Nick Funk

I started a new side project building a network stack for Unity. Then I thought, networking is boring if you don't showcase it doing something.

So what's hard to do in networking? TOWER DEFENSE!

Why tower defense? It has a lot of agents (monsters) that move around constantly and since players control the mazes they build, the movement isn't predictable. So... a lot of chaotic agents going all over the place, basically worst case scenario for predictively optimizing your network state.

So I set out to see if I could quickly prototype a tower defense game with the network stack I built.

It turned out a lot better than I expected. I'm seeing 100's of agents streaming to 3 clients and I'm using ~75 KB/s.

A lot of little networking tips helped, and I had to reference Gaffer On Games a plenty, but the results are great for what little time I've put into this.

I still have some more work to do on delta compression, agent culling, and I want to improve the movement interpolation some more, but all-in-all, this has been a great little project!

So now I have my own network stack, and it's compatible with Unity. More technical specific details to come, but this is a good start. Maybe this will end up on the asset store at some point? crosses fingers

To be productive, we need to stay focused on working. We also need to communicate and stay up to date with our team.

Your daily stand-up or scrum is an obstruction to that flow of getting work done. To prevent disrupting flow any more than needed, we try and minimize the time needed during this ceremony.

What if your scrum or stand-up is running long. Way longer than the usual 10 - 15 minutes. What's up?

There are two things that usually cause this from my experience:

• Not staying on topic in the meeting (i.e. "getting into the weeds")
• Your scrum has too many people

Not Staying on Topic

This is a matter of discipline. Your update should be simple.

Let's say Kim is currently building a login page for a website.

Her contribution could be:

Moving forward with building the login page. I'm not sure if we want to support other login types than just e-mail for now. Sarah, can we meet up afterwards to discuss?

Sarah would reply:

Yep, let's talk after this, we'll get the details hammered out.

Then we would move on to the next person in the scrum call. Short and to the point.

What if you have some suggestions when someone states an open ended problem?

Simple, use these words:

I have some ideas around that, let's discuss it after the meeting.

And then you move on.

Scrum is not for problem solving! It is for reporting in and starting other asynchronous discussions so everyone can get back to working.

Any time you're sitting there hashing out how to solve a problem when 3 or more other people are just sitting there, you're likely wasting a lot of time.

Finally, it's up to your scrum lead to call people out on this. Just gently remind folks with the following:

This sounds like it's getting a bit detailed. Would it be possible to discuss this after the meeting?

Your Scrum Has Too Many People

This is one that subtly creeps in as we try and include all the different roles on a team and keep everyone up to date.

It usually starts with adding your quality assurance point man to the call. Then you add in your managers. Finally, you add in more stakeholders. Then some other developers seem related to this project, so we add them too.

I'm here to tell you, this is a bad idea. Keep it simple. Try and keep it down to the following people:

• Scrum leader
• Developers working on the specific product
• Quality assurance point person
• A product owner / stakeholder

Ideally, this list will add up to about 5 people. Any larger and you're getting into unwieldy territory. It's the same reason why you don't play board games with more than about 6 people, it takes too long.

More importantly, if you have a massive scrum, it's a communication "smell" that maybe your team members aren't communicating effectively outside of scrum time.

You'll notice this is happening when you start hearing the following phrases:

Wait 'til scrum, we'll bring it up then when everyone is there.

Let the product owner know that's missing from the requirements at the next scrum. They'll know how to deal with it.

Do you see the problem here? We're waiting until an established ceremony to bring up issues. Scrum becomes your dumping point for any and all ideas, problems, and troubleshooting.

Because of this, you start adding any and all people who need to be there to hear these details so they're in the know and can respond.

You end up bottle-necking the majority of the communication to within the scrum. It's inefficient and it promotes siloing within your organisation.

How do you combat this? It's more of the scrum lead promoting the good behaviour.

• Keep telling folks to keep things focused and break out into communication outside of scrum.
• Explain to your teams that scrum is just a starting point for communication. They need to be reaching out to communicate outside of scrum more than they do within scrum.

That second point is critical. The most effective teams are those that keep each other in the loop by themselves. It means strong communicative skills that are not tied around a single meeting.

Conclusion

Keep your meetings concise and break any extraneous discussion out into new discussions if you want to keep your scrum on track.

Scrum is just a ceremony. It shouldn't be a replacement for any large portion of communication on a team.

The biggest thing in any action RPG is exploring cool dungeons and killing baddies. As such, I've been working hard to do two things:

• Build interesting random dungeons
• Create cool random encounters (say, maybe a triggered ambush?)
• Make all of the above work in multiplayer

That's right, this post demoes the framework for the new dungeon generator as well as multiplayer AI for monsters in those dungeons.

I took a different blogging approach this time, and decided to not only make a quick demo video, but also a more in-depth walkthrough of these features. Video for that is below:

As always, thanks for reading and watching! More updates coming as soon as I can code them together!

Play with it live at: https://webgl.softfalcon.com

Before I headed to Unite 2018 in LA, I wanted to put together something cool I could use as a talking point to start conversation.

Something that has always caught my eye is ocean shaders. There's a lot going on in rendering an ocean that interests me. There's the shape of the waves, the reflections from the sun and sky on the water surface, depth and how it interacts with the light to produce varying color in the water.

I also wanted to try writing a WebGL layer from scratch in Javascript. So, why not throw together a water shader inside WebGL?

This solved my goals in meeting people at Unite:

• Have something cool to show
• Put the visualization on the user's device
  • This is huge, as it shows deployment chops as well as the "magic" of it "just working" in their hand on their phone
• Make the technical challenge difficult enough that it striked up conversation about how it was done

The result is the ocean shader that you can play with at https://webgl.softfalcon.com.

I'll be doing a write-up of what was involved in creating this at some point so that others can learn from the experience. For now, have fun playing with pretty waves on your phones, just remember to watch your battery life, WebGL is pretty resource intensive on mobile.

I've been playing around with a set of 3D tiles lately using some of the concepts I learned from creating Viking Village.

In a relatively short time, I was able to create a web server (UDP networked) to stream zipped json payloads of terrain chunks.

I'm using Simplex Noise to generate the general features, an edge walking algorithm to add cliff edge tiles, and some evenly distributed clumping to generate some interesting surface details (i.e clumps of trees, boulders, shrubbery).

I'm not sure whether this top down 3D tile based style is something I want to pursue for Viking Village, but I do think I'll turn this into a small multiplayer game. The streaming server for the terrain lends itself well to creating a multiplayer server, and I do love multiplayer gaming.

I'm thinking a rogue-like where players get dumped into the wilderness and have to cooperatively find an evil temple, descend into the depths, and clear it out. I'm sure this tile based setup could be adapted to create Diablo-esque dungeon levels as well as sprawling top worlds. Maybe add some crafting elements to allow players to setup a base camp to return to during their adventures, unlocking shops and other town resources.

I really like the way it's coming together so far, and the new architecture for the server and Unity behaviours is making this a joy to code.

Also, before I forget, I want to mention the incredible people at Kenney.nl for providing the 3D tileset! They have a lot of great low poly stuff compatible with Unity and Unreal. If you're into low poly, they are worth looking into.

More later, thanks for reading!

I realized in my last video that I didn't really demonstrate the login, character select, and terrain loading.

This is my attempt to remedy that mistake. Hope you like the preliminary login and character select screen!

Long time no see folks! I have some big news about Viking Village. The game client has been fully ported over to Unity!

All the same spanning wilderness and kingdom survival, now powered by a AAA game engine!

I made the leap and left behind XNA Game Studios and have finished porting the existing game functionality over to Unity. I figured I should post a quick video showing off how beautiful the procedural terrains look in the new game client.

I'm hoping with all of the creator tools, asset store, and huge community this will help me get Viking Village to the alpha stage more quickly.

More videos to come! Thanks for watching!