That's yuge.
Long-distance travel is gonna require some serious planning, and quite a lot of resources.
- Snipehunter
On this shiny day, let’s talk about people. People need a lot of stuff. They need room, they need food (so… much… food…), they need water… and like, for some reason they want stuff like minerals and wood to “build” with. Worse, they just make more stuff with it all!
Nah, I’m just kidding, people are great.
But the point is valid: When you’re trying to build a world for people, especially one that people are meant to virtually live in, you can’t just make a space and make it look it good, you must consider all the stuff the people need. And when you’re talking about a big world, we’re talking about a LOT of stuff.
I probably don’t have to tell you that Elyria is a big world -- I mean, I've said it before -- but even knowing that, I bet that you’ll be surprised when you’re standing in it. We know this world (all 4 worlds) pretty well, but we’re still constantly startled when we put the camera at Mann level and realize that “little valley” we were looking at is actually big enough to fit Los Angeles in it. Earlier today I was complaining that larger versions of the map had taken away the pieces of coastline that I liked in the lower fidelity maps… Bless the art team’s heart for humoring me, too. They just zoomed the map in to show me that not only were they still there, but at full res they were significantly more nuanced. Actually, I happen to have a color map here, let me show you.
(A color map includes no height information, like squashing a relief map)
Do you see that cape there? When I had this color map zoomed all the way out so I could see the whole continent, I was looking at this little nub and thinking “ahh man, did we lose that feature?” but as you can see, zoomed in to a decent resolution, not only is it there, but it’s actually projecting out of a little bay, a detail that wasn’t apparent in the map voting stages. That’s kind of eye opening – especially when you realize that, even in this image, each pixel here actually represents a parcel of land. And this is a tiny slice of the starting continent; there are more for players to find via exploration when the game launches, each as large, or larger. Not to mention myriad smaller islands. I’ve seen the numbers and I felt like I grokked that, but seeing it is a different thing.
This is true about other aspects of the world as well. Each version of Elyria that a server experiences began as a map and a projection, based on the conditions of that map, of where people would live, what resources can be found at each point in the world, and how they would live. An algorithm crawls every meter of the world and analyzes it deeply to determine how many lives it can sustain, based on what resources are there, access to water, the presence of competing elements in the environment, and more. This is used to generate a likely distribution of the population based on conditions. And, once that’s done, the results are fed into yet another algorithm that looks at places to build settlements, determines how large a settlement can be created, and then how sustainable said settlement would be and it begins to allocate the populace to those settlements. Yet another algorithm then looks at each settlement to determine if a sustainable industry can be founded there and does what it needs to do to set it up, allocating settlement resources and people and generating their backgrounds and history based on those results.
To get back to talking about the sheer scope of Elyria for a second: I had an idea of what results this process would create. But that idea was based an idea that Elyria is a lot smaller than it really is. So, the first time I saw a log entry like this while the process was running, my jaw dropped:
I’m telling you, Elyria is BIG! And this is me saying so after already having said it once! It’s like really, really, big! It’s just this big, massive, gigantic, really big thing. Actually, that’s not fair. It’s not a thing; it’s a world. And bringing it to life is probably the most amazing thing I’ve ever participated in. It manages to strike me with awe at every turn, even though I’m one of the ones writing its brain, laying out its bones, and filling its lungs with breath.
Until next time, stay shiny my friends!
That's yuge.
Long-distance travel is gonna require some serious planning, and quite a lot of resources.
Referral Code: 912EC9
Just to clarify: If this is/was the printout from one of the actual four generated worlds, then there would be 231 cities across those ca. 4 Taiga duchies? Or there would be 231 spaces which could potentially support city-sized populations? Or...?
Posted By Snipehunter at 03:56 AM - Fri May 10 2019
(...)That log entry there is old. It's not precisely reflective of the current numbers (...) That said it is in the ballpark(...)
Discord By Caspian at 03:18 AM - Tuesday May 7 2019
(...)one of the problems we've been struggling with (...) Taiga could support well over 500,000 people. So at some point, the numbers, while accurate, need to be adjusted using new heuristics in order to come up with values which work better from a hardware limitation standpoint.
Could you possibly clarify this? Since this shiny is two days after Caspian mentioned the issue, I am puzzled that you seem to imply that while 400k max and 100k actual population are 'old figures', they would be 'roughly in an accurate scale'. With 5 continents and 16 bioms on the initial one, a conservative calculation would be 2 * 8 * 100000 = 1600000 population per server with the figures you've presented for the taiga. Or am I missing something here?
Sage willing to help with Purity (if you spot me on Discord and have some Plague on your account that could be nullified with a trade, drop me a message on Discord)
Elyria is big. You just won't believe how vastly, hugely, mind-bogglingly big it is. I mean, you may think it's a long way down the road to the chemist's, but that's just peanuts to Elyria.
Thank you Douglas Adams :)
Posted By Logain at 07:00 AM - Fri May 10 2019
Posted By Snipehunter at 03:56 AM - Fri May 10 2019
(...)That log entry there is old. It's not precisely reflective of the current numbers (...) That said it is in the ballpark(...)
Discord By Caspian at 03:18 AM - Tuesday May 7 2019
(...)one of the problems we've been struggling with (...) Taiga could support well over 500,000 people. So at some point, the numbers, while accurate, need to be adjusted using new heuristics in order to come up with values which work better from a hardware limitation standpoint.
Could you possibly clarify this? Since this shiny is two days after Caspian mentioned the issue, I am puzzled that you seem to imply that while 400k max and 100k actual population are 'old figures', they would be 'roughly in an accurate scale'. With 5 continents and 16 bioms on the initial one, a conservative calculation would be 2 * 8 * 100000 = 1600000 population per server with the figures you've presented for the taiga. Or am I missing something here?
Just a guess, but if the original combo of resource availability and resource consumption rates of characters yields a 500k player max, but that is too many for the server to handle in the area, you could lower resource availability or increase consumption rates to make the new maximum 50k for example.
well, it is so glorious to see the world come to life bit by bit, can´t wait to log in to the server and explore the world.
Posted By kajoreh at 05:37 AM - Fri May 10 2019
Not being a computer person myself, I ask just for the sake of clarity...Is it the algorithms themselves that are "malfunctioning" or is it the specific data being "inputted" into them that is causing the map generation delays?
Not that I would have any solutions, but it would be helpful to those of us laymen to have something to hang our hats on regarding the why.
It doesn't have to be anything too technical...just a "physical" reason would be enough.
Thanks.
Without diving into too many details, The algorithms we had designed weren't precisely malfunctioning. But rather, with such a large data set, they were producing results that were untenable from a technical perspective. If the algorithms were truly malfunctioning it might have taken less time to address, in fact. But in order to get a viable result set that still worked with the systems of the game, we had to isolate areas in both the algorithms and the game mechanics where changes could be made that were transparent to the play experience while still paring down the results of the algorithms in significant ways.
hope that helps! :)
Posted By Scorus at 06:12 AM - Fri May 10 2019
Just to clarify: If this is/was the printout from one of the actual four generated worlds, then there would be 231 cities across those ca. 4 Taiga duchies? Or there would be 231 spaces which could potentially support city-sized populations? Or...?
Closer to the second. That log snippet actually comes from midway through the process, where the system has analysed the biome to determine how many people it can sustain, then looked at each parcel and its neighbors to figure out the best locations for settlements, then determined the actual population those settlements could support, but those settlement had not yet actually been placed by the time that log snippet was generated. (It's literally the very next step in the process, but that step doesn't generate anything meaningful in the log)
Posted By Logain at 07:00 AM - Fri May 10 2019
Could you possibly clarify this? Since this shiny is two days after Caspian mentioned the issue, I am puzzled that you seem to imply that while 400k max and 100k actual population are 'old figures', they would be 'roughly in an accurate scale'. With 5 continents and 16 bioms on the initial one, a conservative calculation would be 2 * 8 * 100000 = 1600000 population per server with the figures you've presented for the taiga. Or am I missing something here?
The changes we made to the numbers didn't result in changes to the theoretical max populations (That's the population assuming 100% utilization of all biome resources going into sustaining the population) of the regions. In the original version of these algorithms, run against the map datasets, we were getting significantly more sustainable area overall, so the actual populations were far closer to the theoretical populations, which wasn't going to work for us. A result of the changes we made to the way we consider resources in a biome has the effect of driving actual population down, but doesn't change the theoretical max population at all. In fact, overall Kaynadin is pretty spot on with his guess as to the results of our changes.
Hope that helps! :)
Posted By kajoreh at 05:37 AM - Fri May 10 2019
Not being a computer person myself, I ask just for the sake of clarity...Is it the algorithms themselves that are "malfunctioning" or is it the specific data being "inputted" into them that is causing the map generation delays?
Not that I would have any solutions, but it would be helpful to those of us laymen to have something to hang our hats on regarding the why.
It doesn't have to be anything too technical...just a "physical" reason would be enough.
Thanks.
From what members of the team have said, it seems like the actual generation is taking longer than anticipated. Basically it looks like there's so much data that their system is taking a long time to run through all the computations. So it's basically a hardware limitation
Posted By Snipehunter at 12:22 PM - Fri May 10 2019
Posted By kajoreh at 05:37 AM - Fri May 10 2019
Not being a computer person myself, I ask just for the sake of clarity...Is it the algorithms themselves that are "malfunctioning" or is it the specific data being "inputted" into them that is causing the map generation delays?
Not that I would have any solutions, but it would be helpful to those of us laymen to have something to hang our hats on regarding the why.
It doesn't have to be anything too technical...just a "physical" reason would be enough.
Thanks.
Without diving into too many details, The algorithms we had designed weren't precisely malfunctioning. But rather, with such a large data set, they were producing results that were untenable from a technical perspective. If the algorithms were truly malfunctioning it might have taken less time to address, in fact. But in order to get a viable result set that still worked with the systems of the game, we had to isolate areas in both the algorithms and the game mechanics where changes could be made that were transparent to the play experience while still paring down the results of the algorithms in significant ways.
hope that helps! :)
It does.
The product is delayed, because in plumbing terms, the turd was just too big to flush?
And as soon as it fits down the bowl, it will come out the other end.
And our plumbers are hard at work making it fit.
That makes sense to me.
Carry on Mario and Luigi.
Happy Plumbing Guys ;)
We Are The Many... We Are The One... We Are THE WAERD !!!
So in essence the Actual/Max population figures shown indicate that the Taiga is around 27.4% of it's maximum carrying capacity. And you originally had that percentage higher and brought it down.
That comes out to about 2.55 plots (10,450 sqm) per person. Which still seems dense at 95 persons per sqkm. Real Taiga regions look to be around 20 times less dense, but for realistic levels of player interaction some fudging is likely necessary.
Question do these maximum population calculations take into account the different calorie needs (and diets) of the tribes, I have to assume Brudvir eat more then Kypiq?
Also what kind of population densities do you calculate at the opposite end of the spectrum for a high density Biome. We got some general rankings of Biome densities from highest to lowest but just having a strait raw number pop/plot would be much nicer.
Seneschal for the Hrothi County of Iskar, Recruiter for the Duchy of Aritaur
https://discord.gg/qRQ3Zj6
I mean, the numbers look big, but any idea how many in-game days it would take to traverse the entirety of the continent?
One must not always answer every question, one must always question every answer.
What server can carry that huge map and big amount of players? like can we see 200 or 300 players in same location without problem? hope you have plan for that. :D
Posted By Gazii at 7:34 PM - Fri May 10 2019
What server can carry that huge map and big amount of players? like can we see 200 or 300 players in same location without problem? hope you have plan for that. :D
Plans? SBS don't need no stinking plans....
You must all be quick to listen, slow to speak and slow to get angry. JAMES 1:19 NLT
Posted By Gazii at 01:34 AM - Sat May 11 2019
What server can carry that huge map and big amount of players? like can we see 200 or 300 players in same location without problem? hope you have plan for that. :D
It just works!