|
Yeah that's brilliant doing your coding in the livestream and getting live help and feedback.
|
#
?
Aug 6, 2014 00:36
|
|
|
|
| # ? Apr 20, 2024 02:44 |
|
|
Development and Q&A Livestream today 4PM GMT, 9AM PDT, 12PM EDT As promised, today will be the first of the regular livestreams. Tune in at http://www.twitch.tv/ae_play in about 3 hours (as of this post) --- Iunnrais posted:I really like the merging lanes. Before, it was really jarring, now it looks awesome! Progress! Not yet, but soon. Maybe I will even implement that in one of the next livestreams!
|
|
#
?
Aug 7, 2014 13:57
|
|
|
I watched a decent chunk of the livestream after it was recorded, and I'm worried about the way you're implementing deliveries. It seems to be very closely tied to the concept of cars, and I have a feeling you'll be completely rewriting it later. Here's how I would do it: First, the business that needs the good would request a shipment from the nearest source that has available goods to sell. Then, the source would find a route to the destination via all available transit networks. Lastly, the good would be sent to the destination. This accomplishes a few things. - Goods are only sent to places that need them (a business can fail if there is no demand for its products). - The goods themselves are what is spawned by the shipper, meaning they can take any transit path to their destination - car, train, boat, it doesn't matter. Even taking more than one method of transportation, if needed. - You can easily define goods that have limitations on how they can be shipped - some goods might be too heavy to ship by truck, for example. - You can have factories that unload directly to a train, like this: https://www.google.com/maps/@42.9720729,-78.9256099,206m/data=!3m1!1e3 https://www.google.com/maps/@42.9785925,-78.9250584,240m/data=!3m1!1e3 https://www.google.com/maps/@42.9749998,-78.9304571,514a,20y,180h,41.5t/data=!3m1!1e3 KillHour fucked around with this message at 05:20 on Aug 8, 2014 |
|
#
?
Aug 8, 2014 05:17
|
|
|
Yes, of course I will have to rewrite it. I'm pretty sure it will end up being a generic transportation system almost exactly like you describe. But just getting the drat cars working is hard enough for now! Again, my mantra: implement the most stupid version first, make it smarter later - or I will go crazy.
|
|
#
?
Aug 8, 2014 12:41
|
|
|
Reminder: Tomorrow at 15:00 CET I will do another livestream on http://twitch.tv/ae_play
|
|
#
?
Aug 8, 2014 19:55
|
|
|
I'd love to see different sort of cargos handled differently. Some industries NEED rail access to function, no way around it. Some industries need water transport to function. Some industries and businesses only need tiny delivery trucks or even cargo bikes could work.
|
|
#
?
Aug 8, 2014 20:09
|
|
|
anselm_eickhoff posted:But just getting the drat cars working is hard enough for now! I say this as a software developer who's been doing it long enough that I feel old when I think about when I started: you're doing exactly the right thing, and never let anyone convince you otherwise. (Just don't forget that it's always fine to throw out code that's outlived its usefulness.)
|
|
#
?
Aug 9, 2014 05:23
|
|
|
Avenging Dentist posted:I say this as a software developer who's been doing it long enough that I feel old when I think about when I started: you're doing exactly the right thing, and never let anyone convince you otherwise. (Just don't forget that it's always fine to throw out code that's outlived its usefulness.) To each his own. When I'm doing a big project, rewriting the same code over and over is what kills it for me. That's the biggest reason I don't code any more.
|
|
#
?
Aug 9, 2014 06:01
|
|
|
KillHour posted:To each his own. When I'm doing a big project, rewriting the same code over and over is what kills it for me. That's the biggest reason I don't code any more. The real problem is that you're often not 100% sure of the requirements when you write the first version of a thing, so you'll end up rewriting it no matter what. Might as well make the first version as simple as possible so you don't waste your time. Avenging Dentist fucked around with this message at 06:28 on Aug 9, 2014 |
|
#
?
Aug 9, 2014 06:22
|
|
|
I'd really love to build my hometown in this game and see how it stacks up against the real thing with regards to traffic and stuff. And to also reenvision my hometown as a New Urbanist type city.
|
|
#
?
Aug 9, 2014 18:15
|
|
|
After watching your last stream, I thought a little bit about the current problem. Here's an example of a complex road near where I live:  With a system like this, it's possible that your turn could come up faster than you could merge if you're breaking it up based on stretches of road without lane changes. It might make more sense for the pathfinder to pick a route and then "flood fill" all adjacent lanes that are part of the same road as acceptable for the car to be in. You could have this visualized for debugging by painting allowable end points on the lane segments red or green, depending on whether the car is allowed to go through that point. You would need individual car AI, but it would only be a few lines of code. Edit: Thought about it a little more, and this is probably simpler. Say you have this complicated section of lanes, all going in the same direction:  You need to get a car from the purple dot to the green dot. The first step is to mark all the segment ends you can't go through in red (This is possibly non trivial):  Next, let's say it takes 50M to merge a lane. Draw backwards 50 meters from each red segment:  Next, you have to iterate this to test for segments where two red lines start at the same place. If this happens, you need to extend the "outside" one 50 meters (again, may not be trivial to figure out which one this is):  Now, you have to know which lanes not to merge into. Extend your red lines 50M with a yellow line.  This brings up another test-case - you don't want to extend the top right red line with a yellow line, because it's at a fork in the road. Forks should probably be treated as a separate case and not be included in this algorithm. You're home free at this point. Your car will always end up at the destination regardless of how it merges as long as you follow two simple rules: 1: ALWAYS merge out of red lanes into a yellow or blue lane. 2: NEVER merge from a blue lane into a yellow or red lane. KillHour fucked around with this message at 04:52 on Aug 10, 2014 |
|
#
?
Aug 10, 2014 04:16
|
|
|
KillHour posted:but it would only be a few lines of code. Famous last words.
|
|
#
?
Aug 10, 2014 04:39
|
|
|
Tres Burritos posted:Famous last words. Well, I fixed it.  (Ignore the non-trivial parts) Edit: And I even figured out the non-trivial parts! (And they said it was a mistake to drop out of my computer science degree. Sure showed them!  )You just start at the green dot and work backwards in steps. Here are some pictures. Start red this time:  50M later:  100M:  150M:  200M:  You get the idea. This also neatly handles the forking road case and you can save this with the path info, so you only need to calculate it when you're recalculating paths. Double edit: I can't think of any edge cases where this wouldn't work, but someone is free to prove me wrong. Looping cases like a roundabout can be tested for and handled pretty easily. KillHour fucked around with this message at 05:13 on Aug 10, 2014 |
|
#
?
Aug 10, 2014 04:44
|
|
|
KillHour posted:It might make more sense for the pathfinder to pick a route and then "flood fill" all adjacent lanes that are part of the same road as acceptable for the car to be in. "flood fill" is a really nice way to put it. I was stuck at the problem of which initial path to pick before discovering all the other acceptable lanes, but "flood fill" made it click! It doesn't matter which initial path you pick, if you flood fill, you will always get the same set of adjacent lanes! I also really like that all of this info can just be encoded in the path information and cached temporarily, like the path itself, it doesn't have to be encoded with the lanes itself and meticolously updated when the road layout changes, (I can just invalidate cached path information, just like before) Thanks a lot for your thoughts! The edge case of a loop won't happen because the shortest paths returned by the pathfinder of course never contain loops!
|
|
#
?
Aug 10, 2014 23:27
|
|
|
New update! The Road to Alpha, Week 23 - With a Little Help
|
|
#
?
Aug 13, 2014 01:58
|
|
|
Just had a thought, how 'political' is the game going to be? Are we going to be able to create our socialist commune/fascist police state or are we just city planners?
|
|
#
?
Aug 13, 2014 02:01
|
|
|
Praseodymi posted:Just had a thought, how 'political' is the game going to be? Are we going to be able to create our socialist commune/fascist police state or are we just city planners? Probably not much. There will be ordinances and similar tools that allow a little local politcal control, but nothing more. Of course mods can make the craziest stuff possible.
|
|
#
?
Aug 13, 2014 02:28
|
|
|
Anselm, have you been following Train Fever at all? Obviously a very different sort of game, but they seem to have simple procedural buildings that fit angles fairly nicely. They're ugly and simple, but it works. https://www.youtube.com/watch?v=qwrE8k0NRjo
|
|
#
?
Aug 13, 2014 05:50
|
|
|
Baronjutter posted:Anselm, have you been following Train Fever at all? Obviously a very different sort of game, but they seem to have simple procedural buildings that fit angles fairly nicely. They're ugly and simple, but it works. Can't say I've been following it. Looking at their procedural buildings reminds me that even simple solutions can look pretty good and that I shouldn't overthink everything so much 
|
|
#
?
Aug 13, 2014 13:43
|
|
|
anselm_eickhoff posted:Can't say I've been following it. Looking at their procedural buildings reminds me that even simple solutions can look pretty good and that I shouldn't overthink everything so much Well clearly you don't have a case of train fever, but hopefully you have procedural buildings fever.
|
|
#
?
Aug 13, 2014 15:56
|
|
|
Train Fever looks hella cool. Do we have a thread?
|
|
#
?
Aug 13, 2014 17:42
|
|
|
Vahakyla posted:Train Fever looks hella cool. Do we have a thread? I was thinking of making one but for a game ready for pre-order coming out in a month there's very little media or info on it. Other than a german language beta feature video and a couple fairly uninformative and dated trailers there's really not much to go on other than "germans sure love logistics and train based games". I was tempted to pre-order but I'd need an good solid chunk of gameplay in english to watch first. Also no one has ever shown how signals work and they "promise" signals will be in for launch. I'm sure there will be a thread once it's out. But it's a small team that started with a kickstarter or something, so shows that an tiny indie team can make a complex transport sim with procedural cities.
|
|
#
?
Aug 13, 2014 17:49
|
|
|
Baronjutter posted:I was thinking of making one but for a game ready for pre-order coming out in a month there's very little media or info on it. Other than a german language beta feature video and a couple fairly uninformative and dated trailers there's really not much to go on other than "germans sure love logistics and train based games". I was tempted to pre-order but I'd need an good solid chunk of gameplay in english to watch first. Also no one has ever shown how signals work and they "promise" signals will be in for launch. I'm sure there will be a thread once it's out. But it's a small team that started with a kickstarter or something, so shows that an tiny indie team can make a complex transport sim with procedural cities. I hope I can make ridiculously large and complicated train networks like in TTD and it's not abstracted where trains can pass through eachother like in some other games.
|
|
#
?
Aug 13, 2014 18:02
|
|
|
Baronjutter posted:Well clearly you don't have a case of train fever. After watching that German 50min review video, I have to say I'm quite impressed! I already know a couple things that I will blatantly steal!
|
|
#
?
Aug 13, 2014 21:13
|
|
|
e: oops I am a dumb
|
|
#
?
Aug 13, 2014 21:30
|
|
|
Livestream tomorrow at 4PM UTC!
|
|
#
?
Aug 13, 2014 22:38
|
|
|
anselm_eickhoff posted:After watching that German 50min review video, I have to say I'm quite impressed! I'm curious what aspects impressed and "inspired" you!
|
|
#
?
Aug 13, 2014 23:07
|
|
|
Baronjutter posted:I'm curious what aspects impressed and "inspired" you! - Road Building AI that autoconnects existing roads and creates new ones (won't use that directly but might be useful) - Preview of how terrain will be carved by a new road/rail segment - Display of all resources currently being transported (with icons overlaid on the transporting vehicle) - Procedural terrain - Calculation of shopping/workplace/leisure coverage and distances for individual houses
|
|
#
?
Aug 13, 2014 23:43
|
|
|
Baronjutter posted:I was thinking of making one but for a game ready for pre-order coming out in a month there's very little media or info on it. Other than a german language beta feature video and a couple fairly uninformative and dated trailers there's really not much to go on other than "germans sure love logistics and train based games". I was tempted to pre-order but I'd need an good solid chunk of gameplay in english to watch first. Also no one has ever shown how signals work and they "promise" signals will be in for launch. I'm sure there will be a thread once it's out. But it's a small team that started with a kickstarter or something, so shows that an tiny indie team can make a complex transport sim with procedural cities.
|
|
#
?
Aug 14, 2014 00:21
|
|
|
https://www.youtube.com/watch?v=lxfeBpagvQw&hd=1 Speaking of small time devs and cities, looks like the Cities In Motion crew are making a simcity game too!! Of course this is alpha, but I don't think the lack of procedural buildings is going to change, we'll be stuck with fitting pre-modeled repeating squares into curves and triangles. There HAS to be transit since this is a company that so far has only made transit games. I always felt CiM2 felt unfinished and rushed, like an alpha version of a city building game but they only (barely) finished the transit portion. That said, if they can polish up the ugly spaces between buildings with props and fills it could be a pretty ok game. I loved CiM1 but couldn't get into CiM2, I'm rooting for them but I don't have high hopes. Here's their official feature list. I had no idea these people were working on this! Main features: City policies: Set policies to guide how the city and districts develop over the course of your playthrough. City districts: Personalize city districts with names of your choice for variety and personality. Road building and zoning Unlock buildings and services Taxation: Fine-tuning the city budget and services and setting tax rates to different residential, commercial and industrial levels and controlling what kind of areas are more likely to spawn in the zoned areas Public transportation: Build transport networks throughout the city with buses and metros Outside connections: Make industry and commercial districts flourish with new customers in the neighboring cities Wonders: the ultimate end-game content that the players strive towards Huge maps: Unlock new map tiles with unique possibilities to expand the city Water flow simulation: Add new challenges to water services. Polished visual style and core gameplay Modding tools: Built in feature designed to encourage creative pursuits. I'm curious about districts, I bet they will use the transit zone tools from CiM2 to allow us to define districts and even set different policies within the districts. Baronjutter fucked around with this message at 00:11 on Aug 15, 2014 |
|
#
?
Aug 14, 2014 23:56
|
|
|
 Haha, they know their audience. I wish that they had put a little more effort on making the text look good though.
|
|
#
?
Aug 15, 2014 00:17
|
|
|
Baronjutter posted:https://www.youtube.com/watch?v=lxfeBpagvQw&hd=1 At 0:42 they show the six standard low/high density residential/commercial/industrial options. Would be nice to get something new in that area.
|
|
#
?
Aug 15, 2014 03:09
|
|
|
Those cities in the alpha footage look absolutely gigantic when compared to SimCity. Paradox better have a good system for their transit/roads in the game. I mean, they have made the CiM games which were heavily based on transit systems.
|
|
#
?
Aug 15, 2014 03:27
|
|
|
Baronjutter posted:https://www.youtube.com/watch?v=lxfeBpagvQw&hd=1 Myself, I thought it was a fairly good system (not as flexible as we're seeing being developed in Citybound), aside from the annoying thing where cars would leave an expressway at an off-ramp and then go straight back on it. Besides I kept getting annoyed with the city layouts and trying to improve things once the transit system was earning enough. They have to tweak that a bit but they can concentrate most of the effort on all the policy and development stuff. Anyway, we'll see how it goes. The failure of SimCity 2013 as a decent city builder game seems to have sparked a lot of interest in people trying to do it better. IAmTheRad posted:Those cities in the alpha footage look absolutely gigantic when compared to SimCity.
|
|
#
?
Aug 15, 2014 04:21
|
|
|
Yeah I was pretty bummed to see only 3 zones and 2 densities. There's such a huge difference between detached houses, 4-6 story apartments, and 80 story skyscrapers. Even in simcity4 I always found I wished they had one more density level. Detatched suburban style houses Row/town houses and 3-4 story apartments mid-rise buildings High-rises. The jump from light to medium seemed a big bit. Of course I'd so just love to directly control the FSR and heights via sliders or inputs along with a few pre-sets. Maybe I want detached houses but I want them to be fairly dense with minimal yards. Maybe I want to zone an area for acreages. Maybe I want medium density apartments but with large green-space buffers around them, maybe I want medium density apartments but arranged in a more urban wall to wall style. With procedural buildings it seems like it would be pretty simple to implement. Baronjutter fucked around with this message at 00:17 on Aug 16, 2014 |
|
#
?
Aug 15, 2014 23:52
|
|
|
To quote myself from reddit regarding Cities: Skylines "Yes, competition! This means I can blatantly steal all the things they do right, not do the things they do wrong and add my own. It's like free research and development!" Also, there will be another livestream today, again at 4PM UTC on Twitch.tv
|
|
#
?
Aug 16, 2014 12:55
|
|
|
New Update: The Road to Alpha, Week 24 - Signs of Growth Let me know what you think!
|
|
#
?
Aug 20, 2014 00:24
|
|
|
Ohhh nice. Is the likelihood that a lot upgrading to higher density going to be tied to any kind of land value/desirability mapping? (In the long term, anyway, if and when you put such systems in.)
|
|
#
?
Aug 20, 2014 00:45
|
|
|
The Deadly Hume posted:Ohhh nice. Is the likelihood that a lot upgrading to higher density going to be tied to any kind of land value/desirability mapping? (In the long term, anyway, if and when you put such systems in.) This was going to be my question also! I would think that single home residential would want to balance having empty lots (especially wooded or water empty lots!) nearby with also having amenities like shopping nearby, while also having a decent commute. Then the houses would pop up more organically, not randomly. For denser residential, I think it would be the reverse of wanting to build away from each other... they'd want to keep apartments grouped together instead of spread apart, so that they can build wall-to-wall where possible, and so that you don't get nail houses stuck between big apartments. I'm sure you're planning to have wall-to-wall buildings show up. Do you have any code in place working towards this already, or will wall-to-wall be a separate coding project for later? Asking only because I'm interested in the coding process, whether you're laying groundwork or just getting there when you get there. Alas, I'm at work when you do your livestreams, so I can't join then... Iunnrais fucked around with this message at 01:16 on Aug 20, 2014 |
|
#
?
Aug 20, 2014 01:12
|
|
|
|
| # ? Apr 20, 2024 02:44 |
|
|
The Deadly Hume posted:Ohhh nice. Is the likelihood that a lot upgrading to higher density going to be tied to any kind of land value/desirability mapping? (In the long term, anyway, if and when you put such systems in.) Yes! Iunnrais posted:I'm sure you're planning to have wall-to-wall buildings show up. Do you have any code in place working towards this already, or will wall-to-wall be a separate coding project for later? Asking only because I'm interested in the coding process, whether you're laying groundwork or just getting there when you get there. Alas, I'm at work when you do your livestreams, so I can't join then... The problem is that right now, the lot generation system isn't very good at creating lots that are exactly next to each other. A big overhaul there is planned however, which should make wall-to-wall and non-rectangular lot shapes possible. --- In other news: Livestream today at 3PM UTC
|
|
#
?
Aug 21, 2014 12:36
|
|