From dreyer at tti-c.org Sat Jan 27 10:23:00 2007 From: dreyer at tti-c.org (Derek Dreyer) Date: Sat Jan 27 04:25:38 2007 Subject: [Hyper] Module type systems Message-ID: <7fa251b70701270823s41c31b37led960199207c1a19@mail.gmail.com> Due to popular demand (by which I mean that two people requested it), I will help guide the HYPER group through some papers on module type systems (mostly mine, but with excerpts of related work). To start out, we will look at my work with Crary and Harper on module type systems, which extends and generalizes a lot of work from various people (MacQueen, Harper, Lillibridge, Stone, Leroy, Russo, etc.). Eventually I want to work up to recursive modules and Flatt-style units, and a discussion of my recent work, which exposes a really cool hybrid of the Definition and Harper-Stone (i.e. "type-theoretic") semantics of modules. This can be appreciated competely orthogonally from the type-inference-and-modules business that has a tendency to set Dave's pants on fire whenever I talk about it. :-) So for next Wednesday, please read my POPL03 paper: "A Type System for Higher-Order Modules" by Derek Dreyer, Karl Crary, and Robert Harper. http://tti-c.org/dreyer/papers/thoms/popl.pdf This paper essentially gives a unifying type system within which one can understand a variety of design choices in the ML module system, and that's a useful contribution. However, there are two problems with it IMO: (1) it assumes you already know why you would want to support, say, both applicative and generative functors (this assumption was sort of necessary in order to fit in all the material, but still...), and (2) I find our informal motivation for various concepts in our type system to be unconvincing, although some people (like Andreas Rossberg and Ken Shan) have told me they like it. I attempted to address these two questions in the first two chapters of my thesis, which I wrote from scratch and are completely informal. I am not going to ask you to read those chapters in addition to my POPL paper, but I will suggest that you print out at least chapter 1 for reference during the HYPER meeting. (If you want, you can read it, too...it's very high-level, easy to read.) At the meeting, I will talk about some of the material presented in Chapters 1 and 2 of my thesis that is not in the POPL paper. There's too much material here to present if you have not read the POPL paper and come ready with questions, so please do. Chapter 1 of my thesis (please print and maybe read if you want): http://ttic.uchicago.edu/~dreyer/thesis/chapter1.pdf My whole thesis (if you're interested): http://ttic.uchicago.edu/~dreyer/thesis/main.pdf Derek From dreyer at tti-c.org Sat Jan 27 10:51:21 2007 From: dreyer at tti-c.org (Derek Dreyer) Date: Sat Jan 27 04:53:54 2007 Subject: [Hyper] change of plan Message-ID: <7fa251b70701270851o1536cef9p86fa0ae60c3c4a0@mail.gmail.com> Hi, everyone. After taking a closer look at chapter 1 of my thesis, I realized that there's actually a lot of useful background material here that will make reading the POPL03 paper *much* easier, and there's no way we're going to be able to cover all the informal ideas and formal details of the POPL03 paper in one week anyway. So CHANGE OF PLAN -- for this first week, please just read Chapter 1 of my thesis: http://tti-c.org/dreyer/thesis/chapter1.pdf This is an easy read (the first 5 or 6 pages should be total review for any ML programmers), and it's IMO the best 14 pages I've ever written. :-) We'll save the POPL03 paper for the following week, and then we'll move to recursive modules for the third and fourth week. (If this is too much modules, we can move to something else after the third week, but I think four weeks should be just enough time to present a coherent explanation of the state of the art.) Derek From dreyer at tti-c.org Sat Jan 27 12:02:49 2007 From: dreyer at tti-c.org (Derek Dreyer) Date: Sat Jan 27 06:05:23 2007 Subject: [Hyper] Hyper meeting time? Message-ID: <7fa251b70701271002k29140d50k20a265126fe284e4@mail.gmail.com> Would people mind moving the Hyper meeting time up to 11:00 or 11:30 on Wednesdays instead of noon? I've gotten one request for an earlier time, and I would prefer it as well. Let me know your preference. Derek From jhr at cs.uchicago.edu Sat Jan 27 12:13:43 2007 From: jhr at cs.uchicago.edu (John Reppy) Date: Sat Jan 27 06:16:30 2007 Subject: [Hyper] Hyper meeting time? In-Reply-To: <7fa251b70701271002k29140d50k20a265126fe284e4@mail.gmail.com> References: <7fa251b70701271002k29140d50k20a265126fe284e4@mail.gmail.com> Message-ID: <7304366A-7D04-4BB3-979B-4577D4B5347B@cs.uchicago.edu> The manticore group is meeting at 11am on wednesdays. - John On Jan 27, 2007, at 12:02 PM, Derek Dreyer wrote: > Would people mind moving the Hyper meeting time up to 11:00 or 11:30 > on Wednesdays instead of noon? I've gotten one request for an earlier > time, and I would prefer it as well. Let me know your preference. > > Derek > _______________________________________________ > Hyper mailing list > Hyper@ttic.uchicago.edu > http://ttic.uchicago.edu/mailman/listinfo/hyper > From dreyer at tti-c.org Sat Jan 27 12:20:47 2007 From: dreyer at tti-c.org (Derek Dreyer) Date: Sat Jan 27 06:23:20 2007 Subject: [Hyper] Hyper meeting time? In-Reply-To: <7304366A-7D04-4BB3-979B-4577D4B5347B@cs.uchicago.edu> References: <7fa251b70701271002k29140d50k20a265126fe284e4@mail.gmail.com> <7304366A-7D04-4BB3-979B-4577D4B5347B@cs.uchicago.edu> Message-ID: <7fa251b70701271020k5e0e143dnf03f408883de5d3c@mail.gmail.com> Is this a recent change or was the Manticore meeting at Wed @ 11 last quarter? I'm guessing it will be annoying to have back-to-back meetings, esp. if the Manticore meeting runs over. But we can keep it at Wednesday at noon if people want to. Maybe Monday or Tuesday at 11 would be better? Derek On 1/27/07, John Reppy wrote: > The manticore group is meeting at 11am on wednesdays. > > - John > > On Jan 27, 2007, at 12:02 PM, Derek Dreyer wrote: > > > Would people mind moving the Hyper meeting time up to 11:00 or 11:30 > > on Wednesdays instead of noon? I've gotten one request for an earlier > > time, and I would prefer it as well. Let me know your preference. > > > > Derek > > _______________________________________________ > > Hyper mailing list > > Hyper@ttic.uchicago.edu > > http://ttic.uchicago.edu/mailman/listinfo/hyper > > > > From fluet at tti-c.org Sat Jan 27 12:21:27 2007 From: fluet at tti-c.org (Matthew Fluet) Date: Sat Jan 27 06:24:19 2007 Subject: [Hyper] Hyper meeting time? In-Reply-To: <7fa251b70701271002k29140d50k20a265126fe284e4@mail.gmail.com> References: <7fa251b70701271002k29140d50k20a265126fe284e4@mail.gmail.com> Message-ID: <45BB9827.6000301@tti-c.org> The Manticore weekly meeting is 11:00 on Wednesdays, so I'm against moving Hyper earlier on Wednesday; lunchtime (or earlier) on Tuesday or Thursday would be fine. -Matthew Derek Dreyer wrote: > Would people mind moving the Hyper meeting time up to 11:00 or 11:30 > on Wednesdays instead of noon? I've gotten one request for an earlier > time, and I would prefer it as well. Let me know your preference. > > Derek > _______________________________________________ > Hyper mailing list > Hyper@ttic.uchicago.edu > http://ttic.uchicago.edu/mailman/listinfo/hyper > > > From hammer at tti-c.org Sat Jan 27 15:05:51 2007 From: hammer at tti-c.org (Matthew Hammer) Date: Sat Jan 27 09:08:23 2007 Subject: Fwd: [Hyper] Hyper meeting time? In-Reply-To: References: <7fa251b70701271002k29140d50k20a265126fe284e4@mail.gmail.com> <45BB9827.6000301@tti-c.org> Message-ID: Tue or Thur at noon seem good to me too. For those who don't know, there's a TTI-C calendar online: http://ttic.uchicago.edu/cal/month.php I guess the most notable potential conflicts for people here are the "general seminars" that seem to occur on T/Th at 10am, as well as the faculty lunches at noon on Mondays. -matt On 1/27/07, Matthew Fluet wrote: > The Manticore weekly meeting is 11:00 on Wednesdays, so I'm against > moving Hyper earlier on Wednesday; lunchtime (or earlier) on Tuesday or > Thursday would be fine. > -Matthew > > Derek Dreyer wrote: > > Would people mind moving the Hyper meeting time up to 11:00 or 11:30 > > on Wednesdays instead of noon? I've gotten one request for an earlier > > time, and I would prefer it as well. Let me know your preference. > > > > Derek > > _______________________________________________ > > Hyper mailing list > > Hyper@ttic.uchicago.edu > > http://ttic.uchicago.edu/mailman/listinfo/hyper > > > > > > > > _______________________________________________ > Hyper mailing list > Hyper@ttic.uchicago.edu > http://ttic.uchicago.edu/mailman/listinfo/hyper > From amal at tti-c.org Sun Jan 28 16:07:16 2007 From: amal at tti-c.org (Amal Ahmed) Date: Sun Jan 28 10:09:46 2007 Subject: [Hyper] Hyper meeting time? Message-ID: <1170022036.13990@whale.he.net> How about moving Hyper to 11am on Tuesdays? The only potential conflict (that I know of) is the TTI Theory Seminar which is Tuesdays at noon. Is this a problem for anyone? -Amal > Tue or Thur at noon seem good to me too. > > For those who don't know, there's a TTI-C calendar online: > http://ttic.uchicago.edu/cal/month.php > > I guess the most notable potential conflicts for people here are the > "general seminars" that seem to occur on T/Th at 10am, as well as the > faculty lunches at noon on Mondays. > > -matt > > On 1/27/07, Matthew Fluet wrote: > > The Manticore weekly meeting is 11:00 on Wednesdays, so I'm against > > moving Hyper earlier on Wednesday; lunchtime (or earlier) on Tuesday or > > Thursday would be fine. > > -Matthew > > > > Derek Dreyer wrote: > > > Would people mind moving the Hyper meeting time up to 11:00 or 11:30 > > > on Wednesdays instead of noon? I've gotten one request for an earlier > > > time, and I would prefer it as well. Let me know your preference. > > > > > > Derek > > > _______________________________________________ > > > Hyper mailing list > > > Hyper@ttic.uchicago.edu > > > http://ttic.uchicago.edu/mailman/listinfo/hyper > > > > > > > > > > > > > _______________________________________________ > > Hyper mailing list > > Hyper@ttic.uchicago.edu > > http://ttic.uchicago.edu/mailman/listinfo/hyper > > > _______________________________________________ > Hyper mailing list > Hyper@ttic.uchicago.edu > http://ttic.uchicago.edu/mailman/listinfo/hyper > > From amal at tti-c.org Mon Jan 29 14:34:42 2007 From: amal at tti-c.org (Amal Ahmed) Date: Mon Jan 29 14:34:46 2007 Subject: [Hyper] Hyper meeting time? Message-ID: <1170102882.27425@whale.he.net> Hyper will meet at 11:30am on Tuesdays this quarter. The first meeting will be tomorrow -- fortunately the first chapter of Derek's thesis doesn't take long to read. We don't yet have a room for Hyper. I've tentatively asked about getting a room here at TTI, but if people would like to meet at Ryerson, could someone look into reserving a room there. Thanks, Amal From jhr at cs.uchicago.edu Mon Jan 29 16:29:16 2007 From: jhr at cs.uchicago.edu (John Reppy) Date: Mon Jan 29 16:29:11 2007 Subject: [Hyper] Hyper meeting time? In-Reply-To: <1170102882.27425@whale.he.net> References: <1170102882.27425@whale.he.net> Message-ID: <72FB6038-B9DF-4131-ABFA-5204705E3905@cs.uchicago.edu> I think that we should continue to meet at Ryerson, since PL lunch is at TTI-C. I'll try to reserve 255 again. - John On Jan 29, 2007, at 2:34 PM, Amal Ahmed wrote: > > > Hyper will meet at 11:30am on Tuesdays this quarter. The first > meeting will be tomorrow -- > fortunately the first chapter of Derek's thesis doesn't take long > to read. > > We don't yet have a room for Hyper. I've tentatively asked about > getting a room here at TTI, > but if people would like to meet at Ryerson, could someone look > into reserving a room there. > > Thanks, > Amal > > > _______________________________________________ > Hyper mailing list > Hyper@ttic.uchicago.edu > http://ttic.uchicago.edu/mailman/listinfo/hyper > From dreyer at tti-c.org Mon Jan 29 17:15:45 2007 From: dreyer at tti-c.org (Derek Dreyer) Date: Mon Jan 29 17:15:48 2007 Subject: [Hyper] Hyper meeting time? In-Reply-To: <72FB6038-B9DF-4131-ABFA-5204705E3905@cs.uchicago.edu> References: <1170102882.27425@whale.he.net> <72FB6038-B9DF-4131-ABFA-5204705E3905@cs.uchicago.edu> Message-ID: <7fa251b70701291515w5ceb94f4rff5cc3c47bf80a7@mail.gmail.com> That's fine with me. If people would prefer to have the meeting at noon instead of 11:30 (so as to continue the tradition of eating lunch during the first 10 minutes), that's also OK with me. But I will let someone else be "the decider". ;-) Derek On 1/29/07, John Reppy wrote: > I think that we should continue to meet at Ryerson, since PL lunch is > at TTI-C. I'll > try to reserve 255 again. > > - John > > On Jan 29, 2007, at 2:34 PM, Amal Ahmed wrote: > > > > > > > Hyper will meet at 11:30am on Tuesdays this quarter. The first > > meeting will be tomorrow -- > > fortunately the first chapter of Derek's thesis doesn't take long > > to read. > > > > We don't yet have a room for Hyper. I've tentatively asked about > > getting a room here at TTI, > > but if people would like to meet at Ryerson, could someone look > > into reserving a room there. > > > > Thanks, > > Amal > > > > > > _______________________________________________ > > Hyper mailing list > > Hyper@ttic.uchicago.edu > > http://ttic.uchicago.edu/mailman/listinfo/hyper > > > > _______________________________________________ > Hyper mailing list > Hyper@ttic.uchicago.edu > http://ttic.uchicago.edu/mailman/listinfo/hyper > From adrassi at uchicago.edu Mon Jan 29 18:03:18 2007 From: adrassi at uchicago.edu (Aaron Turon) Date: Mon Jan 29 18:03:20 2007 Subject: Fwd: [Hyper] Hyper meeting time? In-Reply-To: References: <1170102882.27425@whale.he.net> <72FB6038-B9DF-4131-ABFA-5204705E3905@cs.uchicago.edu> <7fa251b70701291515w5ceb94f4rff5cc3c47bf80a7@mail.gmail.com> Message-ID: I would advocate the opposite move, going to 11:00, because I have class at noon. However, I'm not sure whether I'll have time to come every week anyway, so don't move it on my account. Aaron On 1/29/07, Derek Dreyer wrote: > That's fine with me. If people would prefer to have the meeting at > noon instead of 11:30 (so as to continue the tradition of eating lunch > during the first 10 minutes), that's also OK with me. > > But I will let someone else be "the decider". ;-) > > Derek > > On 1/29/07, John Reppy wrote: > > I think that we should continue to meet at Ryerson, since PL lunch is > > at TTI-C. I'll > > try to reserve 255 again. > > > > - John > > > > On Jan 29, 2007, at 2:34 PM, Amal Ahmed wrote: > > > > > > > > > > > Hyper will meet at 11:30am on Tuesdays this quarter. The first > > > meeting will be tomorrow -- > > > fortunately the first chapter of Derek's thesis doesn't take long > > > to read. > > > > > > We don't yet have a room for Hyper. I've tentatively asked about > > > getting a room here at TTI, > > > but if people would like to meet at Ryerson, could someone look > > > into reserving a room there. > > > > > > Thanks, > > > Amal > > > > > > > > > _______________________________________________ > > > Hyper mailing list > > > Hyper@ttic.uchicago.edu > > > http://ttic.uchicago.edu/mailman/listinfo/hyper > > > > > > > _______________________________________________ > > Hyper mailing list > > Hyper@ttic.uchicago.edu > > http://ttic.uchicago.edu/mailman/listinfo/hyper > > > _______________________________________________ > Hyper mailing list > Hyper@ttic.uchicago.edu > http://ttic.uchicago.edu/mailman/listinfo/hyper > -- Aaron Turon adrassi@gmail.com http://www.cs.uchicago.edu/~adrassi/ -- Aaron Turon adrassi@gmail.com http://www.cs.uchicago.edu/~adrassi/ From dreyer at tti-c.org Mon Jan 29 18:07:40 2007 From: dreyer at tti-c.org (Derek Dreyer) Date: Mon Jan 29 18:07:49 2007 Subject: [Hyper] Hyper meeting time? In-Reply-To: References: <1170102882.27425@whale.he.net> <72FB6038-B9DF-4131-ABFA-5204705E3905@cs.uchicago.edu> <7fa251b70701291515w5ceb94f4rff5cc3c47bf80a7@mail.gmail.com> Message-ID: <7fa251b70701291607v2e88448bgf7ded3f3f4dc0ebd@mail.gmail.com> 11:00 is no good for TTI folks because we will be having job candidate talks at 10:00 AM (on Tuesdays among other days), or so I'm told. Derek On 1/29/07, Aaron Turon wrote: > I would advocate the opposite move, going to 11:00, because I have > class at noon. However, I'm not sure whether I'll have time to come > every week anyway, so don't move it on my account. > > Aaron > > On 1/29/07, Derek Dreyer wrote: > > That's fine with me. If people would prefer to have the meeting at > > noon instead of 11:30 (so as to continue the tradition of eating lunch > > during the first 10 minutes), that's also OK with me. > > > > But I will let someone else be "the decider". ;-) > > > > Derek > > > > On 1/29/07, John Reppy wrote: > > > I think that we should continue to meet at Ryerson, since PL lunch is > > > at TTI-C. I'll > > > try to reserve 255 again. > > > > > > - John > > > > > > On Jan 29, 2007, at 2:34 PM, Amal Ahmed wrote: > > > > > > > > > > > > > > > Hyper will meet at 11:30am on Tuesdays this quarter. The first > > > > meeting will be tomorrow -- > > > > fortunately the first chapter of Derek's thesis doesn't take long > > > > to read. > > > > > > > > We don't yet have a room for Hyper. I've tentatively asked about > > > > getting a room here at TTI, > > > > but if people would like to meet at Ryerson, could someone look > > > > into reserving a room there. > > > > > > > > Thanks, > > > > Amal > > > > > > > > > > > > _______________________________________________ > > > > Hyper mailing list > > > > Hyper@ttic.uchicago.edu > > > > http://ttic.uchicago.edu/mailman/listinfo/hyper > > > > > > > > > > _______________________________________________ > > > Hyper mailing list > > > Hyper@ttic.uchicago.edu > > > http://ttic.uchicago.edu/mailman/listinfo/hyper > > > > > _______________________________________________ > > Hyper mailing list > > Hyper@ttic.uchicago.edu > > http://ttic.uchicago.edu/mailman/listinfo/hyper > > > > > -- > > Aaron Turon > adrassi@gmail.com > http://www.cs.uchicago.edu/~adrassi/ > > > -- > > Aaron Turon > adrassi@gmail.com > http://www.cs.uchicago.edu/~adrassi/ > _______________________________________________ > Hyper mailing list > Hyper@ttic.uchicago.edu > http://ttic.uchicago.edu/mailman/listinfo/hyper > From jhr at cs.uchicago.edu Tue Jan 30 09:06:18 2007 From: jhr at cs.uchicago.edu (John Reppy) Date: Tue Jan 30 09:06:21 2007 Subject: [Hyper] room for HYPER meetings Message-ID: <25DFB779-9636-4F0B-9045-8E3E40A6679D@cs.uchicago.edu> We have Ry255 reserved from 11:30-1pm on tuesdays (including today). - John From dreyer at tti-c.org Wed Jan 31 12:28:28 2007 From: dreyer at tti-c.org (Derek Dreyer) Date: Wed Jan 31 12:28:33 2007 Subject: [Hyper] Fwd: [pop-group] Fwd: Letter from TJ Watson In-Reply-To: <6.2.1.2.2.20070131131846.07862680@imap.srv.cs.cmu.edu> References: <6.2.1.2.2.20070131131846.07862680@imap.srv.cs.cmu.edu> Message-ID: <7fa251b70701311028n5234139eycb90cb007a0a927f@mail.gmail.com> ---------- Forwarded message ---------- From: Jeannette Wing Date: Jan 31, 2007 12:20 PM Subject: [pop-group] Fwd: Letter from TJ Watson To: cs-students@cs.cmu.edu, pop-group@cs.cmu.edu, software-group@cs.cmu.edu, specification-verification-center@cs.cmu.edu A request from one of our PhD alum, Vijay Saraswat: He's looking for summer interns and full-time people. ===== X10: Programming parallel machines, productively (http://x10.sf.net ) ============================ IBM Research is looking for top students interested in Programming Languages and Software Engineering, for both summer internships and full-time positions. Our summer internship program provides students with a unique opportunity to experience the research environment that exists here at IBM. The program is highly competitive, and is designed for top students who are interested in tackling challenging research problems. Many of our students have authored papers, or even theses, out of their sojourn with us, and many have contributed significantly to open source projects. What makes IBM Research unique is our proximity to the entire software life cycle. IBM is a leader, not only in the software development industry, but in the deployment, testing, and maintenance of this software at major customer sites throughout the world. We are constantly exposed to real world problems that help raise fundamental research challenges and opportunities, often ones that the rest of the research community is not yet aware of. Our department does leading research in a number of areas, all centered on making program development and maintenance more productive, and making the software produced more efficient, more manageable, and less failure-prone at runtime. A few representative examples of active research in these areas include: 1. Programming Languages ? Modeling/Design (a) Cross-Language Integration: Many programs are written in more than one language. Reasons for this include reuse of legacy libraries, access to platform functionality, and efficiency. IBM research investigates language designs for cross-language integration, aiming at programmer productivity, static and dynamic error checking, portability, and efficiency. Contact: Martin Hirzel (hirzel@us.ibm.com) (b) Programming Models for Networked Applications: Increasingly, application software is being organized around collections of loosely-coupled distributed components, often communicating over the internet. Such components typically transform and query persistent data, communicate asynchronously, manage complex failure modes, and interact across multiple administrative domains. Current programming languages are ill-suited for this domain because they were designed for monolithic applications. We are designing new programming models to better support this emerging class of internet-scale applications through the use of abstractions that are easy to reason about and compose in a distributed environment. Contact: John Field (jfield@us.ibm.com) (c) PL-Database Integration: The goal of the Data Access Language Integration (DALI) project is to facilitate the development of database and service-oriented applications, in particular with respect to better integration of program and relational data. We are developing static analyses that bridge an application with the databases it accesses to allow for example: optimization of SQL-style database queries such as those supported by JDBC and LINQ; slicing (impact) analysis between program variables and database values. These capabilities can be applied to code refactoring, program understanding, coverage tools for testing, performance tuning, etc. Contact: Mukund Raghavachari (raghavac@us.ibm.com) (d) Real-time Java: IBM Research's work in real time garbage collection has created an entire new market, allowing Java to be used in application domains where it was previously considered impossible. Current projects include time-portable programming models, real-time visualization tools, dynamic compilation to reconfigurable hardware, and a lock-free, massively scalable, autonomic real-time garbage collector. We are also building real-time systems in Java: a music synthesizer and a helicopter. More information: http://www.research.ibm.com/metronome. Contact: David Bacon (bacon@us.ibm.com) (e) Virtual Machine Performance: Higher-level and managed languages make software development easier, but introduce serious performance challenges. IBM Research has the unique opportunity to develop novel solutions to these challenges, and watch these solutions have global impact in IBM's production JVMs. Contact: Michael Hind (hindm@us.ibm.com) 2. Programming Tools (a) Parallel Tools: The trend towards multi-core and multi-threaded architectures calls for a new generation of tools to enable a larger class of programmers to effectively develop parallel programs. The goal of the Parallel Tools Project is to develop advanced parallel programming tools in an integrated Eclipse environment that include developments tools, performance tools and tools for the detection of common concurrency related errors. Contact: Evelyn Duesterwald (duester@us.ibm.com). (b) Refactoring and Program Transformation: The work on refactoring at IBM Research has resulted in significant advances of the state-of-the-art, and in the contribution of several advanced refactorings to Eclipse. Current research efforts are focused on refactorings for advanced language features (e.g., concurrency), and supporting program analysis and infrastructure. Contact: Frank Tip (ftip@us.ibm.com) 3. Software Engineering (a) Advanced Test Case Generation: High quality test suites are necessary to validate that applications meet their requirements with the required level of quality IBM Research is developing a radically new approach to automating the creation of test suites using advanced modeling techniques and new model analysis and test planning algorithms. Contact: Amit Paradkar (paradkar@us.ibm.com) (b) Software Architecture and Requirements: Agile methods have taken off of late. However, the realities of globalization mean that application development will remain a distributed activity. We are examining how better tools and methods around architecture and requirements can help distributed teams to capture, share, manipulate, understand, and manage software engineering artifacts. Can we help teams separated by time and space to obtain the benefits of agile methods? Contact: Steve Abrams (sabrams@us.ibm.com). 4. Software Quality/ Analysis/Verification (a) Analyzing Non-Functional Properties: The complexity of applications consisting of multiple components and frameworks has been steadily increasing. The resulting performance and memory footprint problems are often addressed too late in the development process, as an afterthought. Several projects at Watson have made significant progress automating the discovery of problems like memory leaks, performance bugs, and general footprint analysis. However, there is a long way to go: How can these problems be addressed earlier in the development cycle? How can non-functional requirements be incorporated and enforced in the development process? What can be learned from existing practices, to improve future practice? Contact: Edith Schonberg (ediths@us.ibm.com). (b) Deep Program Analysis for Software Quality: We are applying deep program analysis to early bug detection, design/specification recovery, and concurrency and resource management, with the goal of improving software quality and productivity. The central research problem in this area is to scale deep analyses to large applications, while providing precise and focused feedback to application developers. We are addressing this challenge through projects exploring a variety of analysis techniques and application domains. Contact: John Field (jfield@us.ibm.com). (c) Language-based Security: We are working on a number of projects whose purpose is to enforce application security through programming language constructs and/or program analysis. The problems addressed include design and development of automated tools for access control and information flow security enforcement. Contact: John Field (jfield@us.ibm.com). (d) Visualizing the Execution of Distributed Applications: A currently popular approach for reducing cost and complexity is building applications by composing services accessible over a network (SOA). A key challenge in this paradigm is obtaining an understanding of the overall activity triggered by the operation of a composite application. This information is critical for debugging, testing and understanding resource utilization of such applications. Our research is directed towards the creation of new tools, methodologies and environments that provide a deeper understanding of SOA applications. Contact: John Morar (morar@us.ibm.com) 5. Usability: (a) Human Centric Tools: Our tools, applications and services are becoming more and more interconnected and complex. We are looking at ways to make these tools usable and simplify our environments. This ranges from new programming models for web components to end user programming, to User Centered Design methodologies for these new environments, to the pragmatics of large scale SOA implementations, to glancable interfaces and visualizations. Contact: Juerg von Kaenel (jvk@us.ibm.com) 6. Virtualization: (a) Virtual Appliance Computing: Virtual appliances (live software state encapsulated in virtual machine sandboxes) are rapidly replacing install images as the unit of software packaging, distribution and management. This project is exploring ways to use the sandbox as a hook through which external tools can monitor, analyze and manipulate virtual state, and enable radically different paradigms for working with software. Contact: Vas Bala (vbala@us.ibm.com) 7. Web 2.0 / Semantic Web (a) Desktop Widgets, Situational Applications and End User Mashups: Web 2.0 technologies are rapidly reshaping the user's experience on the World Wide Web. A new, post-windows, GUI model is emerging in the form of desktop widgets like those offered by Apple, Yahoo, Google, Mozilla and others. The evolution of the web user experience from hypertext browsing to dashboards of ATOM-driven "live" widgets that exploit multiple, large format LCD monitors and other ambient displays is fundamentally changing the way we interact with information, services and communities over the Internet. In this research effort we are exploring this new world in the context of End User Programming. What must be done to empower end users without training in programming to create, deploy, share and enhance widget-based mashups and other situational applications? Contact: Sam Adams (ssadams@us.ibm.com). (b) Scalable Ontology Reasoning: The use of ontologies and need for semantic retrieval is increasing rapidly. However, the inability of reasoners to scale to large knowledge bases is a major inhibitor. We have been working on promising new technology for reasoning over very large knowledge bases in secondary storage, with excellent initial results. Our technology is based on aggressive summarization. We are interested in making semantic retrieval practical and looking at real use cases to validate our work. Contact: Edith Schonberg (ediths@us.ibm.com). Students interested in summer jobs in programming languages and software engineering at the T.J. Watson Research Center (NY area) should send a resume to the appropriate contact listed above or to Aditya Kalyanpur, our summer internship coordinator (adityakal@us.ibm.com), and also formally apply on-line at the general IBM Research jobs site: http://www.research.ibm.com/about/career.shtml For more information about Programming Languages and Software Engineering Projects at IBM Research, see http://www.research.ibm.com/compsci/plansoft/index.html