Second Life Presence

"To date, IBM owns 50-plus islands inside Second Life -- and this is just that one virtual environment," says Booch. IBM-Rational's virtual Codestation was used at the Rational User conference last year, and Booch says they still use it regularly for lectures and promos. The company also owns some large amphitheaters that are used by South American and Chinese employees to meet for training.

"They actively fill those places for the purposes of training, and its saving money, because we can get people together and not have to fly the m from place to place. A variety of IBM labs have centers in Second Life -- Germany , Israel, India -- where they have sites, they do work there. And IBM has a classic sort of business office too, where they host meetings."

As for Booch himself, he's personally given more than twenty lectures inside Second Life. "It saves me time, and it saves IBM money, so I'm getting active use in that regard. I just did a lecture to a Manitoba group. They had rented a [virtual] IMax theater, and were projecting me on an IMax screen with my Powerpoint, and I gave the lecture that way. This is a common modus operandi for me. At the other end, Bluegrass happens to be a specific instance for a particular problem for which we think virtual worlds are fruitful."

Where Bluegrass Fits In

What problem is that? Explains Booch further, "Bluegrass is a particular experiment to look at the specific problem of temporal and geographically-distributed development teams." Bluegrass is aimed at deepening the semantic density of virtual world interactions, Booch tells us.

Semantic density? Booch elucidates, "If you look at some of the collaboration you see going on inside virtual worlds today, these are what I would call interactions that are of low semantic density – I may have a business meeting, or a discussion or lecture, and we might be sharing docs or Powerpoint slides, and those things are semantic – they have meaning, but they're not very dense, meaning that the information in them is largely in the words and tribal memory of people around them."

He continues, "On the other hand, raw running naked code has high semantic density. It's highly connected to other lines of code, files, req docs, and performance tests. So there's a real density associated with that information. That in itself creates a whole class of problems. How do you grok, visualize, navigate and manage that complex information that’s well connected? It's hard enough in a 2D world."

As Booch says, given 20 million lines of code, traditional 2D tools like Eclipse -- IBM's open-source framework -- can be stressed by such complex information. But there's another issue. "With size like that, we are now dealing with a problem that is no longer how you make individual developers more productive, but how do you make that team more productive?"

Additionally, this issue is further complicated for IBM, because the company has so many geographically-distributed software development teams, in China, Europe and elsewhere. In addition to the cross-cultural distribution that takes place, there's a temporal distribution. Sometimes these teams are twelve hours apart. "People walk away, and all you have left is code," says Booch. "Code is not the whole truth. So what can you do to increase information density around those artifacts so that someone can pick them up later?"

Defining The Experiment

As Booch explains, building a solution for collaborative development is a multifaceted issue. "What are the points of pain of getting people to work together in a distributed fashion on these semantically dense things?" He asks. "There are some technical things we know how to overcome, but it’s the social parts of it that become very difficult. How do you build a sense of trust among a team of people who never meet one another in their entire lifetime? How do you build that level of trust so they can work together?"

These are "water cooler problems," Booch continues. "You have these notions of getting a team in a room and interacting on a regular basis. You can't do that across 12, 24 time zones. Are there things that can reduce the friction of those social interactions to increase the level of trust, and therefore productivity? Yes, we think that virtual worlds can contribute greatly to that."

Providing the opportunity for the water cooler kinds of discussions can create a serendipitous collaborative environment, stresses Booch. "Social networking has shown us that, provided the right mechanics and right social structures, cultures will find themselves. So Bluegrass is an experiment to see if we can do that."

The Bluegrass experiment is part of IBM's larger Jazz collaborative platorm. "We've got a project led by John Patterson and others up in Cambridge who say, 'let's see if we can put these hundred small things together to produce an environment that is frictionless and improves social levels of trust while working on things,'" says Booch.

Since IBM is so entrenched in Second Life, Booch clarified why these efforts are separate. "There are two pragmatic reasons not to use Second Life," he says. "The team just wanted to put their hands on the code for creating a world itself. They really wanted to get their hands dirty, so to speak. The second thing is they wanted to explore security – not personal information security, but if you're taking artifacts that represent key IP, you can't afford to put that in a place where there's any chance of it being stolen. This virtual world lives behind a firewall and is captive -- that’s why were pursuing that. That is why Bluegrass is what it is."

Issues Of Friction

So given that one of Bluegrass' primary objective is to help avoid friction when dealing with issues of "high semantic density," we asked Booch to help define the sort of friction Bluegrass aims to work around. In addition to the previously-mentioned physical distance and time zone problems, Booch says, "One point of friction is, can we create a sense of space? And I think we as humans desire that sense of space and nesting -- our caves, if you will. You get the idea. That place to come to – it's hard to measure, but a very comfortable thing to say, that even though we're scattered around all parts of the time-space continuum, we have a place we name and can call our own and make it home. That reduces connection friction."

For example, Booch says, in the Bluegrass world, there are signboards around the trees that people can tack notes to, leaving a trail for new visitors to follow indicating the current environment and events going on in the space. "I can just wander around and see what's going on -- not unlike visiting cubicles if I were to visit, say, India."

We asked for specifics on how the Bluegrass project fits into the larger sphere of Rational Jazz. "Jazz is a larger project which is name-branded like Bluegrass, and it has both desktop-centric elements and a focus around Eclipse, and web-centric elements, like portals and dashboards. Inherently, though, the core of Jazz is 2D in its presentation. Bluegrass is the 3D persona to that, which says, 'here's the meeting place.' Bluegrass is an experiment in providing a 3D place where projects driven by Jazz can meet."

That being said, however, one of the challenges the team currently faces is the relationship between the virtual world and the real-world interface. "There are some other folks who have done experiments with 3D modeling of UML (Unified Modeling Language) in virtual worlds. There's a group in Japan that I first ran across, and another UK group that's done this. They take UML models -- which is really a 2D rendering I and my colleagues produced. They realized, 'wow, we can add a third dimensions and these models are big -- so how about importing them into a virtual world, so you can fly around them, and do things that way?'"

Continued Booch, "There's some value in putting these 3D models in the world. You'd really like to make things live – I'd like to reflect Eclipse changes in the virtual world simultaneously, and that's really hard, because the membrane between the worlds is still a little tough."

Looking Down The Road

So what's next for Bluegrass? Booch says they're not close to thinking about productizing it yet: "This is why research is doing it -- that’s the domain of research, to do things that are not product related, and discover where the hard problems and opportunities are and laid out in the landscape."

But Booch still has some ideas about the future, looking ahead. "If I were to project and go out on a limb five to seven years out, theres another technology that I would like to see emerge to help make this happen, and that’s the advent of digital paper. So that I could now imagine project rooms in a physical space, that align with digital paper that provides an immersive space either at my desk or in the space itself."

He continued, "In an ideal place, you could imagine projects having these physical offices that are seamless with the virtual. Where we find a way to have a more porous membrane between the virtual world and the real world, the digital paper will help. We'd have artifacts in virtual worlds that correspond to real desktop and server artifacts. These are solvable problems in the next generation."

But in the short term? "I see something closer to where Bluegrass is now -- which is having a virtual reality product that we sell that says, here's another server – there's a virtual world loaded in it behind a firewall, and its just another piece of jazz."

Would IBM open up all that code, then? "I think -- not IBM, I personally think that the way Linden Lab is going, the philosophy is that their value added is not in architecture or source code -- it is to provide the best environment and service around it. So they're happy to open-source all their code. Is that similar? IBMs really jazzed about the notion of open-source stuff because we get our value elsewhere. It's not an unreasonable projection -- but I don’t make those decisions."

Booch thinks virtual worlds infrastructure will be commoditized over time, like any interesting technology. "The value added is not going into the code and tweaking. Right now, there's lots of diversity. We see so many different products, there's going to be a shake-out. You're not going to care about that anymore -- can you provide the best world to do the best things? We aren't quite in that phase, but Bluegrass is a first step in that space."

New Frontiers, New Challenges

Finally, Booch discussed how the advent of virtual worlds is helping bring a whole new slate of issues to the forefront of enterprise. "For us as geeks, this is interesting, because we're so used to dealing with the technical side of things. The virtual worlds aspect is forcing us to consider the social side. There are simple cultural things we get caught on."

For example, says Booch, his team had discussed having some games available in the field inside Bluegrass -- the way employees at Google or Microsoft or other companies might have pinball machines around their real-world offices.

"The team was brainstorming, and came up with a list, and I said, 'wow, these sound like games from a middle-class suburbia upbringing. What about games that a kid in India or China would have grown up with?' We have to think outside the box -- we don’t want to be dead white guy-centric. We're just not used to thinking that way. Material that might to offend people -- or, can the avatars have a head covering, for Muslims -- seem like simple things, but they have tremendous implications. And being IBM that is truly transnational, we have to think about those things."

Booch recalled an example he'd heard from a colleague about how Saudi Arabian women can now meet Imams in virtual worlds -- something they can't do in real life, but can now do virtually, since the Koran only speaks of meeting in the flesh. The bottom line? Even with all of these issues and possible sources of friction, the potential is greater than the obstacles, and IBM will continue to sustain its efforts to pioneer these new frontiers.