Download Research Tools
Performance, architecture, execution, bugs, and programs: these words are heard time and again in the context of a major computer science conference. So it was in Beijing this month at PLDI 2012, the conference on Programming Language Design and Implementation. Terminology and accompanying innovative ideas flew fast and furious as 600 academics, researchers from industry, and students gathered to discuss the latest advances in this fundamental field. PLDI is organized by ACM SIGPLAN (the Association of Computing Machinery’s Special Interest Group on Programming Languages).
Pictured from left to right: Lolan Song, Feng Zhao, Jan Vitek, Haibo Lin, and Judith Bishop
Although this was the thirty-third PLDI conference, it was the first to be held in Asia. Microsoft Research was proud to be the Gold Sponsor of the conference, and to celebrate the occasion, we organized a special open house to introduce participants to the work that goes on in our labs. Dr. Feng Zhao welcomed visitors to the beautiful Microsoft Research Asia lab with an overview of the work of the Beijing-based facility. The participants were then invited to a “DemoFest,” where 14 projects were on display: 13 from Microsoft Research labs around the world, and one from Tsinghua University. Many of these demos showcased the latest ideas about concurrency and the cloud, as well as the benefits of program analysis. There were also several projects illustrating end-user programming, such as TouchDevelop from Redmond and ClippyScript from Asia. See the full list of demos.
With hundreds of computer science academics and students gathered together, PLDI presented a great opportunity to engage in discussion of the hands-on work of writing the compilers and creating the tools that make today’s glitzy devices and snazzy apps possible. Modern platforms and applications demand highly sophisticated optimizing compilers and analysis tools, and the advent of new processor technologies, such as multiple cores, GPUs, and mobile platforms, along with the increasing sophistication of development tools, all require mastery of cutting-edge compiler and code generation technologies. PLDI was the ideal place to connect with students who are drawn to such computer science specialties as hardware specific optimizations, whole program analysis, profile framework and profile driven optimization, working set optimization, static alias analysis, optimized code debugging, incremental re-compilation, register allocation, code security, or SIMD and GPU code generation, vectorization, and parallelization.
Sriram Rajamani explains his poster.
Aside from presenting demos and connecting with friends old and new, Microsoft Research personnel also presented six papers and a tutorial that exemplify the high quality of our research. In addition, Microsoft researcher Rustan Leino and his colleagues were honored for having presented the most influential paper 10 years ago at PLDI 2002. That paper marked a turning point in the field of static checking, describing pragmatic design decisions that promote practicality over completeness. The techniques are now also widely used in various forms in Microsoft’s development tools—notably as part of Code Contracts, which ships with Microsoft Visual Studio.
I greatly appreciated the assistance of our colleagues at Microsoft Research Asia in handling all of the local logistics, especially Lolan Song's team, as well as Stewart Tansley from my team.
With so much science in my head, a quiet walk out to the Beijing’s Olympic Park was a great way to unwind. Beijing and PLDI certainly have a lot to offer!
—Judith Bishop, Director of Computer Science, Microsoft Research Connections
A growing trend in both the theory and practice of programming is the interaction with rich information spaces. This trend derives from the ever-increasing need to integrate programming with large, heterogeneous, connected, richly structured, streaming, evolving, or probabilistic information sources—be they databases, web services, or large‐scale, cloud‐based data analyses. However, as the complexity of programs and information structures increases, the coupling between the two is far from seamless, requiring many manual programming and modeling efforts. These manual processes often lead to brittle programs and thwart the easy application of novel compiler technologies and novel information mastering methods.
Fortunately, the Semantic Web provides rich means for ad‐hoc information structuring with querying and type-inference possibilities, while novel programming languages, like LINQ and F#, lower the entry bar to the information-rich world for the developer. In addition, innovative information mastering methods, such as Hadoop and Dryad, are frequently positioned as functional paradigms, and huge potential exists to combine information‐rich sources with both scalable and traditional programming models.
These approaches were on display this May at the Spring Mindswap in St. Goar, Germany, where researchers from around the world examined information spaces from the perspective of programming, looking for fresh insights into the promise and challenges of the design and applicability of the Semantic Web and new data-representation techniques. This workshop described the state of the art, elucidated the challenges that are required to bridge the gap between current information management and current programming language technology, and delineated concrete ways by which providers of information spaces can better serve the needs of programming languages, and vice‐versa. Of particular interest were the breakout sessions on three critical issues: (1) the handling of data versus schema, (2) the effect of information-rich programming on types in programming languages, and (3) the need to consider data quality. We would like to extend our thanks to Professor Steffen Staab of the University of Koblenz-Landau, who was the primary organizer of the Spring Mindswap.
Now, we want to invite the community to extend these discussions at the First Workshop on Programming the Semantic Web, which will be offered as part of the International Semantic Web Conference in Boston this November. In particular, we invite the submission of papers that discuss and promote the programming facet of the Semantic Web. Abstracts should be submitted by July 24, with papers due by July 31; further submission information can be found on the workshop website.
—Evelyne Viegas, Director of Semantic Computing, Microsoft Research, and Don Syme, Principal Researcher, Microsoft Research Cambridge
What are the big challenges and hot trends in computer science research? How are the academic community and Microsoft Research working collaboratively to use computing to solve some of the world’s most intractable problems? On July 16 and 17, 400 elite academic investigators will explore these questions with Microsoft researchers during the annual Microsoft Research Faculty Summit in Redmond, Washington.
But you don’t have to be in Redmond to benefit from this outstanding event. Selected keynotes and panel discussions will be streamed live from the Microsoft Conference Center, and engaging, informative live interviews with top researchers will be broadcast from Microsoft Studios. You can tune in to the live, streaming broadcasts from 9:00 A.M. to 1:30 P.M. Pacific Time (12:00 P.M. to 4:30 P.M. Eastern Time) on the Virtual Event page. And don’t miss the special closing keynote from David Breashears, “Rivers of Ice: Vanishing Glaciers of the Greater Himalaya,” on July 17 at 4:30 P.M. Pacific Time (7:30 P.M. Eastern Time).
Please join us as we explore trends in data-intensive, data-driven research—what we like to call “big data, big insights”—and as we probe the growing movement toward blending virtual and physical reality through advances in natural user interface. Learn about developments in social media, Internet governance, and the use of technology to combat criminal activity. And see how technology is impacting teaching and the creation of rich interactive narratives. What’s more, you can participate by tweeting your questions and comments during the live broadcasts by using the Twitter hashtag #FacSumm.
The Microsoft Research Faculty Summit is dedicated to expanding the boundaries of using technological development to solve real-world problems, whether social or scientific. From harnessing the power of data for analysis and insights, to algorithms for managing election data and detecting malware, to future digital homes and natural user interfaces, software is experiencing rapid change. The 2012 Microsoft Research Faculty Summit unites academic researchers, educators, and Microsoft researchers, product group engineers, and software architects to explore these and other new opportunities and challenges in computer science research—and you can be part of this exciting event via the live, streaming broadcasts.
So mark your calendar and clean your display screen: the Microsoft Research Faculty Summit 2012 is headed to a device near you.
—Harold Javid, Director, The Americas, Microsoft Research Connections