Saturday, April 19, 2008

Judging Linden

I dropped in today on a salon in Extropia featuring Hamlet Au. The discussion was pretty far ranging and participants included some of SL's best and brightest. One brief discussion thread centered around what Linden Lab can do to enhance newcomer experience. Although some positive ideas were discussed, the general tone had a critical flavor.

Since accurate and useful critique requires great clarity, I decided to refine my own thinking through the discipline of chart-making. Here's what I came up with:

No business (or person for that matter) can do everything well. Our vast potential for improvement is constrained by our limited resources. So we all must make choices about where to focus and what to prioritize. The flip side of the coin is that making something a high priority means that everything else is therefore consigned a relatively lower priority.

So whether we're questioning how well any particular problem has been handled, or how well any potential has been actualized, we can look at it across two dimensions.

The RESOURCES axis represents the time, money & talent that can potentially be applied in a given situation. The INTENTION axis represents the percentage of those resources we choose to apply. I ended up with four quadrants. Ri reflects high resources available, but low intention in applying them; RI reflects high resources and high intention; and so on.

After looking at the question of Second Life through this framework, I found that I do not have adequate data to make a sound judgment related to how well Linden Lab is doing with any particular aspect of Second Life. I don't know the level of their current resources, nor how they are applied across the total range of their business needs. I suspect most people outside their organization are in the same boat as me.

Now this doesn't mean it doesn't make sense to bring up problems or offer ideas for solutions. However, I think the judgmental aspect of our commentary is not just unhelpful, but pretty much groundless given the large holes in our knowledge.

So that's my little rant for the day. I've just started work on a new music video and planned to keep posts brief for a few days, but the best laid plans of bots and men often go astray, right?


Dale Innis said...

Yeah, and...? :)

We humans *love* making groundless judgements. I think it's hardwired in somewhere...

John said...

Well, it's not impossible for experienced people to judge by comparison. LL is a software company, a _graphics software_ company, an _online graphics_ software company, and shares DNA with _game development_ concerns. It shares corporate management and funding with prior companies in Mitch Kapor's sphere of influence and money. This gives educated observers a lot to work with in judging the company's progress towards common-sense goals (e.g., a good user experience and profitability).

The reason user experience gets such attention is that it's been proven, time and again, to be critical for customer conversion. And this is not just true in the individual sense (i.e., of individuals making their go/no-go decisions about whether to use Second Life regularly, after an introdutory experience) but in terms of institutional leverage. There are, for example, many companies that would invest heavily in Second Life presences, were it possible to guarantee a sensible, simple, stable user experience at sim-loads meaningful for commerce. If I could get 1) an email-only registration, 2) a two-minute client download, 3) a media panel in the client letting me project a rock-solid audio/video/chat experience to an audience in a moderated way, 4) alternate clients (e.g., graphics-limited) that run well on older PCs, 5) if I could get, say, 200 people in a sim without everyone lagging out, and 6) if the system were available with six-nines reliability, I could use SL to completely revamp the global events business. It would be "game-changing."

But SL doesn't provide any of that. Which ... in years 1 and 2, maybe ... would have been okay. But in years 3, 4, 5, 6 and 7 (always considering that by any reasonable measure, the company is earning enough money and has enough potential value that investors aren't itchy) begins looking a little ridiculous. It's time for these folks to get it together, tighten up the system, and make it useful for something other than random, small-scale, non-critical social networking and prototypic v-commerce.

Botgirl Questi said...

dale: I know it's , fun, fun. But I think they work against the constructive interests of both the judging and the judged. Like alcohol, if you're a casual drinker it's fun. But if you have an addiction to it, the cost of the fun can be very high, both to the drinker and the community.

john: Thanks for your thoughtful analysis. My reluctance to judge LL by the commercial standards you describe is partially due to my perception of SL as an ongoing R&D effort, rather than a traditional commercial enterprise. I think that "common-sense goals" were not the driving force behind the initial vision and that if they had been, we would have a much more stable, but less interesting virtual world. It seems that LL has been moving over the past year or so to be a more mainstream operation. Shifts in top management may be a sign of changes in that direction. In any case, I appreciate substantive and well-informed comments like yours.

Dale Innis said...

Another thought that occurs to me here: if we see that LL has done X, but not done Y, and X seems to require about the same number of resources as Y but to be far less important, then we can get upset on at least some grounds, without knowing the total resources that they have available.

A friend (and long-time SL resident) was complaining bitterly to me, for instance, about the addition of a "full screen anti-aliasing" control to the 1.20 viewer, which was unnecessary (can be done via the OS and the video driver already), didn't actually work, and was a strong candidate for the instability of the current 1.20 Linux viewers. And at the same time noting that the first Linux release of 1.20 didn't work at all, for anyone, due to a compiler setting.

Even without knowing anything about LL's total available resources, one can say with some certainty that they could have, and ought to have, put less resources into the "full screen antialiasing" support, and more into the "testing the 1.02 Linux viewer at least once before posting it" stuff. :)

Not that the effort put into one could have been put directly into the other instead (different types of resources, after all), so it's not a pure example of the type. But it does seem that one can have a legitimate complain about resource balance, without knowing the total amount of resources.

And I do agree with you on the addiction image. So many people reply to every single LL blog post with "why haven't you made everything perfect yet??!!11!1?" that it must be very tempting for the LL folks to just ignore the comments entirely...

John said...

Ah, but the explanation for that odd behavior is simple: a "graphics programming" resource isn't repurposable as a "general QA/unit testing" resource!

A project the size of SL is actually many projects running in parallel, each employing different specializations (in fact, even software testing comprises about four specialties, all differently-compensated, with different work-patterns and toolkits), and these are hard to prioritize cross-purpose. You might right-size things several times a year, but unless you're willing to fire and re-hire (certainly not nice and often not practical given the difficulty of finding certain mixes of skills) you need to keep people busy to rationalize paying them. And this can easily get one group well out in front of the group next to it. Makes sense internally (since there's a grand plan), but not much sense externally - particularly when what looks like 'obvious stuff' like user experience seems to be getting short shrift.