Sansar 2019 Product Meetings week #6

The Sansar Games Room by Sansar Studios

The following notes were taken from the Sansar Product Meeting held on Thursday, February 7th. The full video of the meeting is available here. This meeting was attended by Ebbe Altberg, so the primary discussion points were around Sansar as seen from the CEO’s standpoint, with some riffing (or should that be Ebbing?) on ideas.

R29 Feedback

The R29 Release was made on Tuesday, February 5th (see here for more). This has largely been positively received. The ability to spawn teleport portals has largely been favourably received, although some experience creators would like an option to block it (to avoid cheating in games, etc.). Ebbe Altberg also expressed some dissatisfaction with it (how it looks, where it appears, what is required to spawn it, etc.), so this could well be changing in the future.

There were some issues with the release, and as a result an update was deployed on Friday, February 8th with fixes for:

  • Frequent crashing issues caused by various factors.
  • Some experiences appeared darker or brighter based on the user’s video card.
  • Unexpected text inputs and modals with no input options may appear when saving store-bought items back into the inventory in edit mode.
  • Saving an object back to inventory that contains an inventory cluster resource would contain unlicensed info and cannot be sold.

Avatar Turning

The latest release has re-enabled a left-handed snap-turn capability for those in VR (the F5 key). This appears to be related to some aspects of VR control  (e.g. a “basic” and “advanced” control mode, although one of them hasn’t really been followed through on.

Strafing vs Avatar Turning

There was further debate on strafing vs. avatar turning. In Sansar, as with SL, the camera is generally positioned behind the avatar and facing in the same direction as the avatar. However, when walking left or right, rather than turning to face the direction of travel, the avatar “strafes” sideways by default with the camera still behind them (see below), unless a mouse / right-click combination is used to turn the avatar / camera.

The default “strafing” motion in Sansar – see as confusing for some new users

While common among various games, this is causing confusion for incoming users (including those from Steam). A suggested solution is to keep the current “strafing” motion the default action, but add a capability to have the avatar / camera turn automatically that users can enable, if preferred.

Ebbe’s Comments

Immediate Focus for the Lab

There are two key areas that the Lab will be focusing on for Sansar over the next (roughly) six months.

“Day Three retention” – increasing the number of users engaging in Sansar beyond their first one or two exposures to the platform.

  • The introduction of the Home Space / social hub apparently increased “day three” retention by some 50%.
  • The quest / rewards / achievements system being considered / developed (see my previous Product Meeting notes) is a part of this work.
    • As per those notes the idea is to boost interest in the platform by propelling users into a space where they are doing and earning things whilst also making friends and discovering other places in Sansar as along the way.
    • This system will over time be opened up to allow experience creators to also leverage it / participate within it, but this will not be possible straight off the bat.
    • First step is for LL to build the required infrastructure needed to provide such a system (back-end tools, possible HUD system, etc.), and then get things running.

Avatar enhancements – LL acknowledge there is a lot can could be done to improve the Sansar avatar (.e.g. attachments on custom avatars, MD clothing on custom avatars, skinning, bone deformation, and so on). Some of these may be small-scale projects, others more complex. Time is being given to determining the order in which this work is to be handled and how it should be progressed, which includes seeking feedback from creators.

Avatars will likely be the first element in Sansar to get a level of Detail (LOD) implementation.

One aspect of avatars related to user retention is providing a flow by which incoming users can easily see and understand how they can access the Sansar Store and buy avatar accessories / custom avatars, how they can use the items they have brought on their avatar, etc. It’s not directly clear to first-time users that any of this can be done, and so thought is being given to making it clearer.

Other Points of Discussion

“Sansar is A Virtual World”

The idea that Sansar is a virtual world was a concept that LL has steered away from, referencing it as a “platform for virtual experiences”.

“World” was seen as problematic, as it suggests a wholly contiguous space, a-la Second Life, which Sansar clearly isn’t. It was also hoped that by positioning Sansar as a “platform”, it would maximise the platform’s appeal to a broad cross-section of audiences. However, the Lab are now viewing Sansar as a “world”. This will feed back into into things like user retention, how some capabilities are implemented in Sansar in order to give consistency of expectation / function / immersion across experiences), etc.

An example of the latter is teleport portals. On the one hand, the Lab would like to see these use more for moving between experiences, but on the other they are currently immersion breaking, due to the experience load screen being displayed. Ergo, attempts might be made to try to lessen the impact of scene / experience loading when teleporting between experiences.

Thought also needs to be given to balancing how far experience creators can limit “expected” functionality within Sansar (e.g. preventing free camming or teleport portal spawning) without confusing users. One suggestion is to add icons to the Sansar client to indicate when core options are disabled at experience level (as with Second Life). Ebbe Altberg appeared unfavourable to this idea, citing people don’t look at the icons – something I would personally dispute; if people are made aware of the icons, they will refer to them.

Atlas Sorting

The Atlas initially sorts experiences in a complex manner. In the first place, those experiences currently being visited are listed. After that, a number of factors come into play to determine the listed order of experiences (has the experience had a high volume of traffic recently? what’s the number of likes it has received? etc.).

This has led to some experiences generating regular traffic being pushed down in the Atlas listing in favour of those that have a “one-time” spike in visitors who only stay for a few minutes. LL is aware of this issue, and is working to adjust the algorithm used to sort the experiences ordering in the Atlas (outside of those currently with visitors) help expose those generating steady traffic engagement much earlier in the default listings.

However, a major effort isn’t being put into refining the Atlas overall, as LL would rather have people moving between experiences and discovering places to go from within Sansar (e.g. using portals to link experiences) rather than hopping between immersion and the Atlas.

Collaborative Building

When the Edit mode was moved server-side it was seen as the “first step” towards collaborative building in Sansar. This is still something LL want to provide, but with the focus on user retention, enabling people to work easily together within Sansar’s Edit mode has now slipped to “some day”.

In Brief
  • Nvidia bug: the Lab has a workaround for the bug in the latest Nvidia drivers that is affecting some Sansar users. It’s not clear whether this fix will be released as part of a patch release or the next major release.
  • There is a bug that means that if avatars in the scene are not moving (e.g. they are seated), they may not be seen by those entering the experience until such time as they do move.
  • As well as LOD capabilities, more intelligent use of textures, etc., to reduce the number of draws, and other back-end changes that can be made to help improve performance within experiences, LL also hope to provide creators with more tools that will inform them and help them to make performance-related choices when building their scenes.