Systems problems are traditionally solved in either hardware or the operating system. Hardware is too expensive to experiment with, while OS-based solutions suffer from the chicken-and-egg problem of moving the world to your improved OS. The virtual machine monitor (VMM) presents a sweet spot between these two extremes. VMM-based solutions can solve a problem in software, once, and yet solve it for all popular operating systems and applications.

I'll be using a new, VMM-based tool called VProbes both to demonstrate the scope of what the VMM can intercept, and and as an example of a feature that would be difficult to implement in other layers. VProbes provides a globally coherent, safe, programmable, interactive window into a running machine, from its CPUs to memory to peripherals.

Quantum information processing presents fascinating new areas of research to computer science as well as other areas of science. This talk will overview the state of the art in building quantum computers, algorithms, languages, and compilers. This talk will also describe some of the challenging problems compiler writers face in trying to create compilers for quantum computers.

ZFS is an open-source file system that brings all the cool features of high-end storage to your laptop. Snapshots, clones, compression, encryption, RAID, remote replication, and end-to-end data integrity are all integrated and trivial to use. This talk will explain how ZFS works and what it can do for you.

Dr. Crispin Cowan, famous Linux security guy and vocal Microsoft critic, now works at Microsoft. What? Has Hell opened a ski resort? This talk will reflect on my first year at Microsoft as a Linux guy. Hell has not frozen over, and I'm having a great time. I will talk about how the Linux and Windows communities are more similar to each other than either community cares to admit, for good and bad, as well as highlighting the differences that I have found. I will also highlight the similarities and differences in the security problems faced by Windows and Linux.

While most of us are stuck asking "When will I finally get to go on a moon-hotel vacation?" or wondering "Just what it is that NASA does anyways?" a new approach to space travel is literally taking flight. Space 2.0, the commercialization of space, promises to finally make that hot zero-g dream of yours come true. This talk will discuss the current state of the space industry and survey some of the most interesting companies and projects in the Space 2.0 movement. As a do-it-yourself bonus, we'll also go throw an introduction to building a rocket and shapeship software development.

Electric Sheep is a cyborg mind composed of 60,000 computers and people connected by software and the internet. It's an open-source, distributed screen-saver that harnesses idle computers into a render farm with the purpose of animating and evolving artificial life-forms known as a "sheep". Dreams in High Fidelity is a painting that evolves. This limited edition fine art was made by Draves picking personal favorites from the screen-saver, and rerendering them in much higher quality. The presentation will include a demo of the artwork and an explanation of the technology.

The modern web browser has evolved from a relatively simple client application designed to display static data into a complex networked operating system tasked with managing many facets of online experience. Support for dynamic content, multimedia data, and third-party plug-ins greatly enriches the browsing experience at the cost of increased complexity of the browser itself, resulting in a plague of security vulnerabilities that provide hackers with easy access to systems. To address the root of this problem, we designed and implemented the OP web browser. We have partitioned the browser into smaller subsystems, isolated each subsystem, and made all communication between subsystems simple and explicit. Finally, we have used formal methods to prove the correctness of the communications between subsystems and the ability to limit the effects of compromised subsystems.

Sprinkled amongst both Generation X and Generation Y is a curious, new population. The Nerd. The Geek. A group which has spent the last two decades moving from the dark technological fringe to a mainstream demographic.

In this presentation, Rands will first explore the traits of the nerd. How is it that we're able to rampantly consume information? How many people do we really know? Why is it that we must solve puzzles? How is it that we're so good at context switching and so bad at relating to people? And is all this nerdery going to help or hinder my career? It's going to help.

In the second part of this presentation, Rands will explain the role of the nerd in high tech and discuss entry strategies into your favorite company. We're going to talk about simple ways to leverage your nerd tendencies to get you in the right door talking to the right people.

Roundy will introduce the strongly, statically typed language Haskell, and discuss some of the advantages of static typing, particularly in combination with type inference. He will introduce the concept of "type witnesses", a programming trick to enable the compiler to enforce correctness constraints far beyond what is gained from ordinary static typing. Along the way, Roundy will explain how darcs differs from its "competitors", and provide a brief introduction to the patch formalism it uses.

Q is a high-performance vector programming language developed by KX systems and is the language used to program the Kdb+ database. The Kdb+ database is designed to work with massive amounts of intraday data in memory and immutable historical data on disk. Its main use case has been to process real time tick feeds and provide fast analysis against years of historical tick data.

This talk will introduce the Q programming language and discuss some of the features that are making it increasingly popular at Morgan Stanley. I'll also talk about some of the challenges of introducing a new language into an enterprise environment.

Since the invention of the transparent animation "cel" in the early 1900's, animation has been a technology driven art form. In the 1990's, the transition from handdrawn 2D to computer generated 3D films brought with it a raft of complicated technical challenges. Travel with us down the production pipeline and explore some of the engineering problems encountered producing feature-length films at DreamWorks Animation. We'll discuss the data structures, algorithms, numerical methods, testing, configuration management and systems design required to create the "illusion of life".

The Facebook development Platform launched its first API over two years ago, in August 2006. Since its launch it has grown to support over 400,000 external developers developing thousands of social applications and has become the most popular social platform on the web.

This is the story of the technical evolution of Platform, from data assessors to a simple REST API to the full-fledged development platform that it is today. We will discuss the interesting challenges we faced in regards to user privacy, the development of the Facebook Mark-up Language (FBML) as well as the inherent performance issues involved in safeguarding the privacy of over +100 million users and how developers can assist us solving them. We'll also talk about the future of Platform and FBML with Facebook Connect and the rest of the open web.

In a well-known cartoon, Randall Munroe characterizes computational linguistics as a field "so ill-defined that" practitioners "can subscribe to any of dozens of contradictory models and still be taken seriously." As with any humor, there is a serious point here. Indeed, it is often quite hard to see what the substantive claims of computational linguistics are, and how one's model could be falsified (other than by presenting another model that has a lower error rate).

Part of the problem is that a large amount of what counts as computational linguistics is essentially natural language engineering, where the point is simply to find something that works. And part of the problem is a widespread view that has computational linguistics as little more than applied machine learning. Nonetheless, there are models and problems that serve as unifying themes in the field. Among these are the widespread use of finite-state methods, recent work on grammar induction and the growing field of computational modeling of evolutionary language change. Each of these is interesting not only as a computational problem, but also because of the light it may shed on human language.

We are all being terrorized on a daily basis. We make bad decisions in the face of poorly defined but vigorously promulgated fears. And, as developers, we end up creating worse code as a result.

Let's stop being reactive. And let's stop being afraid. And let's create some wonderful, liberated, code.

Robots have been around for about half a century. Until about a decade ago, robotic arms were strictly confined to caged robotic work cells to protect the safety of people who might otherwise be harmed by high robotic joint torques/forces. The past decade has seen a couple of trends where a cage for safety is no longer practical. The first trend is robotic mobility. It's not practical to build a fixed cage around a mobile robot, considering, for example, Vecna's BEAR robot, designed to rescue wounded soldiers from an active battlefield, or the iRobot Roomba robotic vacuum cleaner.

The second trend is a growing reliance on robots that are necessarily in direct contact with people, such as medical patients during robotically-assisted surgery or rehabilitation. For example, Barrett's robotic WAM arm reduces the invasiveness of knee-implant surgery to improve outcome and shorten hospital recovery time. But robot surgery requires direct physical contact between robot and patient, again challenging the role of a cage. The talk will cover these examples and others that show the recent trends away from traditional fixed robotic work cells in cages. It will also show innovative ways to improve safety as these trends continue.

It has been said that close only counts in horseshoes and hand grenades--as long as you count nukes as a kind of hand grenade. But what exactly are you aiming at, and how close is close enough? This talk will be about the past, present, and future of Perl, with a few pointers on how to launch your project, how to steer it while you still can, and when to let it fly off on its own trajectory.