Eskil is at the forefront of real-time 3d graphics, leading the Verse and Uni-Verse projects and is a contributor to the OpenGL specification. Following is a short conversation with him before this year's SIGGRAPH conference regarding the significance of the technology he has developed.
I am personally familiar with the Verse project, and have been following your work for some time, would you please summarize for our readers what Verse is, aspires to be, and how it is in use today?
Verse is a network protocol for 3D graphics that allows any applications to share data in real time. So say you have a 3DStudio Max user and a Blender user who connected to a verse server. Both users can now see the same scene in both applications at once, and if one makes a change the other one can see the changes happen in real time. Both users can use this full tool set of their different applications to work with the shared data. By having real time sharing where you don't need import/export or even files to share data among applications the way you work starts changing. You may for instance use a 2D paint tool connected to a 3D tool so that you can paint textures and see the strokes appear on the 3D object in the 3D application in real time. You may have a big project where 10 artists are collaborating on a large model like a city scape, each using their favorite tools and each being in a different location but still they can see what the others are doing by getting real-time feedback in their tools. Verse also helps connect developers because any two developers who independently develop software will have their creations work together as long as the integrate verse support in to their applications. To integrate verse there is a simple open source library that you integrate in to your application to make it "verse enabled" so that it can talk with other verse applications. Verse can be integrated in to any application that uses 3D data. One could say that its the ultimate I/O for computer graphics.
At the moment verse is used for collaborative tasks s like content pipelining, visualization, architecture and so on.
What major uses for this technology do you envision which are either not possible with the current version of Verse, or simply not being leveraged?
Verse is in a very completed state and can do pretty much everything we want it to do. What I would like to see verse being used for are things like Games and interactive 3D. Its something verse supports very well, but all the tools aren't there just yet to make it easy. Verse is not a system that tries to do everything, but rather to connect everything together, so as more things get connected it becomes more and more useful and new fields of usage becomes more and more viable. Right now verse is very strong as a content creation tool, but the usages will grow with time in to simulation, entertainment, computational steering, training and VR.
What is the overall, central goal of the Verse and Uni-Verse projects?
This is going to sound pretentious but the goal is to create a new medium that is ubiquitous, like the TV, film or the web. Computer graphics so far is mostly just a tool, to create data for film and TV. Interactive graphics in games are also very limited, you cant run any game data in any game engine on any hardware, its all very hardwired. A video camera on the other hand can capture images that can be broadcasted on any channel and seen with any TV set. When you get to that point technology becomes ubiquitous and you can reach more people and the creation of content becomes separate from the development of content. For this to happen people need to have a way of sharing data in real time in a non proprietary, royalty free way and that is what verse does. Verse has many uses and doesn't have to be that grand, but that i would say is still the goal.
My personal goal is to get to a place where I'm amazed of what people use it for.
I understand you are funded by the European Union, how long do you expect this to continue, and do you have any plans to renew this arrangement and/or otherwise sustain the project after this first phase of development ends? What do you perceive as the EU's interest in promoting your project?
Verse existed long before EU was kind enough to fund the Uni-Verse project and will exist long after February 2007 when the project ends. Uni-Verse has enabled me and many others to professionally work with developing verse applications and that has been a great opportunity. I think EU recognized the great potential of verse for all kinds of industry uses.
Do you feel a technology like Verse could have strong implications for education of 3d artists and engineers?
I do, It brings a social and collaborative aspect to 3D, sharing while doing is the best way of learning.
I know you have expressed, in-depth, the reasoning behind many aspects of Verse's design. Could you summarize the goals of your approach?
The idea of netwok protocoll for realtime sharing of graphics between any application is novel in it self. But going beynod that Verse has a few quirks that may not be obvious at first glance. Verse is a very low level and built to be very high performance. People expect verse to be this huge ".net" thing that does everything, and they are really surprised [at] how small and low level it is. People expect a huge VR system that does everything, but when we say it is a network protocol we realy mean it, its a network protocol and that is it. We dont want to tell you how to do rendering, interaction or much eles at all. All we want is for people to connect their creations with others. We have Rendering engines that are very cool, too but they are not a part of Verse, they just use Verse.
What is your relationship to the Blender Foundation? Is there a history which led to the inception of the Verse project? Do you expect at any time to share code with Blender?
I meet Ton Roosendaal at SIGGRAPH 2002 and we just hit it off immediately. So since then i have been a big promoter of blender, and blender has been a big promoter of verse. I usually speak at the Blender conference, and Blender is getting native Verse support in the next version. Verse CVS and web pages are also hosted by Blender [Foundation].
What do you feel are the advantages of new-style tools like Loq Airou and Connector versus the relatively stagnant interface style used across a wide range of 3d modeling packages?
I do, I want to create very direct tools where you interact with the data directly rather then through shelf's, tool boxes, menus and sliders. A lot of 3D tools today expect you to rearrange and reconfigure your interface before you use it. Its good that you can, but you lose a lot of time doing that, especially when you have to take time to adjust to new interface. But interface configuration is a bit of a sham too. Usually when [you] configure your interface all you do is move the functionality around. What keys are used what you find in the shelf, how the view ports are divided. It still ends up being buttons you have to press. you can make some repetitive actions easier by making a script, but in the end it doesn't change the paradigm of how you work, it only makes the things you need more often a little bit easier accessible. Spending so much time with the tool and so little time with the art is very uncreative. I want my tools to be like your favorite pen, when you hold it you just cant stop being creative simply because it makes you feel like everything you draw with this pen looks good.
Do you feel these projects could or should eventually interact on a single image plane, or is it entirely intentional that they are separate? Do you feel there are advantages / disadvantages to each UI approach?
I think it is unfortunate that more and more applications try to do everything, with out more specialized interfaces productivity goes down. Back 15 years ago people created graphics applications that where very specialized, very "workstation". Back then a graphics computer cost 50.000 USD plus, so if you had some one using it he or she had to be very competent, was trained and specialized. The software was designed to do one thing very well. Nobody knew how a computer program was sepouse to looks like so people made some very specialized and interesting choices when they designed applications back then. Some amazing software where made back then like Alias and Flame. Today users expect everything to look and feel like photoshop or word. Users dont care to learn applications and they run the applications on the same machine as their kids do their home work. Interface design is only about "Easy" and not about "Productive". Sure a lot of things have improved, and some of the early interfases where jnot very good, but something was lost: Creativity. I design Software for fun so i tend to lean toward beeing creative even though it may mean that my applications are rejected by some. But then again: The average of what everybody wants nobody wants.
Aside from what you are doing today, we're interested in hearing about your ongoing involvement with the field of Computer Graphics, ACM SIGGRAPH, and the SIGGRAPH Conference.
What first drew you to Computer Graphics?
I've always loved two things in life: technology and film. CG happen to be smack down in the middle. CG in turn brought me to programming and games.
Can you name any mentors or strong influences in this area of your professional development?
On a technical side Emil Brink taught me to become a programmer. He has helped me a lot with his vast technical knowledge. Artistically i have many influences, Kubrik, Kraftwerk, Michel Gondry, Chris Cunningham, early Spielberg and the last few years Hendrix has become my house guru.
On thing i did early on was that I worked for a very sucsessful small media company runned by a bunch of compleat morons. They where very enthusiacstic but everything they tuched turned in to chaos, but some how they still made it in the end. That tought me that if they could make it i should realy go for what ever i wanted to do. It tought me that what ever faults I have I should try to improve but never let them stop me.
When did you first contribute to SIGGRAPH?
I have had a BOF for Verse the last 3 years.
What year and city was your first SIGGRAPH? Which was the most intense, and why?
SIGGRAPH 2000 in New Orleans, was my first and most intense SIGGRAPH. I managed to meet and befriend many great people like Jos Stam, Charles Loop, and Paul Debevec. I even got to meet my hero's from BUF fresh from fight club. It was 110 degrees and 100% humidity, the convention center was huge and the nights in the french quarter where long so it was a very draining experience. It took me weeks to recuperate.
I learned something very good about how to attend SIGGRAPH. Always stay after the lectures and talk to the speakers. they usually give away to good stuff about the work in progress after the session. If you are lucky you can get some good cards and meet some great people. The beginning of a lecturer is always about state of the art so usually you can skip that. Last year i attended a single session form end to end (It was the verse BOF i hosted) the rest of the time I tried to only attend the last 15 minutes of a session. The more you attend the less it becomes about lectures and the more it becomes a social experience.
What contributions to SIGGRAPH are you most proud of?
This year by far. We will have great booth, and some really great things to show. it will be in booth #1530 the open source pavilion.
Are you looking forward to or enamored by any recent or near-future developments in CG?
Something I'm very interested is is the area of image analysis. The composting tool haven't really changed that much since the original Discreet Flame came out over 10 years ago. With the advances in hardware, image tracking, recognition, matting, color correction, and lumigraphs I think it would be possible to create software that would revolutionize post production. It something I would very much like to work with in the future.
I know you are involved in ongoing development of the OpenGL specification, what other projects are or have you been involved in which may be significant to our readers?
I have a number of applications I want to write, A Paint tool for matte paintings concept design, an animation tool and Film editing tools. But i have lots to do right now and nobody is paying me so I have to figure out how to pay my rent too. OpenGL takes a lot of my time. My main project right now is a as of yet secret game project where I have been able to use verse as a content pipeline. Outside of CG I'm spending time writing a novel and some TV/Film Scripts.