or was that just in a single zone in a system which has numerous zones?) If it's per system, I would expect the average to be about 30. To me that seems small for an average system in an advanced space faring society. If it is per zone, then we may be talking good numbers.

At least in my own vision, major systems, hubs, should have a similar appearance to major cities and airports. The world's busiest airports handle about 1200 aircraft a day, if a system has about 10 zones, then the current 120 per zone sounds like a nice bit of traffic to give the system that sense of being a major hub.

So, if it is an average of 30 per system, then in what I suspect is an average neighborhood (all systems within 3 jumps) of about 20 systems, then you have a population of roughly 600.. again that just seems small to me, perhaps that's plenty... But if it is by zone, than it would be a neighborhood of thousands.

thoughts?
As I write in LOD Simulation and historical simulation, I believe that planetary colonies should spawn hauler vessels to carry goods between themselves - in this way, interaction between colonies becomes "real". These hauler NPCs would be extremely low-LOD NPCs who basically do nothing except move from point A to point B. They become higher-LOD only if some already high-LOD agent (such as the player) interacts with them.

CPU overhead for such a system? Should not be great. I propose the use of a shortest-path-finding algorithm like Dijkstra's (whose name I can now spell without Googling - yay, I'm a real computer scientist now) to compute the shortest path between two colonies and make this information accessible to all haulers via something like the entity cache.

As for collision detection? Make the computer imagine hypothetical cylinders with a certain radius that run along the trade routes that haulers use. Divide the cylinders at intervals into small tubes, and make it so that if nothing but haulers and static objects are within a tube, collision detection is turned off for haulers. Haulers should not collide with each other and should steer clear of one another. As a further optimisation, don't make each hauler vessel try to avoid collision with other haulers (an O(n^2) problem) - instead, use another abstracted higher-level algorithm to treat haulers like particles flowing in a liquid-like manner across trade routes. See: Supreme Commander's "Flowfield" technology. Edit: this paper may be useful - I've only skimmed the first page of it, though.

Haulers should make systems feel much more busy and alive, and I would be somewhat disappointed if Josh couldn't get at least hundreds of these haulers operating per system. Ideally, I'd like to see them number in the thousands.

Also, I believe Josh is working on a not-very-beefy computer at the moment. I think a lot of us will be playing Limit Theory on much more powerful computers when the game gets released, so hopefully our machines could support more NPCs per system than Josh's can right now.

Edit: Dijkstra's algorithm solves the single-source shortest path problem. I think something like Netchange would be more appropriate, as this is an excellent algorithm that allows the routing problem to be solved for every node in a distributed system, and even accommodates changes in the topology of the network (imagine if events in the LT game world caused some trade lanes to open or close - the haulers would adapt!)
Last edited by ThymineC on Sun May 04, 2014 3:24 pm, edited 1 time in total.
Post

Re: THE May, 2014 Devlog Discussion

#6
"Logical... flawlessly logical."
ThymineC wrote:I propose the use of a shortest-path-finding algorithm like Dijkstra's (whose name I can now spell without Googling - yay, I'm a real computer scientist now)
You weren't around for the fallout from his "Go To Statement Considered Harmful" essay. I barely was, but enough to engender a healthy disrespect for his pronouncements. :thumbdown:

[edited to eliminate self-modifying link]
Last edited by Flatfingers on Sun May 04, 2014 5:21 pm, edited 1 time in total.
Post

Re: THE May, 2014 Devlog Discussion

#7
This link takes me to some random Amazon page. But I can guess from the URL that it's probably an audio clip of Spock being Spock or something.
Flatfingers wrote:
ThymineC wrote:I propose the use of a shortest-path-finding algorithm like Dijkstra's (whose name I can now spell without Googling - yay, I'm a real computer scientist now)
You weren't around for the fallout from his "Go To Statement Considered Harmful" essay. I barely was, but enough to engender a healthy disrespect for his pronouncements. :thumbdown:
Was this just an essay he wrote against the use of GOTO statements? If so, that's great. GOTO statements suck and structured programming is awesome.
Post

Re: THE May, 2014 Devlog Discussion

#8
ThymineC wrote:I can guess from the URL that it's probably an audio clip of Spock being Spock or something.
It was originally. How peculiar. Now, it's garbage.
ThymineC wrote:Was this just an essay he wrote against the use of GOTO statements? If so, that's great. GOTO statements suck and structured programming is awesome.
As usual, it wasn't the what, it was the how: full of sanctimony about what's best for other people.

This was followed by more of the same: Pascal and Modula-2, languages that are the technological version of the infantilizing attitude of "no, you can't have that power, you might hurt yourself."

I'm not a fan of spaghetti code; I've had to debug other people's stuff and it's no fun. In fact, I'm in the middle of rewriting 30-year-old PL/I code that's nothing but GOTOs. I'm also OK with extremely minimal starter languages.

But beyond that, if I have to choose between freedom to screw up and someone else choosing for me that I need to be protected from myself, give me GOTOs.

Yeah, not much of a fan of Dijkstra or Wirth. ;)
Post

Re: THE May, 2014 Devlog Discussion

#9
JoshParnell wrote:Tearin' it up! :) I don't actually want to discuss every optimization that I made today in detail, because it's really just a boatload of technical stuff. I could write for pages and pages about it :D So I'm going to go ahead and not do that. Suffice it to say, there was plenty of profiling, coding, ripping, re-profiling, tweaking, slashing, and all sorts of other fun!

Current Optimization Status: 256+ ships in one system running smoothly (60 fps in calm moments, 30 fps in full-out, system-wide battle).
Having the boatload of technical stuff available to examine is desirable.

Consider making a 2 minute video of the mass ship engagement, those are very good pr.
woops, my bad, everything & anything actually means specific and conformed
Post

Re: THE May, 2014 Devlog Discussion

#11
Katawa wrote:
JoshParnell wrote:Tearin' it up! :) I don't actually want to discuss every optimization that I made today in detail, because it's really just a boatload of technical stuff. I could write for pages and pages about it :D So I'm going to go ahead and not do that. Suffice it to say, there was plenty of profiling, coding, ripping, re-profiling, tweaking, slashing, and all sorts of other fun!

Current Optimization Status: 256+ ships in one system running smoothly (60 fps in calm moments, 30 fps in full-out, system-wide battle).
Having the boatload of technical stuff available to examine is desirable.

Consider making a 2 minute video of the mass ship engagement, those are very good pr.
I also support divulging the technical details, but then again I am a programmer. Can you give us even a little bit of detail?
Image

Online Now

Users browsing this forum: No registered users and 16 guests

cron