The following blog post, unless otherwise noted, was written by a member of Gamasutra’s community.
The thoughts and opinions expressed are those of the writer and not Gamasutra or its parent company.

Prefabs are not a new topic to game developers. These will be referenced in future articles of mine, so here’s an overview to a concept for mastering massive content.

If you’ve ever had to deal with tons of objects across multiple scenes, you’ll inevitably wish you could group similar or exact copies together for quicker management. If you’ve added a desk lamp to 30 different office rooms in 15 different scenes, wouldn’t it be nice to turn up the that light’s intensity once instead of 450 times?

That’s the power of prefabs. 

So, what exactly is a prefab? It’s a collection of game objects used in multiple places across the game. It’s a concept I’ve made sure to use across past projects like Six Days in Fallujah, Prey 2, and Mafia 3.

A Prefab should be a Scene. The only thing that should make a prefab different than a scene (a level, map, environment, etc.) is it exists inside a scene. A large city scene can have several small prefab neighborhoods and those in turn can have repeatable prefab houses inside of each.

Some game engines develop prefabs as a unique feature, separate from a scene’s format. By developing a prefab as an extension of a classic scene means they inherit all features and power scene editing has!

Any feature that exists for a scene should exist for a prefab:

• A collection of models (a bench, a bus stop sign, a garbage can).
• A building with AI navigation and interaction nodes.
• An enemy with a unique change (BadGuy_Type_4 with dual pistols).
• A complex script sequence (depending on project’s script file handling).
• A skybox with all associated FX and animation.
• A Security Door (a door model & logic, security palm switch & logic, emergency lighting & logic, surrounded by cover nodes & path volumes).

Advantages

1. Repeat occurrences only have to be created once. When the original prefab is modified, all instances are updated.
    
2. Multiple people can work on the same scene by partitioning the scene into prefabs. Each user works on their individual prefab, testing their work in the master scene.

Game Usage

Prefabs should be created in the editor as a new scene or from a selection within an existing scene. The most common type is a Static Prefab. They can be saved to the game library for reuse later. When the game is built and run, all static prefabs are ungrouped (sometimes called exploded). All parts should have their name prefixed to the prefab’s unique name for reference. All game logic now points to unique items maintaining unique scripting.

In the editor they can also be exploded into an existing scene to creating unique, free roaming objects, no longer connected to its source.  

Aside from static prefabs, Dynamic Prefabs are another option. These are used to stream in content. If done right, prefabs can be the primary means all content is contained and streamed into the game.

With procedural systems like Houdini, you could even extend prefabs into becoming Smart Prefabs. These feature a set of rules to decide what, where, and when to place a variety objects at different times. This article is only about the core prefab concept though.

Every scene can be made with a number of prefabs.  Take a look at the following diagram as an example:

Now let’s see how prefabs let several people work on the scene: 

The Art Team works on the Ground Terrain A location:

• Artist 1 models Building D with indoor lighting to be used in 2 locations.
• Artist 2 models Building C to be used in 4 locations.
• Artist 3 models Building B to be used in 3 locations.
• Artist 4 models Building A with indoor lighting to be used in 4 locations.

Artist 5 is lighting the whole scene. Some lighting features on-off logic with Gameplay.

The Audio department is working on ambient Sound for the scene, mixed with Gameplay.

The Design Team is providing gameplay to the Ground Terrain B location:

• Designer 3 is creating a generic prefab for assigning AI and stats for a Train Event.
• Designer 2 is creating City Tower logic to be used in 4 separate places in Ground B.

And lastly, Designer 1 is providing gameplay for the overall scene in the Master Scene.

When any person is finished, they check their prefab into Source Control to share with the rest of the team. When the master scene is opened again, any existing prefab is updated with the latest changes. When new content ready, the user should be able to pick them from the Resource Browser to add to the scene without restarting the editor.

Example Scene

Let’s group areas in a scene into prefabs, allowing several people to work on each area:

A Designer takes a city block and creates a series of prefabs from it:

• Each building is named appropriately and saved as a prefab.
• A city block of prefabs is saved as one sector (optional).
• The master scene is now a framework of roads, game logic, prefabs, and sectors.

Original Scene

Scene cut into 3 Game Sectors

Artists individually open each prefab building to edit it:

• They can replace the stand-in model with a final version.
• They can even add new objects like doors, windows, lamps, lights, etc.
• They can open the sector to understand relations between each building prefab.

When a Designer re-opens the scene, all prefabs are updated with changes the Artists have made.

Artist 2 works on an Apartment

within Game Sector 1

Artist 3 working on individual Motel Prefab

How do you set up scenes to use sectors? How do you work on them? How should logic be arranged in them? Before we talk about sectors, we need to understand how the scene played in the game (the Master, Main, or Persistent scene) is organized. Each scene consists of four basic components:

Applying this concept to sectors, we can create a framework for maximum flexibility to add (or remove) parts of the master scene without impacting the master scene itself. Each component can become a prefab for each sector. 

Some prefabs don’t have to be exclusive to a sector; they can span several sectors. This allows one person to check out a prefab to have access to the whole scene. The disadvantage is when sectors are move around in the master scene requiring contents to be moved manually to their new position.

Editing Prefabs

Two methods of prefab editing should exist:

1. Edit the prefab in a new scene.
2. Edit the prefab in the current scene.

Editing prefab “A” in a new scene isolated from original scene.

Editing prefab “A” within the scene for instant feedback how it affects the scene. 

(The fences will need to be modified.)

By editing the prefab within the current scene, it removes switching back and forth between scenes to see your changes.

By their nature, prefabs placed in scenes are instances of the original prefab. If changes are made to the prefab, those changes propagate to all copies in every scene. This can be helpful for example, if you want to change a texture or a light across a hundred copies. You might however only want one small change.  You need a way to remember those individual changes, instead of creating a new prefab for each change, exponentially increasing the number in the game library. 

This can be solved by tagging any differences and re-applying those differences after changes to the original is made. Technically speaking, values unique to any type of object are stored in the parent prefab or scene file to override its original values. When a change to the original is made, all copies reflect that change unless the value has been overridden on the instance.

An “Allow Deltas” action provides a way to maintain these overrides. With a prefab selected and “Allow Deltas” on (set to true), unique values will be accepted and maintained. If “Allow Deltas” is turned off, all values are reset to their default.

Without Deltas, editing one prefab edits all of them.

With Deltas, unique edits remain while

any unchanged values inherit changes.

To prevent the need to open prefabs, another option is enabling Child-picking. This ignores prefab containers and directly picks any object in view, no matter how deep they’re nested. Normal editing of prefabs should be outlined in red to indicate changing this one instance will change all instances of it.  With Deltas on, objects inside prefabs should be outlined in cyan, letting you know any changes will only be applied to the selected object.

Today’s games involve thousands, if not millions, of objects. Managing all that content, making quick changes, and sharing team efforts, becomes a massive burden without a prefab system in place. If you want to spend more time discovering what’s exciting about your game and less time playing 52,000 card pickup, then take the time to invest in a proper prefab feature.

You can find more posts like this on my website at CuriousConstructs.com.

Take-Two CEO Strauss Zelnick has scoffed at the notion that single-player titles are dead in the water. 

In a recent interview with VentureBeat, the chief exec was quizzed about the state of Red Dead Redemption 2 (Take-Two being the publisher behind Rockstar’s open-world western), and was asked what he makes of the perceived shift away from story-driven, single-player efforts. 

His response was decidedly candid, and while Zelnick admitted there are some who believe multiplayer, free-to-play titles are the only sensible way forward, he claims they are very, very wrong. 

“There are people [today] saying that a game won’t work if it’s not a free-to-play battle royale. People really are saying that, and not even tongue-in-cheek. I don’t buy that,” explained Zelnick. 

“Single-player, in my opinion, is not dead, not even close. Companies that feel like they’ll just avoid the hard work of building a story and characters and go right to where the money is in multiplayer, I don’t think that’s going to work. I’d be surprised.”

Although Take-Two and Rockstar have experienced long-term success with Grant Theft Auto Online, the online multiplayer portion of Grand Theft Auto V, Zelnick says there’s “no evidence” to suggest people don’t want a single-player experience. 

In fact, he agrees with the assertion that players only became so invested in GTA Online after buying into the single-player offering, and believes those narrative-driven, lone wolf experiences aren’t going anywhere.

You can hear more from the Take-Two boss by checking out the full interview over on VentureBeat.

We’ve just launched an open beta for Creator Homepages, a new part of the Steam store designed to help players discover and connect with the developers and publishers behind their favorite games. With this feature, you can explore the full catalog of games created by the developers and publishers of games you enjoy and you can choose to follow those creators to be automatically notified when they release their next title.

How does it work?

Any developer or publisher on the platform can now set up a customized homepage to showcase their full catalog of titles and content. Once set up, these homepages can be found by clicking on the developer or publisher name from the store page of your favorite games.

Check out the official Creator Homepage Announcement Page, which includes a list of all the homepages so far. This will quickly become an overwhelming and unusable list as more developers create their homepages, so at the top of the page we’ve also specifically highlighted the developers and publishers behind games that you’ve recently played.

On these homepages, you’ll find standard lists of top-selling or new released titles from that developer or publisher.

You’ll also find collections that the developer or publisher has created to best highlight their portfolio of games and content available on Steam.

A studio might divide its games into collections by genre or franchise, or could choose to highlight their fan-favorite or top-selling games. A developer of only a single game might primarily dedicate their homepage to announcements of new projects.

Regardless of how each developer or publisher chooses to customize their homepage, you can easily follow them to be notified when they release their next title or post an announcement. Newly released titles from developers or publishers that you follow will show up at the top of your Steam homepage and we’ll send you an e-mail to let you know that they’ve released something new (as with other Steam e-mail notifications, you can opt out at any time by visiting your e-mail preferences page).

Creators can also set up unique URLs within Steam for easy reference to their homepage. You can see an example by following this customized URL for the Valve developer homepage: http://store.steampowered.com/dev/valve

Why Beta?

We’re pretty excited to get the core functionality into the hands of players and give developers the opportunity to set up their presence on Steam. While we haven’t worked out all of the smaller bugs or finished adding every feature we’d like to, we decided that the basic set of functionality is worth putting into the hands of players and creators. We still have a number of features that we are considering adding and there are still a few rough edges that need smoothing out, so opening this system up as a beta to players and developers will help us gather feedback and suggestions that inform the direction of those features.

Over the previous few weeks, we’ve worked with a number of developers and publishers on Steam to get their pages set up and help us work out as many of the bugs as we can. As a player, you’ll find that many of your favorite game makers probably already have a spiffy homepage created and customized. But there are still quite a few developers that have not yet had a chance to set up their pages and will do so during this open beta period.

Looking for more details?

Check out the official Creator Homepage Announcement Page.

For game developers, check out the Creator Homepages Steamworks documentation for more details and information on how the system works, the features that are included, and necessary permissions for game creators.

Reporting bugs and feedback

As always, we’ll keep an eye out for your feedback and suggestions as to what you’d like to see added or changed about this system.

Google has added a Guest app to its emergent and currently open source Fuchsia OS to enable Linux apps to run within Fuchsia as a virtual machine (VM). The Guest app makes use of a library called Machina that permits closer integration with the OS than is available with typical emulators, according to a recent 9to5Google story.

Last month, Google announced a Project Crostini technology that will soon let Chromebook users more easily run mainstream Linux applications within a Chrome OS VM. This week, Acer’s Chromebook Flip C101 joined the short list of Chromebooks that will offer Linux support later this year.

While it’s encouraging that Chrome OS will soon support Linux apps is addition to Android, it’s not entirely surprising — since Android and Chrome OS are based on Linux. Yet, one of the first things Google emphasized when it revealed Fuchsia in 2016 was that it’s not based on the Linux kernel.

To some, Fuchsia seemed to be something of a betrayal considering how Linux not only forms the basis for Android and Chrome OS but also the Google enterprise platforms. Why add another Windows or iOS when we were getting so close to everyone sharing a common Linux foundation?

No doubt, Google has some very good reasons for avoiding Linux. One reason may be the age and complexity of Linux. By starting from scratch, Google can escape that aspect and deliver more elegant, up-to-date code with fewer targets for hackers. Google is also baking secure updates deeply into the OS, and unlike Linux, is isolating applications from having direct kernel access.

Open for now

Back in 2016, we thought Google might be skipping over Linux to shift to a proprietary OS that it could control the way Apple dictates all things iOS. That may still happen, but for now Fuchsia is an open source project.

Some also speculated at time that considering the trim little microkernel, Google was bypassing Linux due to its inability to scale downward into the MCU realm. Yet, MCU-based IoT does not appear to be the current focus of Fuchsia. Several reports, including a TechRadar post last week, have said that Fuchsia is intended to replace both Android and Chrome OS, and the combined platform will eventually be called Google Andromeda.

Earlier this year, 9to5Google reported that Fuchsia would include separate UIs — an Armadillo UI for phones and a Capybara UI for desktops — and like Android Things and other new Android variants, would tightly integrate Google Assistant voice technology. Essentially, this is the same idea that was behind Microsoft’s failed plan to offer a common Windows for phones and laptops, or Canonical’s defunct “convergence” version of Ubuntu.

Guest ex Machina

Whatever Fuchsia’s destiny, Google needs to attract mature applications, as well as developers, and the best way to do that is to add Linux app compatibility. The new Guest app, which initially supports Linux-based platforms including Debian, works with the Machina library to accomplish this in a way that goes beyond what you can get from QEMU, suggests 9to5Google.

Google describes Fuchsia’s Machina as “a library that builds on top of the Zircon hypervisor to provide virtualized peripherals that integrate with a garnet system.” Zircon is the Fuchsia microkernel, based on Little Kernel (LK), and formerly called Magenta. Garnet is the layer that sits directly atop Zircon and offers device drivers, the Escher graphics renderer, Fuchsia’s Amber updater, and the Xi Core engine for the Xi text and code editor. Other layers include Peridot for app design, and on top, Topaz, a Flutter-supported app layer.

Machina adopts the Virtio virtualization standard, which is also used by the Linux Kernel-based Virtual Machine (KVM). It makes use of Virtio’s vsock virtual socket, “which can open direct channels between a host operating system and its guest, to allow for conveniences that would be otherwise impossible,” says 9to5Google.

This extra effort will likely enable enable fast mouse performance, automatically adjusted screen resolution, and support for multiple displays, file transfers, and copy and paste, says the story. This appears to be much like the allegedly superior emulation that is expected with Google’s Project Crostini for running Linux apps on Chrome OS. The news of the Guest app follows earlier reports that suggested that Google is building an Android Runtime into Fuchsia rather than depending on emulation to run Android apps.

App emulators should be viewed with some skepticism. Most of the mobile Linux OS contenders promised some sort of Android app compatibility, but they have generally failed to deliver. Still, by building emulation deeply into the stack from the start rather than adding an emulator on later, Fuchsia may well offer Linux developers an emulator that they can live with.  

Join us at Open Source Summit + Embedded Linux Conference Europe in Edinburgh, UK on October 22-24, 2018, for 100+ sessions on Linux, Cloud, Containers, AI, Community, and more.

By Joe Robinson 20 Jun 2018