JavaOne once again brings its load of cool stuff about the numerous Java technologies. Among this year's highlights (EJB 3.0, JDO, AOP, etc), we can find the Looking Glass project which aims at "bringing 3D windowing capabilities to the desktop to offer a far richer user experience for work and play".

On TheServerSide.com, Dion Almaer posted a thread announcing that Looking Glass had been open-sourced for developers to contribute. I have been following the ongoing discussion very closely and I would like to share my opinions on this matter.

Henrique Steckelberg expresses his scepticisms regarding whether it would be a wise idea to redefine the way we interact with computers by making a parallel thought with cars. He asks whether it would be smart to change the "user interface" of a car... While I partly agree with him, I can't help but ask myself whether cars will still have the same user interface 30 years from now... "The 5th Element", anyone? The fact that something is built in some way today doesn't mean that it won't be built another way tomorrow. If things are the way they are today, it is not because it is the best the way to build them, but more because we lack the technologies to build them in a better way.

The same reasoning can be applied to computers. Everybody knows how current operating systems look like: a 2D screen, a virtual 2D desktop, some icons on it, some menus, a keyboard, a mouse, other pointing devices, etc. When windowing systems were invented, there was no 3D technology around and the only physical devices we had to interact with computers were keyboards and mices. Pretty cool for that time. Still, my opinion is that since we are constantly immersed within a 3D world, it should be more intuitive for us to think, act and manipulate abstractions in a 3D world than in 2D world. I say "should" because this is unfortunately not the case. Not yet! We have been forced to evolve within a 2D world when interacting with computers, not because it is better for us that way, but because we had no 3D technology when PARC invented the first windowing system for their Smalltalk environment more than 30 years ago. I could bet that if that had been the case, operating systems would be much different as they are today, maybe a little like what Looking Glass proposes. Of course, I have absolutely no proof of what I advocate for, but in 20 years a lot of things can happen given the exponential nature of the technological evolution curve. Let's see how things will be in 20 years :)

Looking Glass is trying to bridge the gap created by the technology shortcomings we had in the 70s. Looking Glass is just one example of how a 3D desktop can look like. Microsoft is also doing research in this field. Have a look at their project called Task Gallery and you'll see that Looking Glass is not that far off the path. You'll even be able to find similarities between both.

3D has becoming a reality. 3D screen prototypes (Holographic TV, etc) only exist in labs but people like Harold Garner are striving to make this a reality. In my opinion, the biggest challenge is not to develop 3D environments but more to create innovative peripherals that will allow us to evolve within those environments. I fully agree with Cedric Beust and others that evolving within 3D worlds is a hassle for users. But I also agree with Oleg Mikheev that this fact is only due to the limitations of today's peripherals, I'm pretty sure the future can (and will) be different. This all depends on how well 3D desktop creators and peripheral companies can work together. This is no pic-nic, I admit!

Let's move on. Matthew Wilson asks how the wonderful tree abstraction for classifying files could be further improved by 3D environments. Well, Matthew makes a perfectly valid point. There is one small problem, though. I think everyone has already experienced a situation where she needed to classify a file but figured out that more than one directory would be a potential parent for it. That's one of the reasons why shortcuts (or symbolic links if you prefer) have been created. The problem arises when you need to move your files around without updating (all) your shortcuts. Agreed, this is not completely the fault of the tree abstraction itself, but more its bidimensional nature. In the same vein as Bill Coutinho, I think a 3D tree abstraction could much benefit from what topic maps, RDF and other similar technologies have to offer, but I'm digressing, this is another debate.

To sum up, I think that Looking Glass will certainly make its way through and that some other 3D desktop projects will be incepted in the coming years. One of the problems being that such projects will strongly depend on what peripheral companies will put on the market. But again, this is a chicken-and-egg problem, as such companies don't waste resources for inventing things for which there is no demand. Apart form that, I think the success of future 3D windowing systems might come from the fact that we increasingly have more and more information to deal with at the same time as Bruno Girin mentioned. As soon as you need more space on your 2D desktop you buy more screens, and thus, you clutter your physical desktop. So instead of cluttering the physical third dimension, why not try to expand the virtual one? On another subject, my last wish would be that Sun folks change the project name to something more sexy than "Looking Glass". It always helps to have some cool name to mention when evangelizing a technology. Hearing "Looking Glass" doesn't raise any kind of emotions in me!