salec - Slashdot User

Submission + - Why Mobile Linux Is Better Than Windows (osweekly.com)

Submitted by techie on Tuesday October 09, 2007 @03:49AM

techie writes: OSWeekly.com has a new article up that compares Mobile Linux to Windows. The author claims that Mobile Linux is much better than Windows. Why? "On the desktop front, most customer built PCs designed with Windows in mind run great with the latest Linux distributions. But there are still some bumps in the road regarding the efforts behind successfully running Linux on notebooks designed for Windows. Yes, there are tremendous resources like linux-laptop.net to make this a bit easier, but in the end, just supporting the vendors that are trying to make compatible solutions with the OS that we have all come to love is one step closer to coming together as a community to support the cause. Think Linux hardware vendors like System76, Emperor Linux and others.

Submission + - The Sorry State of Desktop Innovation in OSes (osweekly.com)

Submitted by

OSer

on Tuesday October 09, 2007 @03:15AM

OSer writes: "OSWeekly.com's Brandon Watts discusses the stale desktop environments in existing operating systems. Sure, we are all used to and expect a certain way of doing things, but does that mean we should kill innovation? He continues, "I don't always agree with the "if it ain't broke, don't fix it" mentality that is so prevalent today. As a power user, I've grown relatively tired of these same old things. When new OS versions come out, they may be new, but they don't always feel new. I think it's time for Microsoft and Apple to really step it up and show us what the next generation of desktops are going to look like. They may be working on it, but I want to see the results as they currently stand. The next versions of Windows and OS X after Vista and Leopard really should bring us the goods, and if they don't, then I want an explanation."

Submission + - 41,000 PCs seek bird flu cure

Submitted by Stony Stevenson on Tuesday October 09, 2007 @02:26AM

Stony Stevenson writes: An international grid of more than 41,000 computers is offering new hope in the search for treatments for the deadly avian influenza virus. The Enabling Grids for E-science (EGEE) network links ordinary computers in 50 countries to form a single giant supercomputer with more than five million gigabytes of disk and tape storage.

Dr Ying-Ta Wu, a biologist at the Genomics Research Center of the Academia Sinica in Taiwan, explained that computing grids are the fastest and cheapest way to find promising new drugs that might be able to battle the virus. By rapidly sifting through a vast number of possible drug and virus interactions, the grid calculates the probability that a particular drug molecule will block key sites on the virus, thus inhibiting its action.

Submission + - Forget about OpenOffice, KDE readies KOffice 2.0 (computerworld.com.au)

Submitted by Da Massive on Tuesday October 09, 2007 @12:26AM

Da Massive writes: While the industry is distracted by the ongoing tussle between Microsoft and OpenOffice.org over document formats, the KDE project is quietly preparing the next generation of its own office suite, KOffice, which will run on for Linux, Mac OS X and... Windows. http://www.computerworld.com.au/index.php/id;1596080362

Submission + - Author suggests sanity in online piracy debate (baen.com) 1

Submitted by on Monday October 08, 2007 @11:51PM

An anonymous reader writes: I just stumbled upon the Baen Free Library. I've never heard of Eric Flint before, but he's apparently a Sci-Fi/Fantasy author with views on copyright that match the majority position here on Slashdot. The Baen Free Library is a place for authors published by Baen Books to offer free, online, full-text versions of their books as a means of advertising. (Available formats include online HTML, downloadable HTML, a few eBook formats I don't recognize, and RTF.) Baen Books seems to be doing with the Free Library exactly what many Slashdotters think the RIAA and MPAA should be doing with sound and video recordings — embracing the internet rather than fearing it. The only author I recognized on the authors list is Larry Niven, but I haven't read much fiction in a long time, so I'm probably out of touch.

Submission + - Mobile Phone Doubles the Risk of Cancer (expressen.se)

Submitted by

billzerker

on Monday October 08, 2007 @10:22PM

billzerker writes: "Mobile Phone Doubles the Risk of Cancer A Swedish news article discusses the results of a study to be published in the Journal of Occupational and Environmental Medicine. The researchers behind the study claim that the use of a cell phone for more than 10 years doubles the risk of brain cancer. Original article in Swedish is here: http://www.expressen.se/nyheter/1.873064/mobiltelefonen-fordubblar-cancerrisken Rough translation of the article below: Mobile Phone Doubles the Risk of Cancer Swedish researchers sound new cancer alarm. The mobile phone doubles the risk of tumors in the brain. "The study is statistically sound," says professor Lennart Hardell. The regular use of a mobile phone doubles the risk of cancer. It is clear from a meta-analysis compiled by researchers Lennart Hardell from University Hospital in Örebro and Kjell Hansson from Umeå University. "The use of a mobile phone for more than ten years doubles the risk of cancer. It is the radiation against the brain that is harmful," says Lennart Hardell. 16 Different Studies: The compilation has been published in the respected medical periodical, "Journal of Occupational and Environmental Medicine." The report draws from 16 different studies, of which eleven concern the research of mobile telephony for more than ten years. Lennart Hardell has been accused of being the "Swedish master of cancer alarm." But he rejects the criticism. "It is silliness. We have researched the subject for more than 30 years now. We have observed that there is a strong opposition to our research high up in society. It is such a big industry with so much money involved," says Hardell. What is a person to do, should one stop making calls with a mobile? "You should first and foremost buy a phone with a low level of radiation. You should also use a headset whenever possible," says Hardell."

Submission + - Citrix Opens Security Holes in Military, Fed Sites (eweek.com)

Submitted by eweekhickins on Monday October 08, 2007 @10:14PM

eweekhickins writes: Ubiquitous technology that chugs away underneath Web applications is being used to put up military and government GUIs with security holes you could drive a bus through. PDP says "It is 2007, people... it shouldn't be that simple."

Submission + - Fluxbox reaches 1.0

Submitted by

hollywoodb

on Monday October 08, 2007 @09:01PM

hollywoodb writes: "As Phoronix is reporting, the ever-popular Fluxbox window manager has finally reached 1.0. The official announcement is on Fluxbox's SourceForge project page, and more info is available at their homepage. This comes hot on the heels of their 1.0rc, which was announced in June of 2006.
Changes include a new default style, new alternate styles, fixed shape handling, a fluxbox-remote utility, updated languages, and of course more bugfixes."

Submission + - Beautiful Code: Leading Programmers Explain How Th

Submitted by

davidnicol

on Monday October 08, 2007 @08:11PM

davidnicol writes: "Beautiful Code is a collection of thirty-three essays by notable computer programmers, all dared by editors Greg Wilson and Andy Oram to hold forth about a piece of source code they have found notably beautiful, apparently with intention of producing what may become a standard computer science textbook. Such an experiment has not been carried out before.

Were Don Knuth dead, this book would probably be dedicated to his memory. Beautiful Code collects "cultural wisdom" of computer programmers. If one were to consider The Art of Computer Programming to be something of a Pentateuch, Beautiful Code could the accompanying Midrash. Having obtained an e-copy from O'Reilly and processed it with pdf2txt and then run grep -ci knuth Chapter*.txt, I found that not everyone mentioned Knuth. Those who did have a K number greater than zero by their name in the synopses below. (Asterisks before the author's name represent other chapters mentioning the author.)

Speaking of experiments, O'reilly is trying to develop an online community based on the book, with a wiki.

I had several programming projects prioritized ahead of writing this book review, but after leafing through the six hundred pages and reading a few of the pieces in it at random I realized that my professional work would benefit from making reading it cover to cover my primary free-time action item.

To be systematic and thorough, here are quick summaries of the essays, which are more prolix than the editor's version of the same exercise in the preface. Proceeding in this fashion is really horrible book review practice, but there really isn't any cleaner way to approach this particular text.

*** Brian Kernighan (K0) discusses Rob Pike's minimal regular expression engine optimized for pedagogy.

Karl Fogel (K0) tells the origin story of Subversion's "delta editor" internal interface, including an abandoned branch. He introduces a human-factors aspect to software beauty: by having a strongly defined interface, development time which might be squandered on debating interface points can be put to more productive use.

** Jon Bentley (K4), the author of Programming Pearls, holds forth about quicksort, instrumenting quicksort, and generating mathematical forumlae which give the same results. His essay nicely demonstrates the tension between procedural (pseudocode) and descriptive (mathematics) language.

Tim Bray (K0), writing in a tone suitable for a more general audience, uses Ruby's regexps to analyze his Apache server logs from his blog, which he encourages the reader to check out, to build a case for his suspicion that Google uses binary search.

* Elliotte Rusty Harold (K1) draws on experience writing XML verifiers to reinforce the truism that early passes at a problem should focus on correctness, saving speed for later.

Michael Feathers (K0) explains his appreciation for Ward Cunningham's FIT Java testing framework's noncompliance against Java coders' rules of thumb, revealing a tension between closed (final classes, designed-in hooks and extension points) and open (working stub base classes) approaches to designing for extensibility. FIT-derived systems take HTML system documentation with tests embedded in it in tables as input and fill in one table cell with a pass/fail result. By carefully selecting what to hold constant and deferring extensibility to authors of derived classes, FIT allows Java developers, jaded and wounded by corporate infighting, to appreciate the fundamental elegance of a solid, basic, O-O approach.

* Alberto Savoia (K1) plays fast and loose with math jargon. He discusses JUnit as an introduction to an exploration of how to test a binary search function for correctness. He performs a math proof — claiming that a conclusion follows from implications of known-true axioms — without calling it that, and a few pages later uses the word "induction" in the non-mathematical sense. He gives us a quick course in software testing best practices, including smoke testing, boundary testing, and what he calls "test theories" which are statements made about what the tested system is supposed to do, and some techniques for deriving additional test theories off the "happy path." Apparently 2006 was the year that data sets reached the size where the (a+b)/2 method of finding the midpoint between a and b started overflowing using signed 32-bit integers, which wreaked all kinds of weird havoc.

Charles Petzold (K0) uses and endorses dynamic code generation for efficient image processing, reminiscing about the feats of the team who wrote Windows 1.0 in 8086 assembly language and applying similar techniques in his C# code thanks to the System.Reflection.Emit namespace. The results aren't easy on the eyes, but they run in a quarter of the time of equivalent code that makes decisions inside loops. It is not clear if the MSIL compiler is going to be so brazen as to perform loop unrolling on bytecodes emanating from ILGenerator objects; if not, I expect that the optimal acceleration of the dynamic method would occur with the loops unrolled into a modified Duff's device large enough to use almost all of the CPU cache. That would remove many loop bounds comparisons, which aren't that many instructions compared with the operation being performed on each pixel, but which may necessitate CPU pipeline stalls. To save another two percent of the time, and possibly hurt general system performance during the operation. Small processes take less effort to switch between. So never mind. Unless you want to benchmark it for something to do.

Douglas Crockford (K1), who explains why LISP has never been accepted by the mainstream, and why the LISP community doesn't care, proposes to use a technique from a 1973 paper (anticipating object orientation, no less) by Vaughn Pratt ("Top Down Operator Precedence") that modified a technique published by Robert W. Floyd (for whom Knuth wrote a glowing eulogy) to write a parser that can process a subset of ECMA-262 (a.k.a. Javascript) entirely within that subset, which includes functions, objects, and JSON literals. It's all about the left binding power. And the "nud" and the "led," terms which when given to a search engine yield a slightly easier-going version of chapter nine. Null denotation and left denotation. Truthy and falsy. Don't repeat yourself, write a macro. A parser yields a parse tree, and hands that off to something else. A proposal for accommodating the needs of both language designers who want to reserve a lot of symbols and programmers who don't want to have to synonymize around them. Would extending Pratt's technique to support languages where not everything is a functional expression really be as easy as Crockford makes it look? Later in the book, R. Kent Dybvig (K0) explains how Scheme's syntax-case macro expander lets the programmer forget the whole deal about using quasi-reserved names in macros, because it abstracts them out for you, while offering an extra-special syntax for when you really truly do want to interfere with the internal operation of a macro. I am very pleased that O'Reilly sent me this book to review while I have been working on a macro system for Perl, and having read the Dybvig and Crockford articles, Macrame.pm will be better that it would have been had I not read these articles, when it finally does get completed.

Henry S. Warren, Jr. (K1) rants about a fundamental problem, the best way to count set bits. Stunningly, divide-and-conquer approaches using bit shifting beat optimized brute force approaches, although calling them opaque is an understatement. Population count processing is applicable to a wide variety of basic data processing infrastructure problems. Warren's elucidation of the situation is a perfect example of the best practice development process: correct and fast, in that order.

A USABILITY GURU: Ashish Gulhati (K0), who prefers to work remotely from the Himalayas, talks about how helping some social idealists inspired Cryptonite, his end-user-friendly encrypted webmail system, then launches into the history of the project, including architecture implementation details and discussion of security concerns and benchmarking, and how using well defined interfaces made replacing one component with a more efficient one with comparable functionality a relative breeze. Along with Lincoln Steins's chapter, this chapter is a nice introduction to object oriented Perl and how the reusability aspect of a good interface design effectively optimizes programmer time in any language.

BIOINFORMATICIANS: Lincoln Stein (K0), without encumbering his narrative with O-O terminology, presents his Bio::Graphics system for displaying genetics research results as a example of usabilty done right, this time usability by one's fellow programmers. The advantage of using standard interfaces becomes apparent when it became possible to support many graphics formats by drop-in replacement of the GD library, which outputs in only one format, with a different library with the same interface but configurability to output in other formats too. The tension between perfection and deliverability is discussed, as Stein reveals his big wishlist item for his library but opines that the level of rewriting required is expected to prevent its reprioritization, especially as there is an effective workaround. Immediately following that, Jim Kent (K0) tells us how readability and understandability lead to extensibility in the Gene Sorter, including a short tutorial on writing polymorphic objects in C.

MATH GEEKS: Adam Kolawa (K0) uses and endorses the CERN math library, which evolved into LAPACK. The well-commented core routines, in Fortran with bindings for use by other languages, are in regular use after three decades. Jack Dongarra and Piotr Luszczek (K0) discuss recent modifications made to LAPACK to take full advantage of modern hardware.

KERNEL HACKERS:Greg Kroah-Hartman (K0) appreciates the discipline involved in maintaining the Linux kernel, where the acrobats operate in full view of the world without the benefit of nets such as dynamic type checking; Diomidis Spinellis (K0) uses the BSD virtual file system as the springboard for presenting general truths about indirection and maintainability; and Bryan Cantrill (K0) likens software bugs to sewage, in particular an embarrassingly unforseen race condition in a released version of Solaris.

PYTHON MONGERS:* Andrew Kuchling (K0) discusses Python's hash tables, Travis E. Oliphant (K0) discusses the optimized n-dimensional iterator that lets the NumPy python math library do matrix operations without blowing its cache, and Rogerio Atem de Carvalho and Rafael Monnerat (K0) hype the flexibility of ERP5, an enterprise resource planning system that runs on python/zope.

ROBERT HEINLEIN FANS: Ronald Mak (K0) uses J2EE beans to communicate both with Mars rovers and their legions of fans who want to download terabytes.

PUTTING THE GOO IN GOOGLE: Jeffrey Dean and Sanjay Ghemawat (K0) discuss MapReduce. If you don't already know what MapReduce is, you probably know what to do to find out.

THE NETWORK IS THE COMPUTER: Simon Peyton Jones (K0) introduces STM (Software Transactional Memory) which is an allegedly fool-proof replacement for locking discipline in multithreaded Haskell, and then discusses his solution to Trono's "Santa Claus problem" without explaining why the problem is supposed to be difficult.

William R. Otte and Douglas C. Schmidt (K0) manage to abstract concurrency models out of their framework for networked services.

Andrew Patzer (K0) reveals some practical best practices as he tells the story of doing a straightfoward piece of systems integration using RESTful web services.

RADICAL USABILITY Andreas Zeller (K0), apparently a devotee of the Einsteinian craft-a-tool-to-straighten-the-bent-paperclip-from-the-good-paperclip-you-have-found methodology, uses an automated divide-and-conquer approach to discovering which of several thousand patches in the delta between two versions of gdb was responsible for a mysterious loss of function in the ddd graphical front end. Had ddd a debugging mode that logged the full interaction between itself and gdb, the world might not be the better for his automated technique.

* Yukihiro Matsumoto (K0), the author of Ruby, might be annoyed that his essay on the nature of elegance, including the DRY (Don't Repeat Yourself) principle and other similarly useful tips, was not selected as a forward instead of buried towards the end of the collection.

COMPUTER ACCESS FOR ALL: * T. V. Raman (K0) shares his amazement at how straightforwardly the extensible design of emacs Lisp, with its advice system that allows insertion of function before, after, or fully replacing any of the existing emacs functions, allowed his system to adapt emacs for eyes-free use to bush out over the extended development of emacspeak as a hobby project. Takeaways include the advice that when adapting a communication for another medium, it is important to communicate the message, not the artifacts of the previous medium. He is justifiably proud of his eyes-free calendar widget, among others. Bizarrely, there is no mention of Morse Code in Arun Mehta(K0)'s retrospective on a project to develop an improved system for the use of Stephen Hawking, who could only use one button (but he could control how long he pressed it, and could watch a screen as he did so.)

BEAUTIFUL CAN MEAN EASY ON THE EYES: Laura Wingerd and Christopher Seiwald (K0), from perforce, discuss readable layout, including selection of line break points, and reveal the strong correlation between bugs and nested ifs. (Takeaway: prefer case statements, even if it means repeating yourself)

CASTLES MADE OF SAND Brian Hayes (K0) demonstrates the truth to the saying "while in other disciplines the researchers stand on the shoulders of giants, in computer science we stand on each other's shoes" as he flails around in search of, and eventually finds, the easy way to find co-linearity of three points in order to analyze satellite photos of the lower Mississippi river, which tends to change course during high water.

IN CONCLUSION

I think if I had been asked to write a chapter for this book, I would talk about normalized Huffman tables, as discussed in Managing Gigabytes, which Hans Reiser recommended in response to my relating to him a conversation about ReiserFS had by the Kansas City LUG. Beautiful code has a lot of different things happening in it, with snippets of source code in various languages, and nice diagrams where appropriate. It was not a fast read, although some of the pieces do flow nicely, due to the necessity of retooling one's entire brain from one essay to the next.

Were I still a single nerd, and my girlfriend were to present me with Beautiful Code, perhaps as a birthday present, I would not hesitate to select her as my personal and permanent nerdess, all other things being equal.

David Nicol, inventor of the online tip jar and always looking for collaborators with spare tuits, can be contacted through the comment form at tipjar.com."

Submission + - Hillary Clinton Declares War on Space Exploration (associatedcontent.com)

Submitted by

MarkWhittington

on Monday October 08, 2007 @07:19PM

MarkWhittington writes: "Recently, Senator Hillary Clinton revealed her science agenda. Of great interest to people involved in making and debating space policy were the three bullets concerning the space program. To be brief and to the point, Hillary's agenda would be terrible news for anyone who supports space exploration and space commerce."

Submission + - Obama promises Cap-And-Trade plan (marketwatch.com)

Submitted by

FreddyKnockout

on Monday October 08, 2007 @05:10PM

FreddyKnockout writes: "From the article:

Democratic presidential contender Barack Obama on Monday said that if elected he would establish an economy-wide cap-and-trade program that would sharply cut greenhouse-gas emissions by 2050.
It's an interesting twist on a much-talked about proposal. In essence, he proposes providing X amount of credits which permit the release of CO2 and other greenhouse gases. The catch with his plan is there are no freebies, all credits must be purchased through an auction, unlike John Edwards plan of granting some credits and auctioning others. This would, according to Obama's campaign, "ensure that all polluters pay for every ton of emissions released.""

Submission + - BLAST - A science project you may be interested in (blastshare.com)

Submitted by

bloodybonnie

on Monday October 08, 2007 @04:22PM

bloodybonnie writes: "This cool web-based community allows users to participate in the creation of BLAST — a science documentary by director Paul Devlin. BLAST follows a close-knit international team of astrophysicists and graduate students as they attempt to launch a multi-million dollar telescope on a NASA high-altitude balloon — a journey which takes them from the Arctic to the Antarctic. Through the on-line community at Blast.ArtistShare.com, users can get a behind the scenes look at the creation of the film, while also supporting it's creation. Depending on the level of participation, users can have access to streaming video production updates, interactive Q&A sessions with the director and the film's science writer, tickets to the film's premiere, and even the opportunity to have your name appear in the credits of the film. There's also a chance to edit part of the film!"

Submission + - Elephants are terrified of ... bees

Submitted by Beenonymous on Monday October 08, 2007 @04:14PM

Beenonymous writes: According to this article, African farmers have a way to ward off badly behaved elephants that trample over their crops and property. Their new secret weapon? Bees. Researchers from the University of Oxford played recordings of bees to herds of elephants and watched them stampede in terror. Some of the experts speculate it may be that one of the bees has previously been stung up its trunk.

Submission + - SPAM: Three Energy Systems of the Body

Submitted by

rinkjustice

on Monday October 08, 2007 @02:58PM

rinkjustice writes: "Did you know the human body has three separate mechanisms for creating energy, depending on the intensity and duration of the activity? I call them 'Immediate Response', 'Fight' and 'Flight', and it's the protocol by which your body expends energy."
Link to Original Source

Submission + - Cyber Minimalist: How to Work (Almost) Completely (zenhabits.net)

Submitted by

Emerys

on Monday October 08, 2007 @02:37PM

Emerys writes: "Great advice for the multi-computer user. Tips on how to work online without having to install or save many programs to your computer."

Slashdot Top Deals