Jake (theFlying3.14) here, Lead of Tool Development here at Nerd Kingdom. Several powerful systems have begun to come online in the Eternus engine recently. To support these systems we’ve designed several tool prototypes to aid designers in creating content. Today I’d like to share one of the more important systems that are being reused in multiple instances to provide a comprehensive functional experience going forward: the Visual Node Programing platform, or VNP.
VNP is a node programming platform that allows users to script functionality across different aspects of the game. The system is already being used in a few early tool prototypes: the biome tool, the animation web, and an AI behavior scripter. Future tools such as the material editor, shader creator, and quest editor are planned for VNP implementations.
Developed from the MIT licensed ThreeNodes.js – a WebGL shader tool – we heavily reworked the basic data structures and assumptions built into the library. Although there is still a lot we would like to do with it, what we’ve ended up with gives us great scalability.
The Visual Node Programming platform exists as an abstract application that we employ within each tool implementation, customizing it to fit the context. This means when you open biome tool, you will be greeted with a similar experience as the animation web. However, in reality, each tool might need to operate slightly differently. For example, the biome system reads the node graph from right to left, whereas the animation system reads “state strings” from left to right. To accommodate this each implementation of VNP has its own override of several fundamental objects: nodes, connections, workspaces. This allows great flexibility when developing and updating tools developed with VNP.
“So great another node programming tool….”
Over the past several months we have gotten to experiment with a few different approaches to VNP integration. The first approach we took was to build the node graph, save the data models needed specifically for the node graph (like node.x and node.y, etc), and then grab just the data we needed for the engine resource, and send it in one big packet. Of course, this worked until we started building big graphs. Once the save packet got too big to pass between the frontend to the backend we smartened up.
The animweb tool took a different approach: each time a node is connected to the graph, the system evaluates where it is and dynamically adds it to the resource. This resulted in live coding. Being able to edit a resource’s node graph and see it change immediately. It also resulted in a lot of edge cases that are still giving me nightmares. For example, deleting nodes or removing one connection from a node that’s still connected to another field become really tedious.
Our overall goal for user-facing tools is to create simple interfaces that developers at any skill level will be able to leverage. VNP provides a familiar interface for designers as similar platforms are used in engines like Unreal and Unity. While programming with nodes can easier than scripting, this is not our final destination. We decided to tackle VNP first to provide us with a clear functional foundation of what designers need. Since nodal programming lends itself to so many situations, we can provide a consistent feeling experience across the game development workflow. Then later we can develop more specialized tools to streamline certain common practices and make it easier for less experienced devs.
I hope you enjoyed this look at our Visual Node Programming platform, and I’m excited to get our tool suite ready for feedback from our awesome community.
Happy Friday everyone!
I apologize for the delayed update as I was out with the flu last week. It is good to feel alive again! Getting back on topic, our engine team is cranking away on the Eternus engine and tools. Our gameplay team is improving our building methods, islands and more. Art team has been hard at work creating new assets for TUG. You can some images and video examples below.
Meanwhile, @x_nekochu_x is hiding away in the sound booth working on sound production stuff. I have been nagging him a lot to let me help out with sound effects. There’s just something extremely fun about making sound effects. For example, breaking things… many things.
We also updated our website and relocated our blog to the main site. You may find a few older posts with broken links and images but there is not much we can do to fix them.
The community questions have been finalized and we hope to have more details in the future. If you missed out, don’t worry! We can do another in the near future. We apologize for some of the TBD answers as those are still in discussion.
Hey all! This is Northman from Nerd Kingdom here to share some AI bytes with you. Specifically I’d like to talk about Pathfinding and how it relates to TUG and our characters.
Pathfinding is the act of finding the best path between two points in the world. The key word there is “best”. The definition of best depends on the type of game you are making and the type of character you are trying to find a path for. I think a small thought experiment helps to clarify the set of problems pathfinding tries to solve:
Imagine a mountain goat and a human facing a mountain that extends as far to their left and right as their eyes can see. Directly in front of them is a door. On the door is a sign that reads: “Tunnel to the Other Side”. The human does not have any climbing gear and the mountain is far too steep to for the human to scale it without proper equipment. If both the human and the goat want to get to the other side of the mountain what do they do? The goat does not have hands to open doors nor the ability to read. However, the goat is a sure-footed climber and does not have any problem scaling the mountain so it goes on its goaty way over the top of the mountain. Conversely, the human does have hands and can read so they take the tunnel. The path the goat and the human found are both the best path they can muster by their definition of best even if they are both very different.
By Darklich14 (Own work) [CC BY 3.0 (http://creativecommons.org/licenses/by/3.0)], via Wikimedia Commons
*Insert funny remark about goats here.*
“Ropes? We don’t need no stinkin’ ropes!”
So now that we have defined Pathfinding and talked about what “best” means we can look at what tools we have for finding paths. Most pathfinding techniques break up the world into spatial subsections (nodes) and store information about how those nodes are connected (edges). In Computer Science we call a set of nodes and edges a “graph”. Graphs are cool because they have been studied by mathematicians since the 18th century (check out the Seven Bridges of Königsberg). What this means to us is there are well known techniques, also known as algorithms, for dealing with graphs and finding best paths on them (see Pathfinding on Wikipedia). One of the most common algorithms used in games for pathfinding is A*. I won’t get into the details of A* in this post because it gets technical very quickly but the image below provides a good visual representation of a typical A* search.
By Subh83 (Own work) [CC BY 3.0 (http://creativecommons.org/licenses/by/3.0)], via Wikimedia Commons
An illustration of an A* search. The initial red node indicates the starting position and the green node indicates the goal position. The gray polygon represents an obstacle. The goal is to find the shortest distance path from the starting node to the goal node while avoiding the obstacle.The path found is highlighted in green at the end of the animation.
We are currently exploring algorithms and graph representations for our world in TUG but so far we have implemented a navigational grid. In graph terms the grid is made up of nodes that all represent the same amount of space (one square meter) and each node has edges to its immediate neighboring nodes (grid cells): top left, top, top right, right, bottom right, bottom, bottom left, and left. These cells can be blocked by obstacles (shown in the image below in red) or open (shown in the image below in blue). This allows us to run A* searches to find best paths for our characters that avoid obstacles.
Blue Cells: Areas a character can navigate in.
Red Cells: Areas blocked by an object.
I hope you have enjoyed this introduction to pathfinding. Pathfinding is a large topic with many different techniques available depending on the pathfinding problem at hand. If you have any questions please feel free to email me at “northman at nerdkingdom dot com”. We are working hard to refine our pathfinding approaches for our characters and look forward to sharing more with you soon!
Have a great weekend!
Hey everyone! I hope everyone is having splendid week.
We are still hammering away on the engine and tools so there are not many visuals to show yet. However, we do want to keep you all updated with more goods so I will be sharing videos from our tech demo Fridays. Until then, here are a few WIP shots of some of our new tools.
[Biome Tool Graph]
[Biome Generated #1]
[Biome Generated #2]
[Early rough video of our floating island with sample of edges]
[8 blend target]
In a previous blog, we answered many of the planning and lore questions from the community. Today, we’ll focus on the TUGv2 and engine remake questions.
The Engine Remake
Have a great weekend!
Happy Friday! Cambo here to kick the weekend off by bringing back the dev tech blogs. We plan to keep them coming at least once a month as we make more progress on development. In regards to the Q&A, most of them are answered and I’ll be sure to post it in our next update blog. For now, here’s our infrastructure dude, Maylyon!
Hey everybody! Maylyon here with a new non-game related, non-engine related, non-tools related tech blog! Hint: this is your tune-out point if those are the topics you are looking for. 3 … 2 … 1 … Still here? Excellent!
After literally years of silence about Devotus, I wanted to follow-up with a snapshot of where Devotus is today. If your memory is a little rusty, Devotus will be our mod content distribution pipeline to help mod authors create and manage their home-brewed content and deliver it to end-users. To get context for this blog entry, you should definitely read those first two blogs. Without further ado, the “what has been happening?” (aka: “you guys still work on that thing?”).
In case you didn’t know, there were mods on Devotus’s developmental servers from a TUG v1 ModJam in early 2016. Don’t go rushing to find them now; they’re gone. They were sacrificed to the binary gods in order to make way for…
Suspend your understanding that the term “serverless” is a lie because there are always servers somewhere and play along for a bit. The old Devotus architecture was built on AWS EBS-backed EC2 instances running a mix of Node.js, C++, and MongoDB. It looked a little bit like this:
The primary detriments to this approach were:
1. Paying for these servers (even extremely small servers) when nobody was using them,
2. Scalability at each layer of the stack would incur even more financial cost and contribute to…
3. Complexity of the implementation.
Moving to this setup allows us to:
1. Greatly reduce the costs associated with Devotus (especially when nobody is using it),
2. Offload most of the scalability problem to AWS (less work = more naps),
3. Synergize our implementation with the other microservices we have been developing on the Infrastructure team.
Devotus now allows mod authors to create git repositories on GitLab in addition to GitHub. It’s actually been there for a while but wasn’t there in the last blog I wrote. By supporting GitLab and their awesome pricing model, Devotus allows a mod author to choose whether they want their mod’s git repository to be public or private at mod creation time. This choice does not apply to mod’s created on GitHub because their pricing model is less awesome (but still pretty awesome) and I’m cheap (see previous section for proof).
In the “bad old” days (read as “a month ago”), mod download count was just an unsigned integer. Download request comes in, number gets incremented by one. Commence spamming download of your own mod to falsely inflate its popularity! Everybody wins! … Except for the people who want to use the system.
Now, in the “brave new world” days, mod downloads are tracked per-user, per-version. This allows mod authors to track their mod’s popularity throughout its release history and allows end-users to trust that a mod’s popularity is probably because of an amazing mod author rather than a mod author’s amazing spam-bot.
That’s all I have for this installment. I (or somebody from my team) will be back with future Infrastructure updates as we get new and/or exciting things to share. In the meantime, be sure to jot down all those cool mod ideas you have kicking around in your brain into a little leather-bound notebook so that WHEN TUG v2 is launched and WHEN Devotus is client-facing, you will be ready!
Have a great weekend!
Hey everybody! Cambo here and Happy Friday!
I’ll go ahead and update everyone since Ino is lost in the world of Zelda today. The team has been very busy the past month working on our engine tools and gameplay features, and core experiences. I will personally be working on having more frequent blog posts, dev updates, and videos when they are available. I am aiming for at least twice a month and video snippets from our demo Fridays. Importantly, I hope to shift our blog posts to our website in the future.
In the last blog, we mentioned that we would take note of the QA document that was compiled from the community. I got to work and nagged the team leads to answer them, nagging is what I do best. You can find the questions here. Thanks for organizing and bringing this to our attention @Rawr! The leads answered most of the questions but we will wrap up the rest over the next couple blog posts as we have more concrete answers. Get ready for the wall of text!
Radial Menu UI prototype
Have a great weekend and don’t forget to follow the leads!
Engine Programmer Lead @JoshuaBrookover
Gameplay Lead @Scriptslol
To get the latest updates on Nerd Kingdom tech sent right to your e-mail, fill out the form below