TLDR

• OpenXR now has a sensible generic controller profile
  
  • Digital buttons (menu, primary, secondary)
    
  • Analogue trigger & grip
    
  • Joystick axes & touch
    
  • Grip/Aim pose tracking
    
• This is a boring-but-huge deal: no more "laughably unusable" default controller
  
  

STANDARDS WORK MATTERS and this is an small but important milestone, standards work is often not celebrated enough. It is slow careful compromise over countless meetings which many dismiss as useless but standards are the foundation on which empires are built and international cooperation and progress secured.

Shoutout to: Andreas Loeve Selvik (Meta), Bastiaan Olij (Godot), Bryce Hutchings (Microsoft), John Kearney (Meta), Jules Blok (Epic), Nathan Nuber (Valve), Rylie Pavlik (Collabora), Lachlan Ford (Microsoft), Yin Li (Microsoft) for this important work.

Story time: How I once built a sexist control scheme by accident and why the people in the room matter
Take-away: You should care who is in the room when designing new things

Full spec: XR_KHR_generic_controller

New Baseline: Why XR_KHR_generic_controller is a Big Deal

This provides a 90% mapping for the Touch Controller defacto standard which has emerged. Even when designing the PSVR2 controller we felt that pressure from devs not to deviate too far from this faux standard.

Sure, we're all hyped for Meta's wristband (North → CTRL-labs → Meta). Eye-tracking and hand-tracking matter too. But controllers aren't going anywhere. 6DoF is the baseline. Tools exist because they give precision and control. Logitech's MX Ink stylus is proof controllers are not just "technical convenience" and why paintbrushes beat out finger painting.

In the stone age of pre-1.0 spec we didn't know what controllers would win and Vive wands and PlayStation moves were still dominant. So we settled on what was a laughable simple set of inputs. Select and Menu button with a grip and aim pose for pointing and the ability to vibrate. Note: The 1.1 spec has broadened it slightly with pose references but nothing more. It might seem silly now but at the time it was a reasonable baseline.

The previous simple controller was laughable unusable by any reasonable dev. Leading Pico and some others to emulate the touch profile to get a faux default. Which in term led some OpenXR devs to stop trusting runtimes as they would misreport the controller. This also ensures that any future hardware controllers will have at least this baseline and likely future BCI or meta-controllers, brief side fuck you Facebook for tainting that common and useful word with trademark bs, are likely to map to a virtual controller of this subset. So yes this is a huge deal which allows developers in confidence to know what the minimum is across platforms.

Brief aside: My History with Controllers

The Stone Age of Motion Controls

This goes back to 2008 when I did an independent study on the Nokia N95 at university. You have to remember how bloody primitive things were then - the Wiimote had just landed in late 2006, showing that motion control could work for mainstream gaming, but the iPhone wouldn't get a proper gyroscope until 2010, and the PlayStation Move was still two years away. We were working with absolute rubbish sensors.

The N95's accelerometer was capped at a pathetic ±2g range, compared to the ±16g sensors we take for granted in modern VR controllers. My study concluded that while it was suitable for rapid prototyping of motion-based games, the platform had severe limitations. The accelerometer was marketed as revolutionary but was actually quite limited - motion detection wasn't nearly as advanced as the marketing suggested. The processing of motion was complex, the device's low range and lack of absolute positioning meant you could only detect the most basic motions. Still, you could create successful game mechanics if you designed around these constraints rather than against them.

The gaming audience was also getting wise to the repetitive nature of that first generation of motion controllers. We've come an extraordinary distance from those early days of ±2g sensors and relative-only tracking to today's sophisticated 6DoF systems with high-precision IMUs.

Modern Day: When Diverse Voices Matter

I'm proud to have been part of the PlayStation Studio feedback group involved in developing the PS5 & PSVR2 controllers. Working to help refine and design the controllers taught me more than I could ever share - from seeing why we cut or added things, ambitious ideas left aside and spectacular engineering saving features we thought for sure would be cut.

I also advised Logitech when they finalised the MX Ink stylus. Bringing diverse voices into the conversation uncovered a core tracking issue due to ergonomic diversity. How I hold a pen differs from most and more importantly how sculptors hold their tools is very different, causing us to cover a key sensor. They tweaked the arrangement and the issue was solved before shipping. An important lesson in getting feedback and how Logitech doing the right thing early improved their product. The Mx Ink is a great input device and I look forward to the XR stylus being a staple.

It is that story and this announcement which prompted me to write this article, because it perfectly illustrates why the people in the room matter so much for building technology that works for everyone. But first my original motion controller story and where I first learnt this lesson.

Story: How I build a sexist controller

I want to share a story from my early game development career while working at Mere Mortals in 2007 on an evaluation version of the Move controllers as an early 3rd party dev on a Pétanque game. We had been asked to test the hardware, feedback and pitch for funding to make a launch title. This was off the back of a Wii games we had done with Ubisoft called World Sports Party. We did all the ball jokes I assure you, as the only female programmer on the team we did so fucking many. This was fresh off experimenting with the Nokia N95 and some Wii coding though both devices paled compared to the moves. The moves had outside in tracking, a better sensor and higher expectations. So I was tasked with writing new motion control code. You must remember how early this was and the gradual evolution of motion controllers. So we didn't have a lot of good material to work off at the time.

For those unfamiliar the primary input is an underhanded lob action similar to the popular Wii Sports Bowling but with more emphasis on release angle as well as the line. There was also an element of wrist flick for spin. While we have outside in absolute tracking the field of view was small and close to body motion. This meant we lost absolute tracking for sections, falling back to relative 3dof track, and while the sensor was much better than previous it was still often maxed out by the basic lob motion causing loss of relative data. Often at the apex of the parabola which was down low out of the the field of view. You would be amazed at the amount of G pulled by the tip of a stick short in your hand while doing even subtle actions. That combined with the various ways people throw and hold the controller meant I needed some calibration code and various code paths.

I was young and inexperienced so my first attempt was asking the player to hold their hand at their side with controller pointing down. I was proud of my work and a friend Alice had recently joined the company and sat next to me. Also the artist I was working with, Bryony, was able to give frequent feedback. I felt I really had dialed it in and was getting more nuance that our previous g spike model which was quite simple.

Though as I asked a few more to test the controller it was hilariously bad my algorithms fell apart and wider reviews of the work were harsh especially by the QA team. I kept digging into the issue until I decided to do a full survey of everyone in the studio and collect all the motion logging I could to build a better algorithm. So one by one I got a few test throws from every person in the office. Including the office support staff, bosses and any randoms I could.

Immediately a shocking split emerged in the office of those who could throw every time like they intended and enjoyed the controls vs those who could not and were frustrated with it. I poured over the data looked at my photos and questionnaires... it was obvious all the women in the office loved it and the men didn't. The artist I was working with was my friend Bryony. Between Alice, Bryony and myself all of the people close on hand while I was working on it were women. I had made a sexist control scheme. It wasn't like I was checking genitals what was the problem?

My instructions didn't account for gendered understanding of language. The instruction was
> Hold the controller at your side pointing down at the ground
Simply put women understood my meaning and pointing directly down at the ground but men didn't follow them exactly. For them a position of hand at rest with controller pointing down was not pointing at the floor but at roughly a 45 degree angle. Yes all the crude jokes were made about other things held.

The code detected for both calibration cases and the problem was solved and the new control scheme was great. Sadly the game never shipped Sony ExDev chose not to provide the additional funds required to get the game ready in time for launch of the Moves so it was pushed onto the large pile of unpublished work all of us creatives share.

The story illustrates the importance of diversity in the room, and not just by gender.

Why People in the Room Matter

The Sony Experience

I'm proud to have been part of the PlayStation Studio feedback group involved in developing the PS5 & PSVR2 controllers. Working to help refine and design the controllers taught me more than I could ever share. From seeing why we cut or added things, ambitious ideas left aside and spectacular engineering saving features we thought for sure to cut.

While working on the PSVR2 controllers one Japanese product designer and me were typically the only women in the room. I felt privileged and lucky to be in the room but we were often the minority opinion which is hugely valuable and every meeting I felt like my coworkers acknowledged that and gave weight to it, ultimately making decisions based on good data and process. I was often wrong and I am happy to admit that, though rarely I was the only one who was right, and those small handful of validated cases were critical. I thank Sony for that and honestly think they handled it SUPER well and we have come a long way from the tiny one nation team of small engineers setting the tone.

The Numbers Don't Look Great

The diversity of standards groups suffers for understandable reasons. Companies often only have one spot to fill on a committee or maybe two. There is a lot of incentive to keep the room small because it is the work of endless meetings, compromise and tough technical decisions on uncertain data at times looking towards a hard to predict future, or at least it was a decade ago. This intense selection pressure coupled with all the other discrimination and selection pressure from school age, through uni, hiring and promotions means that standards committees, especially technical ones, are about as diverse as a white mayo sandwich with the crusts. Hey it's progress the crust used to be cut off.

Based on an analysis of 111 contributors to OpenXR working groups using public data only. This is some manual google work with AI help to compile and is not a great dataset. DO NOT TAKE THESE AS FINAL FIGURES. Though they are broadly illustrative of the situation and I would love to see a more formal diversity report from the working group.

• Gender: Male: 82.0%, Female: 10.8%
  
• Corporate concentration: Meta, Google, Microsoft represent 32% of all contributors
  
• Geography: USA: 47%, China 13%, Sweden 11% (from 36 verified nationality)
  
  

Gender was inferred from names, companies were listed and geography was AI researched and I chose to leave out sexuality, race, ethnic data as it is sensitive though that matters of course as well. Education background was US dominant with UC San Diego, Columbia, and Harvard leading. No surpise for VR peeps to see San Diego get top spot. Though I could only find reliable education data for about half. How many have accessibility needs or non-uniform biology even things as small as a lazy eye make a difference if it's in the room when decisions on eye tracking are made.

The picture looks better at industrial diversity so this is some great news:
- Industry backgrounds: Gamedev 13.5%, CAD/Manufacturing 6.3%, Hard Sciences 5.4%, Creative Arts 4.5%, Business/Humanities 4.5%
- Cross-disciplinary experience: Contributors span game engines, aerospace, neuroscience, filmmaking, and architecture

This is great because while XR is dominated by gaming it is important a broad range of opinions are heard. Also to avoid too much influence from dominant use cases as XR applications in art and industry expand out and grow.

I ascribe no malice to anyone in this case and generally as almost always with these issues everyone is trying their best and they are good eggs. Though the world we live in is imperfect and affordances for everyone. The problems are systemic and I think OpenXR are doing the best they can, but I wanted to highlight it. At most I would request a diversity report to be published so we at least have better data on the issue than my cobbled together public scan. The Khronos Group themselves did a survey in 2022 but never published the results that I could find and say they are committed to Diversity, Equity, Inclusion, and Belonging. Though they did hightlight

Results from the survey indicated that we need to focus our attention on gender diversity, BIPOC (Black, Indigenous, and People of Color) folk, and disability access. One area we appear to be doing well is with LGBTQ+ identities, which are well represented within our membership.

You want the best people in the room, but it does matter who is in the room, even guys unable to follow simple on screen game instructions deserve to be heard and throw balls.

Conclusion

Overall I wanted to write this article to celebrate what I think is a great step forward for OpenXR. The new generic controller profile represents real progress - a sensible baseline that developers can finally rely on. The contributors who made this happen deserve recognition for their careful, often tedious work navigating endless meetings and competing corporate interests.

But when I went to thank the list of contributors - some I know, others by only reputation and others not at all - I was struck by the uniformity and the memory of that old story. Standards are the foundation on which empires are built, and the people who write them shape the future for all of us.

The world is gloriously varied, and our technology should reflect that. Sometimes the most important insights come from the most unexpected voices, asking questions nobody else thought to ask. Affordances should be for everyone - even guys who can't follow simple calibration instructions.

So here's to the new OpenXR generic controller profile, and here's to making sure next time, the room is a little bigger.

disclaimer: generative ai was used for some diagrams in this article