EDITOR'S EYE

The World of Software Development.

September 2006

Okay, I know I all but promised not to write about grids again until the cows came home. But the following isn't my fault. I didn't receive the award--Aaron Walsh did.

The award in question is the "Teaching with New Media Award" at Boston College. Aaron received the award for his "exemplary uses of technology in teaching"--specifically for his work with the Media Grid, a public utility for digital media.

Why I care about this is that Aaron first wrote about the Media Grid in an article in Dr. Dobb's Journal a year or so ago. In that article, he described the Media Grid as:

... a digital media network infrastructure and software-development platform based on new and emerging distributed computational grid technology. The Media Grid...is designed as an on-demand public computing utility that software programs and web sites can access for digital content delivery (graphics, video, animations, movies, music, games, and so forth), storage, and media processing services (such as data visualization and simulation, medical image sharpening and enhancement, motion picture scene rendering, special effects, media transformations and compositing, and other digital media manipulation capabilities). As an open platform that provides digital media delivery, storage, and processing services, the Media Grid's foundation rests on Internet, web, and grid standards. By combining relevant standards from these fields with new and unique capabilities, the Media Grid provides a novel software-development platform designed specifically for networked applications that produce and consume large quantities of digital media.

Since then, Aaron has cranked up his efforts with the Media Grid, focusing on it as an immersive education learning system that uses 3D technology and digital media to bring distance learning to a new level. Unlike traditional online courses, which involve the delivery of simple Web pages or streaming video, immersive education combines interactive virtual reality and sophisticated digital media with collaborative online course environments and classrooms.

So in a way, I'm not writing about grid computing at all, but instead acknowledging the efforts of a creative individual who has used technology to make the world a little better place. Congratulations Aaron.

Posted by Jon Erickson at 06:16 PM  Permalink |

What with all the chatter about Web 2.0, LAMP, Ruby, and the like, it is sometimes easy to forget that in the real world stuff like COBOL--yes, COBOL--still exists.

That's why it is interesting, although probably not that significant, when EvolveWare, a vendor that builds tools for automating the analysis, documentation, and transformation of applications and databases, pats itself on the back about migrating 325,000 lines of COBOL code to .NET in less than 10 months.

According to EvolveWare, the company did this with using its of S2T (Source-2-Target) Migration Tools for COBOL and Adabas/Natural that provided an automated knowledge discovery rate of 100 percent and a transformation rate above 85 percent (whatever that means). There was no explanation of whether EvolveWare did this for a client, or that it just happened to have time on their hands and 325,000 lines of legacy code laying around.

In the scope of things, however, 325,000 lines of COBOL code is probably a drop in the bit-bucket. According to some reports, even Microsoft said a few years ago that 1 million a year of new lines of COBOL were still being added to legacy code bases. In truth, no one knows how many lines of COBOL code are out there; let's just say "a lot." In any event, lines of code is far from the best metric discussing programming languages.

The one thing that always comes to mind when the topic of COBOL pops up, is something Al Stevens, a senior contributing editor to Dr. Dobb's Journal, once said (and I paraphrase): "You'll never hear me say anything bad about COBOL. It put food on my table for a lot of years." 'Nuf said.

Posted by Jon Erickson at 12:17 PM  Permalink |

Really, I'm not that enamored with grid computing. But it is so darn interesting.

So yesterday I mentioned how the Open Science Grid benefited from a $30 million largess. In today's grid news, CERN reports that the Enabling Grids for E-sciencE (EGEE) project has been able to sustain more than 30,000 jobs a day -- over 1,000,000 per month -- for a period of six months this year.

These computing tasks range from simulations of molecular drug docking for neglected diseases to geophysical analysis of oil and gas fields. Clusters of hundreds and even thousands of PCs around the world, have been executing these calculations -- in total over 25,000 central processor units (CPUs) are involved. Several million gigabytes of data storage in disk and tape facilities also contribute to make EGEE the world’s largest scientific Grid infrastructure.

An offshoot of the EGEE project is the Grid middleware distribution called gLite, which ensures the seamless operation of this global computing facility. In addition to scientific applications, EGEE has targeted a range of business applications for support, including financial analysis. Recently, successful demonstrations have been made of interoperation with other major national and international Grids, such as the Open Science Grid in the U.S. and NAREGIin Japan.

Okay, that's it for grids, for the time being--unless, of course, something grid-worthy happens in the meantime.

Posted by Jon Erickson at 11:41 AM  Permalink |

Grid computing and open science both got a financial shot-in-the-arm with the announcement that the National Science Foundation (NSF) and the Department of Energy's Office of Science are funding a five-year, $30 million program to operate and expand upon the two-year-old national Open Science Grid.

This project taps into the power of thousands of processors distributed across more than 30 participating universities and federal research labs. Among the universities participating are the universities of Wisconsin-Madison, Boston, Columbia, Cornell, and Indiana ; the California Institute of Technology; and the universities of Florida, California-San Diego, Chicago, Iowa, and North Carolina. Federal research centers include Fermi, Brookhaven National Lab, Lawrence Berkeley National Lab, and the Stanford Linear Accelerator Center.

Each member of the OSG consortium is supplying computing power to the shared facility, which runs the gamut from small clusters of computers to thousands of networked processors. Some of the OSG sites offer petabytes (1000 terabytes) of data storage. That level of power is especially useful for large-scale science in particle physics and genomics.

According to UW-Madison computer scientist Miron Livny, a principal investigator of the project, "Grid computing ... has the capability to revolutionize research, but the tools remain challenging for many scientists. Projects such as OSG are working to lower the barrier to individual scientists using advanced computing."

"Grid computing ... has the capability to revolutionize research, but the tools remain challenging for many scientists," Livny goes on to say. "Projects such as OSG are working to lower the barrier to individual scientists using advanced computing."

Livny has participated in two major distributed computing projects at UW-Madison. The Condor Project, as described in this Dr. Dobb's Journal article, uses a network of more than 3500 processors on the UW-Madison campus to support high-throughput computing challenges, while the Grid Laboratory of Wisconsin(GLOW) is a campus-wide facility deployed to serve the Wisconsin scientific community, providing more than 30,000 computer hours a day to researchers.

Posted by Jon Erickson at 09:31 AM  Permalink |

To me, it sounds a little like Monty Python and the search for the Holy Grail. Nevertheless, IBM and the University of Georgia have launched a research initiative to find the answer to the question of what is key to innovation -- great inventions, business models, technology and profit.

IBM has donated $80,000 in systems and software to the University of Georgia’s Terry College of Businesswhere students will work together with IBM Research and Market Intelligence to delve into the topic of innovation. Under the direction of Dr. Srinivas Reddy, who directs the University of Georgia’s Terry College’s Coca-Cola Center for Marketing Studies, students will be provided with case studies from IBM and examine IBM accomplishments in the form of patents, awards, and products that have made an impact by either increasing revenue, saving money or by gaining brand name recognition. The students will also research the steps that have been taken to turn innovations into successful sellers.

Besides studying innovations that have led to actual products, the team will also study a wide range statistics and data from industries other than technology -- such as pharmaceuticals. They will analyze a company’s successes and attempts at marketing certain innovations. The findings will help researchers develop their guidelines and will help them figure out how innovations can affect a company’s profits.

Students will provide IBM with a set of guidelines on how it has come to develop innovation that matters to both the company and its clients. The team will develop the "roadmap" by researching how IBM has found creative, cost-efficient, and convenient solutions and inventions for customers in the past. In addition, a published report will document how certain business practices can turn into profitable innovations

I look forward to reading how products like, say, IBM's PCjr, fit into the "innovation roadmap." Now that really will be an accomplishment.

Posted by Jon Erickson at 09:45 AM  Permalink |

I've just got to get in on some of this grant stuff. You know, grants for projects like the $1,000,000 payout for a Virtual Reality Spray Paint Simulator System to Pine Technical College or the $2,000,000 for the Virginia Community College System web portal.

By these measures, the University of California-Davis's National Science Foundation grant for $750,000 to study how open source software is built seems almost reasonable. The researchers will focus on the Apache Web server, PostgreSQL database, and Python. Team members, led by Professor Premkumar Devanbu, will collect information from the message boards, bug reports, and e-mail discussions to understand how design teams organize themselves and interact.

According to the proposal, open source software defies conventional wisdom about collaborative projects. For example, most office workers know that the slowest member of the team sets the pace for everybody else. But in open source projects, work moves at the speed of the fastest member of the team, and adding more hands speeds things up rather than slowing them down, Devanbu said.

Devanbu and colleagues think that the way teams are organized will be reflected in the resulting software. At the same time, the structure of the software will itself have an effect on how teams of programmers are put together. For example, software that is broken into large chunks of code might need a different approach than a structure of smaller chunks.

This grant seems to be to Devanbu's "Open Source Immigration Modeling" project .

It might be worth mentioning that what is really being studied isn't "open source" per se, but "collaborative software development" which may or may not have anything to do with open source. But when you're filling out the forms to people who write the checks, "open source" has a better ring that "collaborative development" I suppose.

In any event, my real problem with this grant is that I didn't think of it first. Maybe next time.

Posted by Jon Erickson at 09:29 AM  Permalink |

One of my dreams has always been to win a MacArthur Foundation "genius grant." Of course, to win the grant you first have to be a genius, which means that you have to be able to pass classes like calculus and physics, among others. In other words, you have to be smart. Sigh.

The second criteria for winning a MacArthur Foundation grant is that you have to have good ideas; that is, applying your "genius" to something (hopefully) practical. By both measures, Luis von Ahn, a Carnegie Mellon University computer scientist, is deserving of one of this year's genius grants.

What qualified von Ahn as a MacArthur recipient is an online, multi-player game he created that also makes the Internet more accessible to the visually impaired. The game, called Phetch, is an Internet scavenger hunt in which players use a search engine to look for images that fit certain descriptions. In the process, they produce and verify captions for unlabeled images from the Web. These captions can then be used to enhance the Web-browsing experience of blind people.

Phetch is one of several "games with a purpose" that von Ahn has developed. The first such game, The ESP Game produced key words for images that could be used to aid image searches. Peekaboom, produces images with objects labeled and highlighted in a way that could be used to train computer vision systems..

Phetch, which von Ahn developed with students Shiry Ginosar, Mihir Kedia, and Ruoran Liu, is designed for three to five players. The narrator writes a description of an image that has been randomly retrieved from a set of 1 million images from the Web; only the narrator can see the image. The other players ("searchers") use a special browser to search for it within that set of a million images. Each round lasts five minutes. Narrators receive points for each successful search and lose points if they pass on describing images believed to be too difficult. The first seeker to find each image receives points and becomes the narrator for the next round. In pilots, players spend an average of 32 minutes with the game and some have played for 10 hours or more in a single session. von Ahn and crew calculate that 5000 people could produce explanatory descriptions of all of the images indexed by Google in just 10 months.

Congratulations to Professor von Ahn for a job well done. And as for my chance at a genius grant, well, there's always next year.

Posted by Jon Erickson at 09:02 AM  Permalink |

Every now and then, you stumble across a software system that you never think about. Such is the case with a Linux-powered helicopter simulator being developed by Mitsubishi Heavy Industries. And no, you can't run it on your PlayStation 2.

The SH-60K Patrol Helicopter Team Trainer (PHTT) simulator will serve as a training platform for the new SH-60K patrol helicopter that Mitsubishi is building for the Japanese Maritime Self-Defense Force. The SH-60K will be carried onboarddestroyers and equipped to carry out missions ranging from anti-submarine warfare and anti-surface warfare, to surveillance, transport, and rescue operations. Now how cool would it be to sit in that simulated cockpit for a while?

At the heart of the simulator is Concurrent's RedHawk implementation of real-time Linux and its iHawk high-performance PCI-based hardware platform for time-critical data acquisition, simulation, and industrial systems applications. iHawk symmetric multiprocessors (SMP) feature from one to eight Intel Xeon or AMD Opteron CPUs and up to 32 GB of memory. One factor which led Mitsubishi to settle on Concurrent was the requirement to capture RGB graphics at extremely high rates. RedHawk Linux guarantees that a user application can respond to an external event in less than 30 microseconds on a dedicated processor. RedHawk Linux is compatible with the Red Hat distribution.

Now the real question is: Will it be out for either XBox 360 or PlayStation 3 in time for Christmas?

Posted by Jon Erickson at 10:13 AM  Permalink |

Scheme has always had a special place in the Dr. Dobb's annals. (That's "Scheme," as in a minimalist programming language that's a dialect of Lisp, not "scheme" as in figuring out how to get spare change from your pocket to mine.)

Maybe it's because XScheme, a popular and elegant implementation of Scheme, was written by former Dr. Dobb's technical editor David Betz. Of course, XScheme was only one of several languages David contributed to the open source community years ago. There was also XLisp, Bob, DROOL, AdSys, and others. But it's XScheme which stands out in my mind. If you want to find out more about XScheme, you can download the source code to it here.

Or maybe it is because of the conversations and interaction that Dr. Dobb's has had over years with Scheme's co-author Guy L. Steele Jr. Developed by Steele and Gerald Jay Sussman in the 1970s, Scheme was introduced via a series of now-famous papers known as the "Lambda Papers." You can also glean a lot about Scheme in particular, and programming languages in general, from Guy Steele's article "Thoughts on Language Design" and an interview entitled "A Conversation with Guy Steele". I'd also like to point out that Guy Steele Jr., in addition to being a Sun Fellow, Distinguished Engineer, and Principal Investigator at Sun Microsystems Labs, is a recipient of Dr. Dobb's Excellence in Programming Award.

The point of all this is that a new draft of the Scheme standard has been released by the Scheme workshop. After four or five years in the making, R6RS is now available for your review and public comment. The six-month public comment period is now open and continues until March 15, 2007. Congratulations to the steering committee and all others who worked hard to get this draft out. A job well done.

Posted by Jon Erickson at 09:42 AM  Permalink |

At the risk of becoming embroiled in the morass of discussions about U.S. immigration policy, I ran across an interesting study conducted by the Society for Research in Child Development.

In a nutshell, the study reveals that children of immigrants are more likely to pursue math and science in college than students from the same ethnic groups whose families have been in the U.S. for generations. Additionally, the study finds that the pursuit of math and science is not isolated to one immigrant group, but exists for children of Latino, Afro-Caribbean, and European immigrants. Clearly, children of immigrants pursue math and science as pathways to upward mobility.

According to Vivian Tseng, author of the survey and a program officer at the William T. Grant Foundation in New York City, one in five children in the U.S. has an immigrant parent. Tseng surveyed almost 800 college students between the ages of 18 and 25 about their aspirations for the future. She also collected data on their majors from the large, urban university they attended. She defined students from immigrant families and children of immigrants as those with at least one immigrant parent. She compared those children to students born in the U.S. who also had two parents born in the US.

According to Tseng, one reason these children of immigrants are more likely to pursue math and science than their peers whose families have been in the U.S. for generations is that the students she surveyed had higher economic aspirations and were aiming for better paying occupations than their later-generation peers. "These findings complement previous studies by other researchers," she said. "In interviews, immigrant parents, especially those working in low-wage, low-status jobs, channel their greatest hopes for upward mobility in this new country to their children. They tell their children that they must do well in school so they can have better lives and more satisfying, better paying, and higher status jobs than their parents."

Tseng's study suggests that children of immigrants fare well in ways that are important for the U.S. economy. "At a time when the U.S. economy is facing demands for highly educated workers in technology and science," said Tseng, "children of immigrants may well contribute to our nation's changing workforce needs."

I don't know what the answer is to the immigration challenges the U.S. faces, but I can tell you that education for all benefits all. And when it comes to technology, a strong background math and science is fundamental. That's not rocket science.

Posted by Jon Erickson at 10:18 AM  Permalink |

According to Ambrose Bierce, history is an account, mostly false, of events, mostly unimportant, which are brought about by rulers, mostly knaves, and soldiers, mostly fools.

Okay, but that doesn't mean that the recently released video The Microprocessor Chronicles isn't worth watching. The four-hour documentary, which chronicles the history of microprocessors in general and Intel CPUs in particular, was co-produced by Stanford University Libraries, Walker Research Associates, and Panalta Inc., and answers questions such as "Why didn't Intel patent the microprocessor?", "How did Intel convince IBM, Compaq, HP and others to place 'Intel Inside' on their computers?", and "How was Intel's microprocessor selected for IBM's personal computer?".

"The microprocessor has become truly ubiquitous," says Rob Walker, an engineer who conducted many of the interviews forming the backbone of the documentary. "Today everyone in the West owns dozens of these tiny computers. Yet the microprocessor has only been around since 1971, and has been ubiquitous only in the last decade. This program provides the context for that phenomenal growth by examining in depth the technology, business and personal stories of the pioneers."

Based in part on this oral history project, The Microprocessor Chronicles includes commentary and interviews with the likes of Gordon Moore, Ted Hoff, Stan Mazor, and Federico Faggin tell the story of the first commercially available microprocessors, the MCS-4 and then the 8008. The DVD delves into the competition between Intel and firms as Motorola, Zilog, and AMD. It also explores the advantages/disadvantages of RISC architecture, which, compared to its predecessor, the CISC architecture implemented by Intel which uses simpler but faster instructions to execute the same task. Researchers at IBM, the University of California-Berkeley, and Stanford produced the first RISC microprocessors. In the final chapter, Stanford President John Hennessy, who led Stanford's RISC research, speculates on the next advances in microprocessors.

All profits from the $49.95 DVD go to Stanford to support continued research and chronicling of the history of the semiconductor industry through the Silicon Valley Archives Project, an oral history of the Silicon Valley.

Of course from G.K. Chesterton's perspective, history is only a confused heap of facts--a position that Bierce would likely agree with, albeit grudgingly (Bitter Bierce didn't like to agree with anyone, anytime, about anything). But thanks to projects like The Microprocessor Chronicles, Bierce and Chesterton would be eating crow, if they were still around.

Posted by Jon Erickson at 04:38 AM  Permalink |

Since there are awards for just about everything else, my guess is that there are annual awards for press releases. Multiple categories, of course, ranging from the stupid to the sublime.

Don't get me wrong. Good press releases are, in fact, very useful and it takes real talent to write good ones. There needs to be lots of hard information. Enough use of buzz words to catch your attention. A minimum of self-serving congratulations. Good quotes. Who to contact. One page in length. That kind of stuff. In all likelihood, there's a checklist or forumula for writing them and, assuming that press release writers might stumble across this missive, there's every chance someone might share that forumla with us. If I had my druthers, like Dragnet's Jack Friday, I'd ask for "just the facts ma'am." But that's probably asking for too much.

For instance, I have to hand it to IBM's press release writers who generally do a good job. Certainly in a recent spate of press releases they hit, if not home runs, at least doubles and triples in the headlines alone. "Reduces Costs, Improves Access". Not bad. "Breakthrough Technology". Good. "Diverse, Distributed Environments." Suggests hard information to come. "Grid Computing Ecosystem." Ditto.

However, once you get into the press releases themselves, it's time to go to work, separately that chaff from that wheat. "... first-of-its-kind encryption technology and services that deliver the world's first enterprise-class solutions for securing consumer and corporate data privacy." "...unsurpassed levels of security" and "history-making." On the edge of over the top, but acceptable.

Other press releases also make use of the superlative. "New software," "transform," "dynamic," "more efficient," "less power and space,"reduces complexity" and the like.

When all the wheat and chaff are separated, what it comes down to is that IBM made some interesting announcements. The company announced:

• It is offering encryption on tape drives (yes, tape drives are still used by a lot of enterprises for a lot of good reasons). According to IBM, early measurements show no appreciable degradation to performance during the reading and writing of encrypted data. Encryption in the drive also allows data compression, reducing potential impact on the media, and the encryption-enabled tape drive can also process non-encrypted workloads.
• New software that lets clients move their existing technology infrastructures into dynamic, virtualized environments for use in grid computing environments by extending the capabilities of its Tivoli Workload Scheduler to virtual computing environments -- automatically clustering, scheduling and managing different types of workloads across a virtualized environment.
• A Grid Ecosystem program to help IBM Grid Business Partners in building skills, offerings and opportunities related to grid computings.

All in all, this is pretty interesting stuff and the press releases don't necessarily stand in the way of figuring this out. That alone probably is worth an award or two.

Posted by Jon Erickson at 08:56 AM  Permalink |

My how time flies. Although it was before my time, here we are at the 50th anniversary of the hard-disk drive. Yes, it was on September 4, 1956 that IBM released the first commercial hard disk--the IBM 350 RAMAC disk drive.

The 350 RAMAC (short for "Random Access Method of Accounting and Control") had about 5 MBs of storage on its 50 24-inch diameter disks with 100 recording surfaces. Each surface had 100 tracks and the disks spun at 1200 RPM with a data transfer rate of about 8800 characters per second. The system was housed in a cabinet that was 60x68x29 inches and was leased for about $36,000+/year (or $7000 per megabyte per year).

In constrast, today's typical hard-disk drives will fit in your fob (okay, that's the official name for the little watch pocket in my Levi's), provide 300 GB+ of storage, and spin at 7200 RPM or more. Oh yes, and today's hard-disk drives sell for under $100.

The nice thing about anniversaries is that you get to reminisce and prattle on about your own experiences. So bear with me.

My first hard disk was a Winchester Radio Shack TRS-80 drive that was about 24x24x6 inches in size. It housed about four or five 8-inch platters that provided 8 MBs of storage. The master drive sold for $3500, but you could daisy-chain slave drives ($2500 each) for additional capabilities. To tell the truth, I could barely pick up the behemoth. However, I had to at times, since there was a little hole on the undercarriage of the case that gave you access to a locking mechanism. You see, the hard-disk platters had to be locked down prior to moving the drive, or you risked damaging the unit.

Of course, I couldn't afford the monster even if I had wanted it. The reason I even had it was that I was writing the service manuals for technicians who had to install and repair it, and another for users who were suckered into buying it. Then there was the software manual--the drive would only work with TRS-80 Model II computers running TRS-DOS with a modified operating system. And as you might expect, the disk drive didn't outsell remote control cars in Radio Shack stores that holiday season.

So the next time you pop a 4-GB iPod into your pocket, just be glad it doesn't use a 350 RAMAC.

Posted by Jon Erickson at 09:44 AM  Permalink |

Believe it or not, .NET 3.0 is barreling down the tracks at top speed. Jeez, it seems like just yesterday that .NET 2.0 appeared on the scene.

Of course, this doesn't mean that .NET 2.0 is going away. Quite the contrary. .NET 3.0 relies on .NET 2.0, it doesn't replace it. What .NET 3.0 does replace--in name only--is WinFX. Putting this another way, .NET 3.0 is the next-generation Windows development API that will be released with Windows Vista. When is that? Your guess is as good as mine, but probably sometime early next year.

So you're probably wondering how I came to be such an expert on .NET 3.0.

That's easy. Yesterday afternoon I spent an hour or so chatting with Juval Lowy, an acknowledged expert on .NET 3.0, among other Windows-related subjects. Our conversation focused exclusively on .NET 3.0 and was, well, very enlightening and highly fascinating. Even entertaining.

What precipitated the conversation was a netseminar entitled ".NET 3.0: It's Just Down the Road" that I moderated and in which Juval did most of the talking. (I've learned that the trick to successful netseminars is for me to keep my trap shut. Everyone seems happier that way, and listeners end up learning more.) All in all, about 1000 listeners signed up and 500 or so eventually showed up. For those who couldn't make it to the session, the event is archived and you can catch up on what Juval had to say at any time.

Juval explained what .NET 3.0 is, focusing on its core parts--the Windows Communication Foundation (WCF), Windows Presentation Foundation (WPF), Windows Workflow Foundation (WF), and Windows CardsSpace (WCS). Particularly in the Q&A session with readers at the end, he went on to describe how it relates stuff like Windows 2000, .NET 1.1, Visual Studio 2005, and the like.

In fact, the netseminar is just part of a program by Dr. Dobb's to clearly explain what's what in terms of .NET 3.0. We'll conducing a series of roadshow-like conferences on .NET 3.0 starting in October. The series, aptly christened as Dr. Dobb's .NET 3.0 2006 Roadshows" will kick off in Los Angeles on October 9th, then move to Mountain View, CA, on to Downers Grove, IL, next to Reston, VA, and on to New York and Waltham MA. Speakers will include Juval, Michele Leroux Bustamante, and Brian Noyes, all noted .NET 3.0 experts.

In one of the audience polling questions we put forth in yesterday's netseminar, we asked when listeners expected to start serious .NET 3.0 development. If I recall correctly, about 58 percent expected to be doing so by, say, mid next year.

Like it or not, .NET 3.0 will be happening and it will offer new opportunities and challenges for developers. The netseminar and roadshow events offer a great opportunity for you to jump ahead. Hope to see you there.

Posted by Jon Erickson at 10:36 AM  Permalink |

It's been a year and a half since Herb Sutter published his prescient essay "A Fundamental Turn Toward Concurrency in Software" in Dr. Dobb's--and nothing's slowed down at all.

Just the opposite, in fact. By whatever technology it goes by--multi-threading, multi-core, parallel processing, or whatever--concurrency has gained momentum and is becoming the norm in software development. Why just this week alone, I've posted on Dr. Dobb's Portal articles such as:

• " Development and Optimization Techniques for Multicore Processors".
• "Dual-Core CPU Buyer's Guide".
• "Leveraging Multi-core Processors with Graphical System Design Tools".
• "Multicore Solutions Proliferating".

Then there are articles in recent issues of Dr. Dobb's Journal such as:

• "Inside the VSIPL++ API".
• "Multithreading, Java, & OSGi".
• "Threading & .NET".

And all that just for starters.

My point here is not just to impress you with the extent of currency coverage in Dr. Dobb's (well okay, in part it is), but to underscore that a grasp and understanding of concurrency is now fundamental to software development. Criminy, today's news is that even Apple has moved its entire iMac family to Intel Core 2 Duo processors.

If you're looking for even more info on multi-core based software development, another good place to start is with some presentations that PARC (formerly "Xerox PARC") put together earlier this year as part of the PARC Forum Series and has since made available in video and audio formats. The presentations include:

• Herb Sutter speaking on "Software and the Concurrency Revolution".
• Marc Tremblay, Sun Fellow, Vice President, and Chief Architect, Sun's Scalable Systems Group on "High Performance Throughput Computing".
• John Crawford, Intel Fellow, examining "The New Paradigm of Multicore Processors: Changing the Focus from Frequency to Instructions per Clock".
• Kunle Olukotun of Stanford University discussing " Architecture Support for Parallel Programming".

There's a trove of other audio/vidoe presentations in the PARC archives, ranging from AI to Nanotechnology. But the Multicore Computer Forum Series in particular is a gem and worth spending some time with.

Posted by Jon Erickson at 09:42 AM  Permalink |

Old editors never die, they just turn into "project liaisons" and "collaboration gurus." Old Dr. Dobb's editors anyway.

Of course, former Dr. Dobb's editor Eugene Eric Kim is a long way from the blue-hair and assisted-living crowd. But he is officially known as "project liaison and collaboration guru" for Doug Engelbart's recently launched HyperScope 1.0 project.

HyperScope is a Web-based implementation of Engelbart's 1968 NLS/Augment (oNLine System)--"a high-performance thought processor that enables you to navigate, view, and link to documents in sophisticated ways" so as to "make more advanced browsing capabilities available in existing tools, and to engage community participation."

More specifically:

HyperScope gives you the ability to change how you view a document... and how you address parts of a document. You can embed viewspecs in an address, and you can use these addresses to link to or jump around a document.

For example, every paragraph in a HyperScope document has a location number, an address corresponding to the paragraph's location in a document. For example, the second paragraph in the top-level of a document has the location number 2. To link to this paragraph, you can use the address:

http://foo/bar.opml#2

To jump directly to the second paragraph while viewing a document, you can click on the Jump button, type 2, and press Apply.

You can see the location numbers for every paragraph, you can click Viewspecs, check the Numbering checkbox, and click on Apply. You'll notice that this view corresponds with the viewspec m. To link to the second paragraph of the document with numbering turned on, you can use the address:

http://foo/bar.opml#2:m

You can also jump to this particular paragraph and view by clicking on Jump and typing 2:m.

Rather than trying to explain and understand this, it's probably better just to run the HyperScope demo, or just download the system from the HyperScope site.

It is also interesting that HyperScope is implements Englebart's Outline Processor Markup Language (OPML) which is now in Version 2.0 draft specification (written by Dave Winer).

The purpose of the format is "to provide a way to exchange information between outliners and Internet services that can be browsed or controlled through an outliner."

To be truthful, I'm still sorting through all this to understand what it is about. That said, I tend to pay attention to the stuff that Englebart comes up with since it always ends up being both interesting and important.

If nothing else, I'll try to find out from Eugene Kim what a "collaboration guru" is and does. Alas, knowing Eugene the way I do, it will probably mean a free lunch for him and a padded expense tab for me.

Posted by Jon Erickson at 09:15 AM  Permalink |

Hi, my name is Jon and I'm a puzzle-holic. Crossword, Sudoku, Jumble word. You name it, I play it. On airplanes, in front of the fireplace, at coffee shops. Give me a sharp pencil, a good eraser, and 15-minutes of quiet time and I'm ready to roll.

I like to tell myself that I'm pretty good at times, but I admit I'm no where as good as WebCrow, a computer program that took on--and soundly whipped--dozens of human competitors at last month's European Conference on Artificial Intelligence.

Both on-site and an on-line competitions took place, with eight different winners in the general catagories of Overall, Bilingual, Italian, and American. Participants were asked to solve five different crosswords--two Italian, two American, and one mixed). A time limit of 15 minutes was imposed for each crossword, 90 minutes total. The American-style puzzles were from the New York Times and Washington Post, and the Italian ones from La Repubblica and a local Italian paper. WebCrow competed from its " the computing department at the University of Siena. WebCrow was written by Giovanni Angelini, Marco Ernandes, and Marco Gori.

According to the authors, problems like solving crosswords from clues have been defined as AI-complete and are extremely challenging for machines since there is no closed-world assumption and they require human-level knowledge. Interestingly, for the first time since AI's kick-off, there is a first nucleus of technology, such as search engines, information retrieval and machine learning techniques, that enable computers to enfold with semantics real-life concepts. The goal of WebCrow's authors is to design a software system whose major assumption is to attack crosswords making use of the Web as its primary source of knowledge.

As described by Tom Simonite in an article in New Scientist, WebCrow uses four techniques in parallel to find possible answers to a clue. Two involve looking for the clue or a near match in a database of solved crosswords or using a dictionary. Another uses rules known to work on a kind of Italian clue with two letter answers and the fourth technique is to search the Internet. WebCrow performs a search using key words extracted from the clue. It can usually find the answer by looking at the small previews that appear with the search engine results, but it can scan whole pages if necessary. Words of the right length that crop up most often in the results are taken to be possible answers. A list of possible solutions to the clue is produced by combining the suggestions generated by each technique. When possible answers have been found for each clue the software uses trial and error to find the combination of interlocking answers that best fills the grid.

All in all, WebCrow did okay, finishing first and second. Detailed results are posted at the WebCrow web site.

And no, you won't find my name listed among the winners.

Posted by Jon Erickson at 08:55 AM  Permalink |