Orcmid's Lair |
||||
|
2003-02-07Weekly Theory Puzzles. Some of the fun that theory-edge has been having lately is thanks to Vaughn Pratt's puzzles. I don't follow this closely, but it is marvelous to watch how a body of experience around Chu spaces and related topics is building up. And the puzzles are drawing out some high-powered lurkers! 2003-02-05ACM: Ubiquity - Genesis Of An Anthill: Wireless Technology And Self-Organizing Systems. Oh boy, a new Espen Andersen article in ACM Ubiquity. Data StructuresThis week's theme. There is fascinating material on sparse matrix approaches and also on frameworks for components that provide sparse matrix implementations, indexing and hashing implementations, and so on.&nbps; Interesting as a double-edged sword, too.Sparse matrix softwareCSE 373: Sparse Matrices Project. An assignment that was carried out in my neighborhood, linked by Nicola. There are some interesting speedup cases -- use of sparse matrices in operations such as multiplication by a constant, other constructs that can be sped up when taken as mass operations. Examples to think about.Guide and Reference: Sparse Matrix. Part of yet another library. Here is a nice illustration of the kinds of Sparse Matrices that are of importance in storing well. There seems to be a standard list of 7 types. From classmate Nicola Daize. Boost C++ Libraries. The overall project and all of the links to other activities, subprojects, etc. Boost Libraries. And the overall collection of libraries (found by slowly trimming the URL from the sparse matrix package). CLAPACK (f2c'ed version of LAPACK). Well, there is documentation to read here, because I have no clue what it is to f2c a version of LAPACK! Boost Basic Linear Algebra. The uBLAS Basic Linear Algebra C++ template that deals with BLAS levels of functionality and is being tested with CLAPACK. More things to find out about. Sparse Matrix. Hmm, an example of Sparse_Matrix template from the Boost library. More from David Read. This is a hairy lot of stuff. MTL: The Matrix Template Library. More interesting material and a library that provides high performance using the C++ Standard Template Library for its approach. Thanks again, David. I finally noticed that template libraries don't do what component frameworks do, and it will be interesting to see where and how component frameworks can come close to this. It might not matter to the serious number cruncher who is willing to compile a program that fits the job, but I would like to see what is possible using dynamic interfaces and bindings for component libraries. Just curious. Sparse Matrix Storage Formats: J.�Dongarra. A cataloging of storage formats for sparse matrices, depending on the kind of regularity that is known to exist in the matrix and how one intends to access the material (essentially). More from David Read. CIS353 Sparse Matrix Module. A Case Study in Layered Abstraction. That topic is dear to my heart, and we have a worked example here from Thomas Hain, at the University of South Alabama. Another find from David Read. IML+++: Iterative Methods Library in C++ . Yet more interesting material on sparse matrices and other fascinating things. There is a SparseLib++ too. This was found by classmate David Read. It has been a long (a very long) time since I did much heavy lifting with matrix manipulations and with various storage schemes such as queues, binary trees, dispatcher task control blocks, etc. It is a lot of fun to dig into it again, even if only for a part of a week. 2003-02-03Java Tutorial: Collections. Here's Joshua Bloch's tutorial on Collections. I need to look at how it all works. The Java[tm] Tutorial. This is really nice. I was pointed here by classmate Thomas Eichberger, as part of our discussions on sparse arrays (to flex our muscles in the unit on data structures we are going through). I think the collections tutorial is about what I want to understand, because it seems to show how to overload [ ] or how to define a class of object as something that is indexed, and I want to see how that happened. ITPACK 2C: A FORTRAN Package for Solving Large Sparse Linear Systems by Adaptive Accelerated Iterative Methods 1. Another of Martyn's citations on sparse matrices. Here, to ITPACK. Templates for Solution of Linear Systems: Contents. This citation was provided by classmate Martyn Atkkins as part of a discussion on storage structures for sparse matrices. There is a discussion of that here, along with a healthy bibliography. Bruce Eckel's Free Electronic Books. Bruce does his writing in public (but more nicely than I do) and makes electronic books freely available. The ones of interest for my studies are his Thinking in Java (now at edition 3r4.0 in press) and Thnking in Patters which he is still working on. I want to figure out this Pattern business, so this is perfect timing for me to be sent back here by a classmate. I will thank him. Bruce Eckel's MindView, Inc. Exceptional Learning Experiences. Ooh, I have been here before and I have great admiration for this site and the approach reflected here. I didn't recognize the name when classmate Thomas Eichberger recommended it the other day. I am a Bruce Eckel wannabe. Competitors Shape Strategy to Gain Edge in Web Services. A New York Times on-line article, on the jockying going on with regard to Web Services, how they will serve enterprises, and whose platforms they will operate on. There is an interesting dinstinction between middleware as a platform-neutralizing device and the .NET approach which is being more positioned as more of Windows. 2003-02-02Software EngineeringDavid Parnas materialsSoftware Engineering Research Group. The SERG at McMaster University. This is one of the places where David Parnas can be found. SERG has 8 faculty and 10 graduate students. 7 of the 10 are asian. The remaining 3 have East European surnames. I do not mention this for xenophobic purposes, but to marvel at the fact that we do not appear have US students with the will, determination, and preparation for successfully entering this program. There are 43 alumni of the program, and 28 of those have asian names. (It would appear that there is a strong tilt to Chinese participation in software engineering programs, based on the small sample that today's students represent.)Publications: David Lorge Parnas. The Parnas bibliography. It is remarkably similar to the one in the back of [Parnas#222], the [Hoffman2001] compilation with D. M. Weiss. There are differences, and I suspect these are owed to the fact that this a numbered list in HTML, so that prior insertion will upset the apple cart. It looks as though there have been some corrections but that the list is sacrosanct. Good idea, since it provides a clean citation device for Parnas papers. Dr. David Lorge Parnas. The Parnas homepage at McMaster University, furnished by classmate Puneet, who had classes in Software Engineering during his undergraduate studies, 10 years ago. The degree studies of Professor Parnas (P.Eng) were all at Carnegie Mellon university, and one might suppose this is where he came into contact with Alan Perlis, who Parnas identifies as one of the key influences in his career.
|
|||
