Now that the newly-minted Linden Lab CEO, Rod Humble, is not quite so much “the new guy” as he was the first time that I interviewed him, I’ve been nosing around the PR corner of Linden Lab to see if I could get him back for a second round.
While there was a lot of calendar juggling involved, Humble and Peter Gray (Linden Lab’s PR manager) managed to find a little time in a busy schedule for us to get together and talk.
On this occasion – though I didn’t have as much time as I would have preferred – we were able to go in-world into Second Life and talk in real-time, reclining in some beach-side deck-chairs; a pleasantly non-businesslike setting. During the course of things, we got just a little distracted here and there, on some fascinating topics which I very much hope to revisit in future, but the core focus of the discussion was about usability.
Usability is one of the three drums that Humble has been beating since he settled into the big chair at the Lab: Service, usability and lag. Usability is one of the most interesting ones, since many times in the past it hasn’t been clear whether Linden Lab even thinks of usability in the same terms as the rest of us.
So, that’s where we began…
Speaking as a user, quite often I’ve been moved to wonder whether the Lab defines ‘usability’ in the same sort of way as the rest of us.
RH: Sure, well I suppose it’s a philosophy. And it’s that Second Life is magic, but Second Life stops it being magic.
That might sound a bit weird, but let me elaborate. There is so much in SL that I adore and so much of it seems just out of reach. Either by UI, or service or a lack of seamlessness.
So in general, I want to make it all just work and work well. That’s usability in my view. So it’s a broad term.
Making it do what it says on the tin?
RH: Yeah, exactly.
To me, Second Life is endlessly fascinating. The people and the creations, the serendipitous encounters and all of that. It just seems like you slam into a wall just about every time you turn around, or try to take a step.
RH: Yes, that’s exactly it. And when one of those walls is removed it just feels like a Summer breeze. Everything feels better. So I know its getting a little tedious by now when I keep harping on about service, usability and lag. Its all a part of the same thing in my view. I want to get all that working and do it this year.
The other part is new users. Poor usability hurts us all because it stops a significant number of new people joining and trying it out….
New users keep us all going.
“it killed me seeing thousands of people every week simply fail to even get to meet anybody in-world, or even move one meter” |
RH: Yeah they do, and the number of new users we get trying Second Life is amazingly high compared even to large MMOs. There is something about Second Life that is evergreen – it’s truly stunning – so it killed me seeing thousands of people every week simply fail to even get to meet anybody in-world, or even move one meter.
Some fell at registration (that’s better now). Some fell at navigation (that’s a bit better now). Many still fall at finding anything to do that they want. That’s the next focus.
Although I will say even though it may not show up as big numbers on concurrency, a larger number of those new users are already sticking, which is nice to see. It shows that we can move the dial.
Basic Mode is a part of that.
RH: Right now Basic Mode is deliberately a separate thing so it can iterate very, very fast without impacting the full viewer. But that’s changing fairly soon, I think, when they become integrated.
Better integration of basic/advanced mode in future iterations – so we won’t have to restart the viewer to switch between them?
RH: Yes, that’s correct. Ideally it should be seamless. Even more ideally it should just add and remove elements as you need them. And have a intuitive layout. That’s a bigger issue, though.
You know I’m going to have to bring up the Viewer 2.x UI – and also that Viewer 2.x seems to be unaccountably slower than 1.23 was, with identical settings – you didn’t really come in until 1.23 was already gathering dust a shelf somewhere.
Sluggish is the word that comes to mind, compared to the late 1.x codebase, at least on many systems.
RH: Ah, I didn’t know that. I will ping folks and see what i can dig up. No reason we should ask folks to use something slower that’s for sure.
As for the UI layout, it is… well… to be changed 
Which leads to the question… how, exactly, do you evaluate potential changes to the UI?
RH: Right now the process is simply finding very obvious things that most intelligent people would find unintuitive or burdensome. I admit that’s a stop-gap measure.
More long-term when it comes to actually add things and hopefully increasing features? I don’t know. My instincts are to solicit a lot of feedback, then go quiet and implement, then return for further feedback. But I am open to other approaches. With search for example the team did a bunch of AB testing which was handy.
How do you see third-party viewers fitting into that ecosystem? I mean, is the widespread use of TPVs a good thing for Second Life and the Lab, or is it a sign that the official viewer isn’t currently meeting the needs of many users?
I think our friends who work on 3rd party viewers are an asset to usability.
Second Life has many features, and I think it’s wonderful that our many different kinds of users can have options in how this kind of thing is presented.
My opinion is that the transition of code bases has been rough over the past year or so. My hope is that by the end of this year that transition will be mainly behind us and the focus can be on expanding a unified code base that benefits all our customers.
Are there any usability items tentatively slated for… oh, say by the end of the quarter?
RH: Oh a bunch!
I think the next thing you will see from us are a much improved login screen, new newbie avatars and chat fixes – I run into group chat issues every week. Today in fact.
That obviously needs to be fixed. I’m not going to commit to dates just yet though. I hate disappointing people.
That’s all we had time for, but I’m hoping to follow up with Humble on service and lag in the coming weeks.
As I mentioned, there was quite a bit of ancillary chatter in addition to what you see here, and I came away with a better understanding of the Lab’s new CEO, and his drive to overhaul the more delinquent aspects of the Second Life service. I get the feeling that Humble isn’t just gunning for the low-hanging fruit, as the Lab has done traditionally, but is determined to tackle the hard, awkward stuff, whatever it takes.
And that’s a good feeling. Given more time and future occasions, there’s so much more that I would like to interview him about.
Since the beginning of Humble’s tenure, the Lab seems to be accomplishing more and faster than I can recall it ever doing. Not entirely without its problems, to be sure, but things sure are happening, as the Lab seems to be ploughing its way through long lists of delinquent tasks and feature-requests at an unprecedented rate.
“If you do what you’ve always done, you’ll get what you’ve always gotten.”
Win or lose, Humble doesn’t seem satisfied with the Lab just continuing to do what it has always done, and I think many of us would agree with him there.
Back in 06/07, there were a number of articles ranking Second Life’s then-V1 user-interface as among the worst ever. All these things are subjective, of course.
If two-thirds of users prefer the V1 UI in TPVs, then (presumably) the converse is true and one-third prefers the V2 UI in the official viewers instead (of course you could argue that some people use the TPVs for the features rather than for the UI, but then that would imply that the V2 UI isn’t as much of a problem – let’s keep things simple, and think about the UI for a moment).
33% is a big slab of users. If it were me, I’d be taking notice if it were just ten percent. With a one-third/two-thirds split, I’d be either looking for ways to accommodate both groups, or looking for some alternative option, like the customisable design that Rod mentioned above.
Considering that viewer 2 has been the default download for more than a year now and that people have to actively seek out the alternatives and to live without features such as media on a prim and multiple clothing items, achieving only 33% of market share is a sign of a pretty big failure in my view.
I don’t think the degree of failure is really relevant. What you’ve got is roughly 30% of the users (that’s one in every three) who would potentially feel bad about the notion of (for example) switching to (or back to) the V1 UI. That’s too large a number to ignore.
Whether it was a success or failure that got us into this situation, we’ve now got two camps – each far too substantial to be ignored – that don’t like the UI that the other one prefers, and that’s the situation that needs to be dealt with.
I completely agree that the degree of the failure is not as relevant as why it happened. I think the reason is unwillingness to honestly listen to feedback and to just do pro forma beta testing without taking notice for improvements suggestions.
I was really stunned in the case of the chat feedback. The request was simple and with a user story as they required. I wanted the ability to dock local chat with group and private IMs in a tabbed view like in viewer 1. I often go to places where there is quite an active local chat and conversations often spill over to private IM. Having to use a mouse to switch between the two different forms of chat was making the whole experience much more cumbersome that it needed to be and was impediment to use of the new interface.
Q Linden flatly refused that this was the case and said that they did internal testing and determined that keyboard switching alone made it too easy which led to people typing wrong messages to wrong IMs. He’s vetoed any attempt to introduce the ability to *optionally* dock all chats together. He wrote a whole long blog post about how having customizations and options were bad for you.
Oz Linden refused to even look at how any feature was done in v1.
You cannot have this attitude and come up with a successful product or improve the user experience. Quite the contrary.
I’m not advocating that the whole UI gets reversed, but to cherry pick the parts that worked well in v1 and incorporate them into v2 which has many new features people want.
With V2 being the default download, there must be a large proportion of V2 users who are very new, and maybe the statistic which really matters is the number of new users who stay with both SL and V2.
I wonder what the Lab sees in the numbers for users who have signed up over a month ago, and still connect? How much of the TPV market is the older users? Is the V2 retention rising faster, or slower, then general user-retention.
I don’t know what the current set-up is in the Lab for this sort of data collection, but maybe they need a statistician, and one who has experience of survey design as well as the maths. (I’ve seen what my brother charges per day for contract work. I doubt they think they need a statistician that much.)
I do like his comments on the future of the whole “Basic Mode” premise. The current implementation just seems so wrongheaded, but making it more seamless and allowing you to customize the UI more to your needs would be ideal.
I will be watching very closely on these new “newbie avatars” he mentions. The starter avatars impact how we build to a great extent. We live in a world of giant avatars almost entirely due to 7-8′ tall starter avatars and poor avatar creation tools. It prevents us from unlocking SL’s full creative potential.
Not to mention the starters are the “face” of SL to new users. If you start new users in ugly avatars with arms that are too short, heads too small and legs that are frightfully long it is going to leave a very bad impression. Yet this is precisely what LL does, and has done since the beginning. Correcting this would be a sign of an impressive turn-around within LL.
[...] interview with Rod Humble, Linden CEO, in which he says a few things about viewer development. See Behind the curtain: Talking with Rod Humble about usability. This interview is well worth the read. « Second Life Mesh Update Week [...]
They are going to INTEGRATE stupid mode and unusable mode? I was hoping they would dump the one that won’t let you access inventory.
*face palm*
I don’t remember having any major gripes with the v1 GUI, sure it had a few quirks, but nothing big.
Well, except when they decided to switch it to dark text on light background; but that was easilly worked around.
Regarding the communicate window issue, i didn’t had it, not sure why, but i easilly undocked local chat and arranged my windows the way i wanted, while at the same time hearing people complain it couldn’t be done; i clearly remember testing both the current official viewer and a TPV, i think it was the Cool Viewer, that was supposed to have things the old way, and i managed to dock and undock localchat and arrange the windows just as easy on both, while still hearing people complaining they couldn’t do it anymore. The impression that it gave me was that just because the defaults were changed from one way to the other and some minor details in the visuals of the GUI were changed people got stuck thinking LL had removed functionality when they in fact had just changed the defaults while keeping the customization functionality intact. Though perhaps my memory is innacurate, or for some other reason i somehow skipped the troublesome versions catching the fixed version fresh out of the oven while most others hand’t updated yet; dunno really…
Good work, Tat. Take whatever opportunity you can to bring some news out of the LL cave. Any insight we can gain is better than no insight. This sounds like very positive news to me. We were on hold for a long time, but this year the news has become quite upbeat. I’m for showing more patience since things are improving!
[...] 1. Following up from her February interview with Linden Lab CEO Rod Humble, Tateru Nino has a great chat with Mr Humble on Second Life’s usability. [...]
Usability. I wish some website interview would talk about usability from a standpoint of people with disabilities and ask the hard questions. When will people with disabilities stop being ignored by Linden labs? They have the wonderful world called second life for people with disabilities can do things they normally can’t including talking to people and socializing. I have a disability Duchenne Muscular Dystrophy and the only way I can navigate the world of second life is through the speech to text dictation system. I use the software built into Windows Vista and also in windows 7. And the special mouse where I can move my fingers on. Others are not that fortunate to have hand movement. Version 1.23 was wonderful I could use my speech dictation system all be it glitchy. To walk around I could use my mouse with the onscreen moving control on the second life viewer. I am a builder, scripter, and helicopter enthusiast giving rides to others. When 2.0 came out imagine my horror to have some arrows missing from the movement controls specifically left and right stride. Try building without stride, try flying a helicopter on life without being able to turn in a controlled fashion. Luckily a friend of a friend knew a programmer with Linden Labs and managed to convince them to put it back. What was their reasoning for removing it? Users were getting confused. For how long is my question. In the technological age of the Xbox and cellphone controls yeah right. It was wonderful that version 2.3 now have all the controls back. What a relief until I find out voice dictation does not work it keeps asking like there isn’t a text area there. I can use it on any program on my computer except the second life viewer. I guess my issues isn’t popular enough. That sounds to me like they don’t care how wonderful second life is to people with disabilities. I shouldn’t have to hire or beg someone to make a third party viewer that will work. I’m not asking for preferential treatment just a viewer that works like any other program. A viewer like 1.23 or like 2.0 but with all the movement controls.