2016 SL project updates 34 (2): TPV Developer meeting

Arranmore; Inara Pey, August 2016, on Flickr Arranmoreblog post

The majority of the notes in this update are taken from the TPV Developer meeting held on Friday, August 25th. The video of that meeting is embedded at the end of this update, and references to it are indicated through the use of time stamps in the paragraphs below. My thanks as always to North for recording and providing it.

This is not intended to be a transcript of the entire meeting, which featured discussions of some situations specific to individual region rather than SL as a whole. However, key discussion points have hopefully been highlighted.

Server Deployments

There were no scheduled deployments for week #34.

SL Viewer

Release Candidates

[00:15] Both of the current release candidates – the VLC Media Plugin replacement for Windows and the Visual Outfits Browser viewer, have elevated crash rates. Each is liable to receive a further RC update in week #35 (week commencing Monday, August 29th).

[00:53] There will be a new Maintenance RC viewer appearing as well.

64-bit Viewer

[00:58] Progress is being made on the 64-bit versions of the viewer, and these are expected to surface (Windows and Mac) as project viewer quite soon.

Bento Project Viewer

As noted in my Bento update #24, the Bento project viewer was issued in what is hoped will be the final project viewer iteration before it progressed to Release Candidate status. Version, released on August 25th, includes the following updates among its changes:

  • The latest viewer updates body size less frequently (starting / stopping an animation) so vertical height repositioning should be less jarring.
  • During mesh upload, SLM files are by default not created and not used. This can still be overridden by changing the debug setting MeshImportUseSLM
  • The show bones display has been modified to use colors differently, distinguishing between joints that are skinned to, joints that have position overrides defined, and all other joints.
  • Animation of collision volumes has been fixed.

Voice Updates

[02:17] The voice updates are progressing, but is not expected to appear in a project viewer soon due to some remaining issues which need to be resolved.

[03:56] There are some server-side voice updates currently on Aditi undergoing test. When these will be moved to the main grid is subject to further discussions with Vivox, which are due to take place in the forthcoming week. When these updates are enabled on the main grid, they will be gird-wide.

[33:10] Oz offered a reminder that in-world / region issues are unlikely to be related to / cause Voice issues in the majority of cases where the latter are being experienced, as the only way the simulator and its host server are involved in voice is when a user is initially connecting to Voice (the simulator provides the channel addresses to the viewer). After this, all communications are between the viewer (via the SLVoice plugin and the Vivox voice servers.

[34:10] As a part of the ongoing work to better monitor and control Voice, the Lab will be introducing improved logging on exactly what the viewer is doing when connecting to Voice.

Exceptions Handling, etc

[02:44] As noted at the last TPV Developer meeting, the Lab is going to be making some source code changes, including a clean-up of asserts in the viewer and how exceptions are handled (the latter to try to prevent the viewer crashing as a result of encountering an exception), and further rendering pipeline clean-up. These will be showing up in a project viewer at some point in the not-too-distant future.

Accessing Crowded Regions & Increasing Max Number of Users Per Region

[08:30] Trying to access a crowded region for an event is often a matter of pot luck. when full, you have no real choice but to get re-trying via teleport or crossing  a region boundary. As a result, the Lab has frequently been asked to add some form of access queuing system to regions.

While a queuing system has been considered, the Lab feels it brings with it may questions around how it work to present an optimal solution for users. As such, while they are open to proposals for a queuing system, optimally, they’d rather work towards trying to increase the number of avatars regions can support. This is something also being looked at from time to time (see my May 27th, 2015 SL project update as an example), although there is no work actively being carried out on it right now.

Abuse Report Categories

[17:00] As indicated in my last TPV Developer meeting update, the Lab will be introducing a new region capability to handle abuse report categories. Once implemented, TPVs will be able to adopt the use of this capability, rather than having abuse categories held within the viewer, where they may get out-of-step with the categories supported by the Lab.

There is also a reminder in the meeting that abuse situations should be reported ASAP after the event, so that the Lab can refer to the related simulator logs during investigations, and that there is a L$10,000 bounty payable on SEC JIRAs which identify actual or potential abuse vectors  / exploits which the Lab can act upon to close.

AMD Graphics Issues

[26:10] Many users with AMD GPU (notably the RX400 series) have been experiencing glitches and “tears” appearing in their world view  – see BUG-20057 and FIRE-16829. User Hurana Ugajin reports she is having a dialogue with AMD on the issue, and a solution may be forthcoming soon. When / if available, a request has been to post it to the SL technology forum.

Region Resource Allocations

[36:19] A question is asked on whether it would be possible to allocate simulator resources (script usage, etc), on the same basis as object land impact – thus allowing resources to be assigned / capped by parcel size, etc.

The Lab does not believe this can be done without significant changes to SL, as a lot of the underpinning resources are not location-based (e.g. scripts are run on the basis of where they are). Instead, the Lab is focused more on penalising individuals for excessive use of resources (e.g. if you run a script which abuses resources in a region, all your active scripts are throttled, not just the offending script). Such abuses can also be AR’d by those noting them.

Other Items

Aditi Access

[04:42] Some users are apparently being told that in order to access Aditi, they require Payment Information On File (PIOF) and to have completed a tutorial. These are actually part of the requirements to be able to upload mesh. No special requirements are required to access the beta grid other than your user name and password.

There is , however a problem with new user accounts not showing up on Aditi when they should, and a request has to be passed to support to enabled them. This is down to changes in how the Lab is handling account data, which has left Aditi in a state of flux, but a fix for this is in the pipeline.

Account Management

[06:22] The account work mentioned above is part of ongoing back-end work the Lab is carrying out related to personal data is handled and stored even more securely, and help resolve some account management issues. One outcome of this will be a single change to a user’s password will be applied pretty much immediately to both Agni and Aditi.

Large Texture Over-Use

[43:40] A continuing problem in SL is the over-use of large (1024×1024) texture on every surface, no matter how small. While this is a bad content choice, it has been suggested that to discourage this, LL should consider charging differently for different texture sizes. This has been discussed by the Lab, with the admission that given the already low cost for uploads (L$10 per item), there is a certain lack of conviction that introducing a nominal fee scale based on image size will do much to discourage the practice.

Avatar Complexity Calculations

[48:18] As noted at the last TPV Developer meeting, the Lab will be carrying out further refinements in how Avatar Complexity is calculated. A suggestion put forward at this meeting is that as well as having a per avatar limit, if a scene-wide limit could be defined / set. This is viewed as an “interesting” idea.

2 thoughts on “2016 SL project updates 34 (2): TPV Developer meeting

    1. Yup. I’ve previously covered the provision of 64-bit Havok support (here, for example) when the 64-bit viewers appear.


Comments are closed.