Tuesday, May 3, 2011

Not That We're Keeping Track or Anything

Some eight months ago former Linden Lab CEO Philip Rosedale laid out a set of concrete goals for the development and advancement of Second Life. Although Rosedale has once again stepped down as CEO—with former EA exec Rod Humble coming on board a few months ago—I thought it might be useful to take a look at what's happened with these goals. Despite perpetuating an open-and-accessible atmosphere, in my experience Linden Lab is consistently obtuse and non-committal about how it plans to manage and extend its virtual world platform. These concrete goals were a rare articulation of the Lab's goals and were made after the Lab went through a significant downsizing that saw many well-known and veteran employees exit. The goals bear revisiting.

Mesh imports will enter beta by the end of 2010

Status: accomplished. Mesh imports have been up and running on Linden Lab's beta grid (Aditi) for some time. However, Linden Lab has not released a timeframe for when they plan to bring mesh imports to the main grid, nor have they articulated the "costs" of mesh—since meshes can be very complex,unlike sculpts some single-object meshes are likely to cost more than one prim apiece. Also, since creators will want to use mesh to create elaborate vehicles and environments, Linden Lab may increase the maximum size of standard prims to accomodate large meshes. Standard SL prims are currently limited to 10m³; reports have Linden Lab considering a new limit of 64m³. Other reports have Linden Lab waiting until third party viewers like Phoenix have viewers available based on Linden Lab Viewer 2.x code—viewers need to be updated to support mesh, and that pretty much means no viewers based on the older 1.x codebase will be able to view mesh objects. But the bottom line: mesh is in beta and interested users can work with it.

Users will be to create and edit their own Display Names

Status: accomplished. Display Names have been out on the grid for some months, and as a result Linden Lab has limited all new avatar account to the last name "Resident." In my personal experience, user adoption of Display Names seems low; even people using viewers that support Display Names often turn off because it reduces the usability of their Friends List, means they have to put up with a variety of "clever" names that use Unicode characters and special symbols in an effort to stand out from the crowd. And, when the name system fails, folks with display names just see avatars named "??? ???". My viewer doesn't support Display Names, but I wrote a little script that will tell me someone's Display Name. I find it's quite common for SL users to address people by their usernames rather than their Display Names, implying they are not using the Display Name feature in their viewers.

Group chat and sim crossings will be "fixed" by the end of 2010

Status: not accomplished. In my experience, group chat is even worse now than it was at the time Philip Rosedale promised it would be fixed by the end of 2010. The last two Buccaneer Bowl games have been significantly impacted by group chat failures, and what little group chat I engage in is, ironically, mostly people apologizing and complaining that their comments are out of order or missing due to group chat failures. Linden Lab apparently has been experimenting with converting the group chat to XMPP technology; however, recently reports have those tests producing no significant improvements.

Similarly, sim crossings have not improved. Linden Lab has rolled out a change that, when users cross between sims or teleport to a new sim, the data associated with their avatar is compressed (using gzip) on the server they're leaving, then transfered to the new sim where it's unpacked, saving significant bandwidth. However, if I have that correct, the change doesn't seem to have produced a significant improvement to sim crossings—there is a minimum lag of several seconds walking between sims—and the new systems now seem to mean teleports are a crapshoot: you can teleport, but you might not get anywhere, or (worst case) you might find yourself logged out of Second Life instead of arriving at your destination. I'm basing these comments not just on my personal experience, but observing avatars near me. My home parcel is located on a sim boundary, above a Linden railroad that other SL users often "ride" across a multitude of sims for fun. I regularly see avatars get "stuck" at the sim boundary near me when they try to ride the railroad across it—and my sim is comparatively quiet and underutilized. Don't believe me? Hop one of the bizarro vehicles that roam the Linden Roads and see how far you get before crashing.

Linden Lab will be shutting down the Teen Grid

Status: accomplished. Linden Lab shut down the Teen Grid and moved 16 and 17 year-old avatars to the main grid, where they are permitted to visit regions rated "General," but not "Mature" or "Adult." So far as I can tell, the sky hasn't fallen as a result of letting teen users onto the Main Grid; however, as an avatar who has elected not to participate in Linden Lab's age verification procedure, I have noticed an increase in the number of locations I am unable to visit. Some SL users I know remain deeply concerned about the presence of minors on the main grid, and apparently land on Linden Lab's adult-only continent (Zindra) continues to command a premium compared to mainland.

Linden Lab's Viewer 2 will adopt a scrum development process

Status: accomplished. Linden Lab has been kicking out betas and "official" release of Viewer 2 with regularity, so much so that they're up to Viewer 2.6 in their official release channel—they ought to be considering the question of whether they're going to run with Viewer 2.10 after 2.9, or whether they're going to round up to Viewer 3.0. The frequency of Viewer 2 updates has enabled Linden Lab to bring new features into the Viewer quickly; however, some of those released have had significant stability issues, suggesting the Lab hasn't magically solved the traditional problem of scrum development: quality assurance.

Viewer updates updates will be background downloads

Status: accomplished. Updates to Linden Lab's Viewer 2.x now download in the background whether users want them to or not.

Assets (textures, inventory, etc.) will be fetched using HTTP

Status: accomplished, I think. Linden Lab's HTTP Assets project has apparently rolled out, however, I have not personally observed any improvement in texture or inventory loading, while I *have* observed an increase in frequency in which my viewer's local cache (of textures and other stuff I'm "using") becomes corrupted, resulting in a crash or causing the viewer to crash on launch.

Linden Lab will eliminate new user orientation

Status: plans changed. Linden Lab does seem to have eliminated new user orientation, but "Help Island" and the infohubs are still up and running. Instead, Linden Lab has implemented "Basic Mode" and "Advanced Mode" in Viewer 2.x—Basic Mode restricts users to a tiny handful of approved destinations in the virtual world; users can only select from canned avatars, and can't buy or create anything. To do any of that stuff, they can switch to "Advanced Modes" at any time.

Linden Lab plans to make an iPad client for Second Life.

Status: no update. I haven't heard a peep about this from anyone since the day it was "announced."

Not that I'm keeping track, but I count that as one failure (sim crossings and group chat) out of eight items that had measurable deliverables—and considering the new Viewer "Modes" as the equivalent of eliminating new user orientation. The iPad app was a pie-in-the-sky thing; it hasn't been done, but nothing specific was promised, either.

What else has come along?

OK, so Linden Lab has delivered on a lot of what it promised back in August. What else has Linden Lab done along the way?

Avatar "soft physics"

The Linden Lab official viewer now supports butt, belly, and boob jiggle.

Web profiles

Linden Lab has converted avatars' in-world profiles to a Web-based platform, meaning avatars' profiles are now visible to the entire Internet (unless a user specifically makes their profile private). This change has not been without ramifications: for a while, Linden Lab was publishing all information in an avatar's profile to the whole Internet regardless of inworld privacy settings (including hidden groups); after that, setting a profile to private broke specific group management functions, making many common group management tasks impossible. (Here's the official bug; you have to have an SL login to view it. Linden Lab claims to have fixed it.)

Marketplace emphasis

Linden Lab is increasingly emphasizing its Web-based SL Marketplace as its preferred way to sell virtual goods, de-emphasizing the idea of shopping "in-world"—actually teleporting to an in-world store, finding the thing you want, and buying it directly. In some ways the Web-based store is a convenience move: many SL users apparently prefer shopping on the Web, and considering how frustrated I get shopping in-world, I can understand that. However, I also consider the Marketplace highly problematic: it seems to be consistently gamed by creators offering hundreds of versions of the same item—making it difficult to impossible to browse—search seems useless, and there's often not enough information available to make an informed decision. However, Linden Lab's stress on the Marketplace seems financial: they get a five percent cut on every sale.

Website revision

Linden Lab rolled out another major shift to its Web site—now dubbed a "Community platform"—that attempts to present SL's forus, Q&am;A, blogs, and knowledge base into a single unified presentation. (SL Wikis, the Marketplace, and user account management remain separate.) Among the new changes, users can get ranked based on "his or her contributions to the platform."

In-world search still deeply broken

I hate to flog a dead horse, but in-world search hasn't worked for me in either my antiquated Snowglobe viewer or Linden Lab's official viewers since October of 2010. Apparently it does work for other people, but remains deeply flawed, with complaints that people, parcels, businesses, and locations can't be found even when searching for their exact name. As a result, I have to rely on group notices (see group chat failures above) and out-of-band sources to learn when events I might be interested in attending are scheduled to happen. The only way I can find SL businesses by name is to use Google—and hope they have an out-of-band blog or other resource that includes a pointer to their in-world location.

So why did I run through all this? I'm sensing that the day Linden Lab blocks viewers based on the old 1.x codebase is coming sooner than later. At that point, I will be faced with a decision to use Linden Lab Viewer 2.x—which, in addition to a deeply problematic interface, is burdened with deep privacy concerns—use a third-party viewer based on Linden Lab's Viewer 2 code (same privacy concerns, plus a third party in the loop), or…leaving SL altogether.