Fieldnotes from the Metaverse — Metaplace
In 2006, several virtual world veterans created Metaplace as a platform to democratize the development of virtual worlds in the spirit of the open web.
Setting the scene
Web2.0 was a perceptional shift in how the web operated. Instead of a publisher-reader relationship, websites turned into platforms where everybody could be a publisher.
Initially, the web was synchronous. Every interaction had to be first requested by the user’s browser, processed by the web server, which then sent a new web page for the browser to reload and display. In effect, every user interaction resulted in a new web page.
However, also in 1999, Microsoft Internet Explorer 5.0 introduced XMLHttpRequest (originally called IXMLHTTPRequest) that allowed browsers to make requests in the background. Web servers would only send data, processed by the already loaded web pages, making the web asynchronous. For end users this meant that websites now behaved more like traditional applications than websites, hence the term “web apps”.
Websites started emphasizing user-generated content, ease of use, open and participatory approaches, and interoperability between platforms. The trend was first recognized and called “Web 2.0” by Darcy DiNucci in 1999.
With the growing popularity of Web 2.0 culture, the XMLHttpRequest object became the de facto standard in all major web clients at the time, including Internet Explorer, Mozilla, Safari, and Opera 8.0 by the mid 2000s. Tim O’Reilly and Dale Dougherty then popularized the term at the first Web 2.0 Conference in late 2004.
The initial Web 2.0 culture valued openness and interoperability between platforms, spawning many decentralized standards and protocols to handle many different data types and contexts. By 2006, the web was a mesh network of smaller platforms (sites, forums, blogs, or other user-generated content), working together via exchange formats.
Here are some examples:
- OpenID provided an open standard to decentralized authentication
- Gravatar provided consistent avatars and visual profiles across the web
- RSS allowed to syndicate web data independent from user interfaces, becoming the basis for news feeds and podcasting
- Backlinks tracked links across platforms in a contextual manner, comparable to citations
- Last.fm introduced “scrobbling” as an open standard to exchange musical preferences
- Webrings acted as multi-dimensional curated content catalogs for the web
With all this in mind, a group of virtual world veterans came together to enable users to build virtual worlds the same way, powered by Web 2.0 technologies, protocols, and culture.
The first prototype of Metaplace was developed in early 2006 by Raph Koster in his in his spare bedroom. He took a MUD server, added custom, HTML-like tagging to the room descriptions and changed its output to HTTP. The result was displayed on a dedicated client, which could also run as ActiveX plugin in an Internet browser.
In July 2006, Koster founded Areae to further develop and market the platform. It later changed its name to Metaplace, Inc. because “nobody could agree on how to pronounce or spell it”.
Metaplace was born out of two ideas.
Creating virtual worlds and MMOs should be simple again.
From a technological point of view, virtual world servers had turned into giant, monolithic infrastructures, purpose-built for just one game world. Realistically, there was no re-use of virtual world servers (nor clients) anymore — a stark contrast to previous MUD and MOO servers that could easily be cloned, extended, and re-used by everybody.
Developing virtual worlds and MMOs thus took a lot of time and resources, meaning that only companies with deep pockets could develop virtual worlds. Companies that naturally wanted to avoid risks by re-creating what they knew already worked, leading to less innovation and experimentation.
Areae argued that Web 2.0 offered a new way with a robust set of protocols, extreme scalability, and reusability, while having low cost and low barriers to development.
Virtual worlds should be first-class citizens of the Web, next to text, pictures, video, and audio. Just another kind of media on the Web that could be used whenever the user context makes sense.
This viewed virtual worlds well beyond the context of games, but as a natural part of daily digital usage. But to achieve that, virtual worlds also needed to integrate themselves into the technology (rituals, processes) of the real world.
In a GDC talk, Koster highlighted this difference:
“When you look at the Web, the idea of cramming the Internet into one spatial world is starting to look not only silly, but pointless. We like using our iPhones to get directions while in downtown San Francisco. The idea of having to log into a virtual world and walking inside that world to get where the virtual map is seems ridiculous — it kinda sucks as a user paradigm.”
It’s not surprising that Raph Koster was one of the participants in the Metaverse Roadmap study in 2006, which also separated the Metaverse as a concept from representation and modality, thus including text-based, audio, video, 2D and 3D experiences on any type of screen, headset, or indeed sensory device.
A Metaplace server was fundamentally a web server.
Every object in Metaplace had a unique URL, representing either a static resource (image, sound, video, ..) or dynamic content. Browsers could access the content just like any other webserver: Showing files, pulling and interpreting HTML as websites, running scripts, showing directory listings. Or users could pull up the Metaplace client in their browser and view the resources within virtual worlds.
This was fully aligned to how the web was already operating: File-based resources, collections are lists of URLs, supported by script-based logic. It didn’t start with a virtual world that would write bridges or interfaces into the web (like for example Second Life did), but it was the Web.
As a result, web developers could easily pick up the technology and create Metaplace worlds of their own. They could also integrate them into their existing web pages. An example of this was the ability to add Metaplace worlds to Wordpress blogs or phpBB forums. Regular web URLs could also be added to the world RSS feeds to access and show content from the web within the worlds, for example to set up news feeds or ad services.
As part of the web, Metaplace worlds also integrated with the social networking services of the time. Communication between users was done via a chat system called “Metachat” that was accessible through all the major social platforms, from the Metaplace forums to MySpace to Facebook, even Second Life. Users could scrobble the current ambient world music to Last.fm.
The reverse was also true: Metaplace worlds could tap into existing web services. Not just scrobble to Last.fm, but pull the users preferred playlist to influence the virtual world music. Grab the geolocation of the user, their current weather and match the virtual weather to it. Need monetarization? Integrate with PayPal.
Areae opened the public Beta describing their platform with: “The power of Second Life, with the ease of The Sims, on the Web.”
Challenges and eventual pivot
Metaplace got a lot of press attention early on, however that didn’t translate into user numbers.
By 2006, gamers were used to polished MMOs with theme park-like experiences. Already spoiled for choice with World of Warcraft, Guild Wars, Dungeons and Dragons Online, with Lord of the Rings and Warhammer Online around the corner — there wasn’t really an appetite for another MMO that they needed to build for themselves first.
A lot of casual gamers were experiencing their first social games & virtual worlds through social media and Facebook games. However, playing asynchronous games with friends and posting scores to their timeline didn’t automatically translate into creating games or worlds for themselves and their friends — the barriers to entry being too high.
Metaplace lacked games to attract core or casual gamers and it didn’t have the developers because it didn’t have the audience.
There were also issues with the Metaplace client and graphics.
In 2006, there was no native support in browsers to render any real-time graphics. Java and Flash offered real time 2D and 3D support, however they required additional plug-ins and with questionable performance and no mobile support. In the already mentioned GDC talk, Raph Koster states that they started out with a standalone client, then moved to Flash, then pivoted to a standalone client again, after finally reverting to Flash. Out of options, Metaplace had no choice but to pick a technology that was against the intent of being a first-class citizen of the Web.
While the goal was a modern 3D client, the eventual Flash client was only able to render 2D isometric graphics. The default game assets had a distinct comic-like art style that was often compared to Habbo Hotel, Club Penguin and other virtual worlds targeted at kids and young teenagers, which put off older players and creators.
The result was a tight but small community of creators and players that didn’t really grow. In the end, Metaplace as a platform failed to gain enough traction to be a viable product.
Metaplace Inc. announced to close all consumer facing services, including the public world building tools and the world hosting service, on January 1st, 2010. Instead, the company would pivot towards using their tools to create several successful social games on Facebook, gathering well over one million active players per month.
In July 2010, four years after Metaplace was founded, the company was acquired by Playdom.
If you want a deeper dive into the history of Metaplace, I recommend Raph Kosters book “Postmortems” (2018), which covers not only the technical architecture but also has a frank postmortem of why it didn’t work. There are also a lot of articles about it on his blog.
Metaplace was around for only three short years. It was never a polished product and maybe tried to be too much for too many audiences. People were confused as to what it was and who it was for. It didn’t have much marketing support and didn’t gain a critical mass of players. And so, gamers and even many industry insiders sadly have forgotten about it.
Metaplace was the closest we got to an actual Metaverse infrastructure: A platform — maybe even a standard — to generate a giant network of virtual spaces, capable of encompassing every sort of virtual — maybe even real — environment, object, and actor.
According to founder Raph Koster, the aspiration for Metaplace was to “make online world elements like dynamic, graphically shared space, avatars, and virtual currency part of the standard code which drives the web.”
This included the ability for players to own and make their own spaces and objects. The ability to pull in data from the real world into virtual worlds and vice versa. Worlds could be linked to one another with players being able to jump between them — changing their identities but keeping a core identity between them.
Metaplace solved (or at least started to solve) very hard problems in the Metaverse space that are still debated today.
Metaplace satisfied Toni Parisi’s Seven Rules of the Metaverse and was in line with the concepts outlined by Matthew Ball’s Metaverse Primer — 16 years before they authored them. And as many VCs, startups and established companies race today to create a new Metaverse stack, I can still hear its faint echo, asking “What-If?”
Then again, Koster and company are at it again with Playable Worlds. Their aim is to “Build a truly cloud-native persistent world that evolves and changes over time. Our goal is an online community made up of many different playstyles coming together in a rich world.”
Godspeed, Playable Worlds!
Dusan Writer commented in 2011: “But it leaves a question: are virtual worlds places? Or will the technologies that enable 3D spaces become so ubiquitous that we’ll stop thinking of them as distinct places? Because in Raph’s view, the tools and technologies to create 3D artefacts, the system for managing your avatar and identity should be EXPRESSION-agnostic. In other words, we should have the tools for creating content and then be able to seamlessly publish that content to cell phones, browsers, Flash, separate clients — whatever, it’s not the viewer, it’s in the engine from which content is derived and creating standards and tools for expressing the content from that engine.”
Raph Koster replied: “FWIW, virtual worlds are definitely “places” in my mind. But to me, clients and devices are merely windows that look onto that place. That doesn’t preclude rich 3d “windows” — I merely happen to think that multihead, flexibly represented VWs are the future.”
About the series
The term “Metaverse” is currently claimed by many groups, driven by different incentives. Some groups attach the term to specific technologies (for example VR, AR, XR, Digital Twins or Blockchains), others see it as a future vision or narrative (sometimes dystopian, sometimes utopian). Some groups talk about the coming Metaverse, others argue that it already exists.
“Fieldnotes from the Metaverse” is a series that discusses the history, visions, perspectives, and narratives of the Metaverse: Specific milestones, their immediate impact and how they shaped the discussion going forward. The goal is a holistic and inclusive view of the Metaverse space, separating visions, signals, trends, and hype.