Conceptual Origami

Understanding Complex Datasets through Matrix Decomposition

2011-12-31T20:48:00.000-08:00

David Skillcorn's book on matrix decomposition techniques is superb. I especially enjoyed his coverage on non-negative matrix factorization (NNMF) techniques and eigendecomposition (i.e. spectral techniques). I would recommend the book to those interested in data mining and knowledge extraction. The techniques cover a wide range of media and are not simply restricted to relational datasets and textual documents. The treatment of PageRank is concise and articulate: demonstrating the deep relationship between graph mining and learning techniques and matrix decomposition(SVD amongst others) techniques that make search engines such as Google and Bing possible. As a reviewer summarized, "The author explores the deep connections between matrix decompositions and structures within graphs, relating the PageRank algorithm of Google's search engine to singular value decomposition. He also covers dimensionality reduction, collaborative filtering, clustering, and spectral analysis. With numerous figures and examples, the book shows how matrix decompositions can be used to find documents on the Internet, look for deeply buried mineral deposits without drilling, explore the structure of proteins, detect suspicious emails or cell phone calls, and more."

The link provides the complete text of Skillcorn's book in the form of a PDF.

Sacrifice and Transcendentals

2011-12-31T20:36:00.000-08:00

"Sacrifice is nothing other than the production of sacred things ... religion the search for a lost intimacy. Sacrifice destroys an object's real ties of subordination; it draws the victim out of the world of utility and restores it to that of unintelligible caprice. ... The sacrificer declares, 'Intimately, I belong to the sovereign world of the gods and myths, to the world of violent and uncalculated generosity ... I withdraw you, victim, from the world in which you were and could only be reduced to the condition of a thing, having a meaning that was foreign to your intimate nature. I call you back to the intimacy of the divine world, to the profound immanence of all that is.'"

-George Bataille, Theory of Religion

Ecstasy of the Beautiful: Beyond Symbolic Exchange

2011-12-31T20:19:00.000-08:00

"We have become completely absorbed by models, completely absorbed by fashion, completely absorbed by simulation. ... Our whole culture is in the process of shifting from games of competition and expression to the games of risk and vertigo. Hence, we move to the form of ecstasy. Ecstasy is that quality specific to each body that spirals in on itself until it has lost all meaning and thus radiates as pure and empty form. Fashion is the ecstasy of the beautiful: the pure and empty form of a spiraling aesthetic. Simulation is the ecstasy of the real."

-Jean Baudrillard, Fatal Strategies

Technologies of the Self

2011-12-14T08:02:00.000-08:00

"My objective for more than twenty-five years has been to sketch out a history of the different ways in our culture that humans develop knowledge about themselves: economics, biology, psychiatry, medicine, and penology. The main point is not to accept this knowledge at face value but to analyze these so-called sciences as very specific truth games related to specific techniques that human beings use to understand themselves. As a context, we must understand that there are four major types of these technologies, each a matrix of practical reason: (I) technologies of production, which permit us to produce, transform, or manipulate things; (2) technologies of sign systems, which permit us to use signs, meanings, symbols, or signification; (3) technologies of power, which determine the conduct of individuals and submit them to certain ends or domination, an objectivizing of the subject; (4) technologies of the self, which permit individuals to effect by their own means or with the help of others a certain number of operations on their own bodies and souls, thoughts, conduct, and way of being, so as to transform I themselves in order to attain a certain state of happiness, purity, wisdom, perfection, or immortality."

-Michel Foucault, Technologies of the Self

My Interview for Apache Hadoop on Windows Azure

2011-12-13T22:15:00.000-08:00

This is a link to a video interview that I did earlier this week in preparation for the developer preview of Apache Hadoop for Windows Azure. I founded this project at Microsoft and have shepherded its architecture and engineering efforts in addition to developing the overall framework for integrating OSS into Microsoft offerings. I have been joined by a cast of tremendously talented and inventive developers. The opportunity to work with members of the open source community and especially those involved in the Apache Hadoop project(s) has been a rewarding experience. I personally view 21st century global economic markets being transformed by two complementary and mutually amplifying trends: (1) competitive differentiation based on analytics and (2) the emergence of cooperative, crowd-sourced open development of software tools and utility-priced scale-invariant clouds. The globalization of software and cloud-delivered computational platforms is rewriting the fundamental laws of economic competitiveness. Like the emergence of machinery to accelerate the rise of industrial capitalism in the 19th century, analytics-/data-driven competition and loosely-federated open markets for software are transforming the modern corporate IT environment and this trend is accelerating.

Hadoop is at the forefront of the trend and our service on Azure is an important step in the seamless integrated flow of scale-invariant analytics and information derivatives into classic business intelligence, data warehousing, and statistical analysis toolchains and practices.

The project, codenamed isotope, was done to provide a rich port of Apache Hadoop onto the Windows Server and Windows Azure platforms. Cloud-scale computation on elastic services is delivered through Azure and the ability to easily integrate any data sources and big data workloads via an elegantly intuitive portal and interactive shell. Our developer preview is the first manifestation of the longterm roadmap that we have for bridging the worlds of self-service business intelligence to cloud scale analytics delivered to users on any device. The elastic map reduce and analytics service will be available at www.hadooponazure.com. The summary of the talk is included: In this interview, Alexander talks to us about how big data has already had a profound impact on businesses. Big data problems require new class of NoSQL technologies and tools like Hadoop. Hadoop has a vibrant ecosystem for processing big data using commodity hardware. Microsoft's Hadoop distribution will democratize big data tools even further by making them more accessible and easier to program for people who don't know anything about clusters. Our Isotope team has adopted the OSS development model, and will be contributing code back to the community. Hadoop on Windows Azure will be available in CTP in December, and GA will be in March.

For me, one hashtag says it all: #Slide26.

Vector Processing Languages: The Future of Big Data Analytics and Real-time Business Intelligence

2010-12-31T14:57:00.000-08:00

When I think about the systems being designed to address the big data problem, partially a data storage, access, and query problem (10%) but mostly about analytics (90%) at scale, latency, and performance thresholds only imagined by previous generations - I immediately think of vector processing systems and elastic, on demand, computing fabrics.

One of the best systems for the vector processing aspect that I have seen is q/kdb+ from kx systems. Q/kdb+ was developed out of K, itself spawned by A(+), and previously founded on Kenneth Iverson's seminal language APL, so it is the exemplar vector processing language. Another vector processing language familiar to many is Mathwork's MATLAB. MATLAB, short for Matrix Laboratory, definitely shows its intellectual ancestry in APL.

Why vector processing? Is not the world to be owned by dynamic, object-oriented, and functional programming genres and languages? Why would so many hedge funds, investment banks, and savvy individual traders use systems such as q/kdb+ and others instead of these "obvious" choices proffered by large software enterprise players such as Oracle, IBM, and Microsoft? There are a number of factors militating against this that run deep within the big data movement, and I shall only summarize a few of them here.

First, nested vector data models and higher-order processing such as those provided in q/kdb+ and other vector processing systems are uniquely appropriate to processing massive, time series ordered (partitioned) datasets. Unlike relational data models, which lack intrinsic ordering (unordered set vs. ordered vector semantics - remember?), vector models are natively adapted to operating over massively sized, dimensional data with intrinsic partitioning (time). Stock trading data streams are a good example. This is evident in the financial services industry and anywhere where time series analytics (understanding what is in your data) and predictive analytics (regression, what is likely to happen in the future based on that information) are keys to competitive differentiation.

Second, vector processing engines scale extremely well since they are designed for the in-memory, high speed performance with zero wait states and blocking semantics. Most vector processing engines, such as q/kdb+, are interpreters running at extreme speeds in a single thread of control and with no complex lock managers trying to outsmart the runtime vector engine. This is in complete contradistinction to the prevailing wisdom of RDBMS systems, which were designed for solving many concurrent access, I/O bottlenecks, and interleaved indexing operations that simply no longer apply to the big data world. Vector processing engines are already way ahead of where streaming and in-memory tuned engines such as Streambase, Coral8, Vertica, SciDB, and others wanted to get to. I suspect that we will see old school RDBMS vendors try to reposition some of their antiquated technologies in this vector processing space. Although this may be laudable and slightly quixotic, it is still like trying to rebuild a Ferrari sports car starting with the design of a Yugo. Why, in a sane world, would you start from that design position? You would not.

Third, vector processing languages are highly amenable to parallelism and high performance concurrent execution. A key part of this is the minimal unit of operation of these languages: the nested, n-dimensional vector. This, and not the scalar, is the unit of operation. This is a subtle point: most languages are designed to operate over data one scalar value at a time. To deal with anything more complicated requires a great deal of error-prone iteration, typically nested. What requires two or more iterative looping constructs in a language such as C# or Java is simply one operator (or higher-order function) in a vector processing language. An example would be computing the average of a vector of test scores: in an q or APL-like language this becomes a simple one liner: sum aVector / tally aVector. That is the entire program. Taking these powerful monadic and dyadic operators and partitioning them across an elastic fabric of processors, caches, and storage is simplicity itself since so little context needs to be migrated in the form of closures and execution context. This hidden support for scalable parallel processing is another of the values of vector processing languages. Few Algol-60 derived languages can boast this form of inherent parallelism. In a big data world, this is critical for success and scalability.

Finally, I posit that vector processing engines, when wedded to elastic, on demand compute and storage platforms, will become the most attractive computational and modeling option for data scientists (quants) trying to do data-driven science and business. Combining a small vector processing core with a scalable, elastic fabric such as Heroku, EC2, Google, or Azure will usher in a new era of vector fabric processors. They will have all of the algorithmic sophistication of classical high performance in-memory vector engines such as q/kdb+, vhayu, and timescale but combined with the infinite CPU, storage, and scale-out of cloud computing. Being able to interface these fabric-scaled vector processors with rich visualizations of the sort one encounters in R, Excel, MATLAB, Tableau, and other real-time data graphing exploration vendors makes for a great marriage of equals. You do not have to be locked into a proprietary visualization system any more than you have to be locked into a proprietary cloud. There are too many excellent choices available. I have addressed several in my recent blogs.

This marriage of vector fabric processors and the economies of scale that cloud computing will provide should create a fundamental economic disruption in both the cloud computing vendor landscape as well as across many verticals: from finance to insurance to computational bio-pharmacology to retail to leisure. We are already seeing the convergence of medium-scale data warehousing workloads for analytics and map reduce computational models (AsterData, Greenplum/VMWare, Netezza/IBM, etc.) This is only the harbinger of a far more systematic transformation in the way machine learning, predictive analytics, mass-scale simulation, and continuous (near-line) business process optimization is done in the next five years.

Imagine the three screens of the future analyst workstation as depicted above: (left) historical data warehousing, (middle) real-time business intelligence, and (right) predictive analytics, simulation, what-if? analysis, and regression modeling. All three screens (modes) can be incorporated in one unified platform for information and (actionable) insight with the power of vector processing engines combined with scalable cloud fabrics.

Interested parties are invited to try out q/kdb+ (free evaluation trial of 32bit edition) or one of the many high quality APL implementations available online. I would recommend the following APLs to try: APL-X and Dyalogic APL. All three are excellent offerings and will demonstrate the power of higher-order vector processing, for scenarios ranging from statistical analysis, portfolio strategy optimization, risk management, search engine optimization (SEO), to analytical CRM for eCommerce. The old cliche of back to the future may well prove to be prescient in regards to APL and its many children.

Cloud fabric plus vector processing plus scalable storage (ala MongoDB, Membase, et al) is the equation for the future of big data analytics and continuous, online machine learning. Say goodbye to Java, C++, C#, and the old scalar-at-a-time programming languages derived from Algol-60.

Welcome back to the serene elegance, conciseness, and high performance parallelism of APL (minus the custom symbology) in the big data world.

Coltrane: The Brilliance of a Uniquely American Genre

2010-12-31T14:09:00.000-08:00

It is hard for me to write about music: it is one of those aesthetic experiences that resist paraphrase and reduction to a linear or serial discourse. Three of my favorite musicians, of the non-electronica genre, are John Coltrane, Miles Davis, and Thelonious Monk. Coltrane, or trane as he was sometimes referred to, has a special place in my heart. For instance, Blue Train, which is quintessentially hard bop, is one of those recordings, like Kraftwerk's Computerwelt, that I never get tired of listening to wherever I happen to be. I am not suggesting a hierarchy of masters between the three, or between the many other talented Jazz musicians west and east coast, that created a unique American art form we call jazz. It has spawned everything from acid jazz to downtempo to funk to rap.

In fact, jazz may truly be said to be the cultural and artistic capstone of American 20th century musical art. It influences every genre, especially its riotous offspring rock and roll. But to truly appreciate Coltrane, Davis, And Monk - just look to the recordings and jam sessions of the late 50s and early 60s. This is when all the templates for popular music were laid down. You do not have to enjoy them as trailblazers or genre definers but merely for what they were: brilliant musicians, composers, and arts. Art is its most powerful and elemental when it transform the particular and subjective into something beyond itself: when it reveals the playful beauty of our (very mammalian) species. We love to play and art is play raised to the infinite degree.

Coltrane's amazing musicianship, compositional skills, charisma, and ability to overcome the many demons of a segregated post-war American society only add to the triumph. But when it comes to the music, you really can leave all that behind - this is the place of the numinous, the ineffable, and transformational.

Do yourself a favor and listen to Blue Train, Giant Steps, and My Favorite Things. You will see what I mean about art at its apex. Do not ask me to explain it any further, I shall just refer you to Miles Davis' definition of jazz.

Bio-Inspired Artificial Intelligence: How Self and Complexity Emerge in Biological Systems

2010-12-31T13:55:00.000-08:00

Although the connections between biological systems and computing have been many in many other places, I found this primer, Bio-Inspired Artificial Intelligence, especially interesting for its chapter on immunological systems. The contention that the basis of identity (self vs. other) and even intelligence has its evolutionary basis in the adaptive evolution on immune systems is just too fascinating to ignore. The authors spend considerable time working through basics of immunological response and conceptualizing what they term shapespace. I think this is great text to selectively tip into for this topic alone - since so much of the literature for computational biology is focused on bioinformatics (strong matching), genomics, proteomics, and them a fast shift to systems biology.

I actually think the relatively primitive immunological system (pattern matching engine) is a superb place to ask some fundamental questions on what we think of as the self and how it evolves in a dynamic, evolutionary ecosystem constantly deluged by bacteria, prions, viruses, electro-nmagnetic radiation, DNA replication errors (cancers), and even more elaborated bio-programming systems (namely: memes). I like the approach, since it posits a very small number of non-supernatural and metaphysical entities to account for both the emergence of intelligence, collective coordination, cooperation, and the explosion of biodiversity.

As one person put it: Natural evolution is a good analogy to this method–the rules of evolution (selection, recombination/reproduction, mutation and more recently transposition) are in principle simple rules, yet over thousands of years have produced remarkably complex organisms.

Enter the Void: Sunyata and CGI in Gaspar Noe's Latest Film

2010-12-29T09:36:00.000-08:00

Śūnyatā(शून्यता)(Sanskrit noun from the adj. śūnya: "zero, nothing") is often translated as "void" and adds the appropriate resonance (spiritual, philosophical, and aesthetic) to Gaspar Noe's latest cinematic triumph Enter the Void. This is one of the rare instances of film that actually makes the strongest possible case for the distinction between the Real(tm) and phenomenal. The effect is profound as it is disorienting and nauseating in places. Noe has summarized his film's plot as, "the sentimentality of mammals and the shimmering vacuity of the human experience." Weaving in the Tibetan Book of the Dead, massive doses of GHB and DMT-induced Dreamtime visualization, more-real-than-real overhead shots of Tokyo at night, trademark hyper-brutalist sexuality, and a poignant tragedy to underscore the maelstrom of sensations that pull the viewer along, Noe has once again crafted a new language of visual expression. Even more brilliantly, he has taken the tools of kitsch, Hollywood CGI, worked to death in lackluster juvenilia such as Star Trek and Jurassic Park, and transformed the Real (our experience of the "void") into a hallucinogenic and shimmering tableau that never fails to disorient, fascinate, and remove the artificial irony of the sophisticated viewer.

You have no choice but to submit to Noe's art.

The link takes you to a montage of the incredible CGI that was created for Enter the Void. From the neon-drenched Tokyo to inter-uterine fertilization to Cthulu-esque higher-dimensional lifeforms floating through ceilings and walls, his CGI-rendered camera truly explores the ineffable recesses of the human experience.

This is technology in the service of a turbulent vision that turns viewer into the disembodied, floating soul of the main character as he drifts from scene to scene awaiting reincarnation. Seldom has technology produced such a profound spiritual and emotional effect. Noe has established not only a distinctive language of expression but quite probably a new genre of generative art. CGI has finally been used to fuse sentimentality and the void.

This may be the best, inadvertent, recruiting film for Buddhism ever created.

Kling Klang: The Electronic Garden of the Musikarbeiter

2010-12-29T08:45:00.000-08:00

Describe how Kling Klang looks like inside?

Ralf Hütter: It's an electronic garden. We like to perform electronic gardening.

The location of the building is a well kept secret?

Ralf Hütter: Sometimes somebody comes all the way from Japan. But the doors are tightly closed and we are never inside on a regular basis. We concentrate ourselves onto the music and not on what's happening outside. We call ourselves Musikarbeiter. Everything with us is about the music.

What a brilliant summary of the ethos behind the electronica. Hutter is one of the co-founders of Kraftwerk and its spokesmodel.

Insen: Microsound Engineering and Classical Music Performance

2010-12-29T08:06:00.000-08:00

Ever since being profoundly shocked by the all electronic soundtrack to Forbidden Planet at the age of eight and the generative-procedural audio workout of Morton Subotick's Silver Apples of the Moon (Nonesuch was such a brilliant label for experimental electronica) - I have had a life-long love affair with computationally and electronically designed, composed, and performed music. Needless to say, Kraftwerk has been the creative influence of my life. Computerwelt, Tour de France (Single Version 1983), and Mensch-Maschine are the apex of popular electronica - their influence reverberates through the decades in ways that are too innumerable to specify. From Detroit to Tokyo to London to Berlin, the Kraftwerk ethos and exacting approach to crafting electronica in a myriad of genres has influenced more people than just about anyone I can imagine.

One of the most extreme research and performance tangents of electronica has to be microsound, the mathematical exploration of tonality at the ultra-low latency range. Championed by Curtis Roads and others, microsound is a sound designer's and mathematician's aesthetic paradise. Granular synthesis was developed to explore this space, providing a much needed antidote to the clash of the titans dueling between the subtractive-analog and frequency-modulation (FM) schools of sound design. With the advent of fantastic virtual synthesizer technology (both Steinberg VSTs and software such as Reason, Ableton, Renoise and Reaktor) along with sound design laboratories like Pure Data (Pd) and MAX/MSP, the aspiring sound designer has free rein to chart the methematical and audio realms that escape the classical chromaticism that many have been conditioned to accept as the minima and maxima of musical expression. As Jacques Attali once suggested in Noise: the Political Economy of Sound, where we make the distinctions of noise versus music is a cultural and political segmentation - it is not innate in the frequency spectrum. A microsound composer like Carsten Nicolai, who typically records under the nom-de-audio of Alva Noto, is a great example of the marriage of technology, mathematics, and a fine sense of symmetry - albeit in a different dimension than one typically encounters in either to classical repertoire or in popular music - which is typically (and dismally) a strict subset of the former. Nicolai's collaborations with the talented Ryuichi Sakamoto for Vrioom, Insen, and other mash-ups of microsound and classical piano are an exemplary entry point into the realm of mathematically generated and designed audio experience. The description of the Insen release is worth reviewing.

Insen is the second album in an ongoing collaboration between Japanese composer Ryuichi Sakamoto and German electronic artist Carsten Nicolai (here credited as Alva Noto). The album's core sound is a blend of Sakamoto's impressionist piano melodies and Nicolai's digitally processed beats and sounds. Released in 2005 by Nicolai's Raster-Noton label, it follows the duo's debut album Vrioon, which was named album of the year in 2004 by The Wire magazine.

Insen is, by my estimation, the finest of the collaborations. Aurora and Logic Moon are simply mesmerizing. You can focus on macro-structures (piano chord progressions) or on micro-structures (sound shapes and tonal textures) and still be carried along in the aesthetic experience. This ability to straddle both ends of the continuum makes Insen so noteworthy. I have enjoyed the solo efforts by Nicolai too: For, Unitxt, and For2 stand out. For (Katsushito Hokusai) is absolutely a simple, Zen joy. If you enjoy electronica and designing and performance of noise/music (generative sound is the catch-all term I have for it), try out Insen or one of the other compositions.

MongoDB: Scalable Storage and Computation without Schema

2010-12-29T05:42:00.000-08:00

MongoDB is one the best of the new crop of post-relational or post-schema-driven persistence engines designed for large scale cloud storage. Like CouchDB, another document-centric and schema-less cloud scale engine, MongoDB also offer several robust features: MapReduce support over its data, attribute-level indexing, auto-sharding, in-place updating, and high-availability. What MongoDB does not have is equally interesting: mandatory schema and SQL-like restrictions on data access and programming. Written in C++ and designed for multi-language access (Java first and foremost it seems), MongoDB is what I would term an instance-oriented cloud store: instances (documents) can be highly variant in structure and the system elegantly scales and continues to perform. A lot of this has to do with the manner in which the auto-sharding and replica management happens behind the scenes. This automagical behavior is reminiscent of the old(er) world of RDBMS storage that has dominated the enterprise computing space from the late 1980s through early 2000s but has jettisoned many of the sacred cows of SQL-based storage: homogeneous structure, convoluted graph-oriented operation (subqueries), and scale limitations based on the design of locking managers that were designed for I/O characteristics that simply do not obtain in the cloud space.

I have had a chance to play with MongoDB, Membase, Hive, Cassandra, and CouchDB and can definitely say that this feels like the inevitable direction of cloud scale storage and computation (see Spark, Dryad, MapReduce, etc.). Microsoft has proprietary auto-sharding storage systems (SQL Azure and Azure XStore) which I shall write about on another occasion. All that said, schema-free storage, indexing, auto-sharding, and high performance make for a compelling offering in MongoDB.

Spark: Big Data Computing with Scala

2010-12-29T05:24:00.002-08:00

Spark is a new distributed computing model built on to of the Mesos cluster operating system. Both projects are done out of UC Berkeley. The interesting thing about Spark is that its concept of resilient distributed datasets (RDDs) and who the underlying Scala language with its support for shipping closures around the network delivers a scalable and high performance model when compared to the state of industry: i.e. Hadoop MapReduce. Spark's architecture naturally allows for efficient in-memory operation as well as caching - so that iterative algorithms such as those that one often encounters in machine learning, matrix algebra, mathematical optimization, and multi-dimensional indexing can really see a significant speed-up. The published performance/scale figures show that as iterations increase (in the case of say a logistic regression problem), Spark is 10-15 times faster in execution than a standard Hadoop MapReduce job operating over the same HDFS datasets. In-memory and cached intermediate state make a big difference in the distributed big data computing space.

A nice write-up of the logistic regression case is provided below from Matei Zaharia's website.

In more interesting applications, users probably need to read a data set and potentially transform it before performing calculations on it. For this purpose, Spark provides a second type of distributed dataset -- a file in the Hadoop Distributed File System (HDFS). Currently, only text files are supported. The HDFS file looks to the programmer like a collection of records (in text files, each record is a line). However, operations on it run at the nodes that contain each block of the file, as in MapReduce.

The corresponding parallel program in Spark (on top of Scala) to implement the logistic regression is the following:

// Read data file and transform it into Point objects
val spark = new SparkContext()
val lines = spark.hdfsTextFile("hdfs://.../data.txt")
val points = lines.map(x => parsePoint(x)).cache()

// Run logistic regression
var w = Vector.random(D)
for (i <- 1 to ITERATIONS) {
val gradient = spark.accumulator(Vector.zeros(D))
for (p <- points) {
val scale = (1/(1+Math.exp(-p.y*(w dot p.x)))-1)*p.y
gradient += scale * p.x
}
w -= gradient.value
}
println("Result: " + w)

This is an impressive programming model and one that definitely bears watching as Hadoop MapReduce, Microsoft Dryad, and other frameworks (like Sphere and Twister) vie for dominance in the BigData world that has emerged.

Node.js: Scalable Cloud Services

2010-12-28T15:59:00.000-08:00

Sometimes, it is easier to let a framework speak for itself. In this case, node.js. It combines the zero-blocking design of several event-driven server architectures with the number one web programming language Javascript. The syntax is straightforward and easy to parse. Readability is a key adoption driver and hence why several continuation- and meta-programming-based languages and frameworks just do not fare very well. Node.js is one to watch (and use). I have excerpted some of their materials below to demonstrate the ease of use in building scalable server-based.

An example of a web server written in Node which responds with "Hello World" for every request.

var http = require('http');
http.createServer(function (req, res) {
res.writeHead(200, {'Content-Type': 'text/plain'});
res.end('Hello World\n');
}).listen(8124, "127.0.0.1");
console.log('Server running at http://127.0.0.1:8124/');
To run the server, put the code into a file example.js and execute it with the node program:

% node example.js
Server running at http://127.0.0.1:8124/
Here is an example of a simple TCP server which listens on port 8124 and echoes whatever you send it:

var net = require('net');
net.createServer(function (socket) {
socket.write("Echo server\r\n");
socket.on("data", function (data) {
socket.write(data);
});
}).listen(8124, "127.0.0.1");

Node's goal is to provide an easy way to build scalable network programs. In the "hello world" web server example above, many client connections can be handled concurrently. Node tells the operating system (through epoll, kqueue, /dev/poll, or select) that it should be notified when a new connection is made, and then it goes to sleep. If someone new connects, then it executes the callback. Each connection is only a small heap allocation.

This is in contrast to today's more common concurrency model where OS threads are employed. Thread-based networking is relatively inefficient and very difficult to use. See: this and this. Node will show much better memory efficiency under high-loads than systems which allocate 2mb thread stacks for each connection. Furthermore, users of Node are free from worries of dead-locking the process—there are no locks. Almost no function in Node directly performs I/O, so the process never blocks. Because nothing blocks, less-than-expert programmers are able to develop fast systems.

Node is similar in design to and influenced by systems like Ruby's Event Machine or Python's Twisted. Node takes the event model a bit further—it presents the event loop as a language construct instead of as a library. In other systems there is always a blocking call to start the event-loop. Typically one defines behavior through callbacks at the beginning of a script and at the end starts a server through a blocking call like EventMachine::run(). In Node there is no such start-the-event-loop call. Node simply enters the event loop after executing the input script. Node exits the event loop when there are no more callbacks to perform. This behavior is like browser javascript—the event loop is hidden from the user.

Heroku: Elegance and Simplicity for Cloud Computing

2010-12-27T12:27:00.002-08:00

Of the many cloud platform as a service (Paas) architectures that I have seen, I have to admit that Heroku leads the pack in terms of elegance and simplicity. Its core concept of a dyno (the packaging, deployment, and billing unit) is easy to cost and understand. You understand exactly what goes into it: from RubyVM to POSIX services. Built out as a Ruby (on Rails) deployment and execution platform, the four panel architecture cartoon/film strip is definitely an expository technique that other PaaS businesses should emulate. being able to concisely and clearly articulate your architecture and why it was designed the way it was is subtle marketing technique. If you are ready to start, an interactive two sliders web application dynamically immediately explains how much running dynos will cost you: simplicity is a feature. Consider computing the true cost of deploying a multi-tier cloud application on other platforms. Finally, the deep integration with git makes deployment, versioning, and packaging simple and fast. Git is a wonderful tool and my preferred source code control system.

Ruby is an intriguing dynamic programming language that marries Smalltalk meta-programming, block closures, and a rich framework ecosystem with a syntax that is, mostly, reasonable. Although every surface syntax requires some adjustment on the part of the neophyte, Ruby is certainly within one standard deviation from the norm in terms of orthogonality and symmetry. What I think is extremely positive about Ruby is its framework ecosystem: Mongrel, Ruby of Rails (RoR) and Sinatra being just three of the superb production quality frameworks developed by the community.

I posit that the quality and number of 3rd party, community, frameworks and tools developed for a language is a strong indicator of its relevance. Single vendor vertical stacks all tend to possess a disconcertingly myopic worldview: insular and ultimately over-engineered to a single set of requirements.

Let us see how the Saleforce.com acquisition of Heroku plays out. I think Salesforce.com has a great platform to penetrate far beyond their their current APEX community/audience.

Set Theory and Philosophy

2010-12-25T11:02:00.000-08:00

Michael Potter's survey of set theory and its philosophical interpretations - platonist, constructivist, formalist, etc. - is thought provoking and incredibly balanced. Its eschews polemics to perform a deep survey of set theory, the attempts at its axiomatization, and the ontological implications of supporting some or all of the classical axioms used to corral the complex set of ideas embodied in term set. It is also a great preparation for those that become intrigued by the insights offered in non-well-founded set theory. Potter's work is a great place to start, I only wish it had been published two decades ago.

After Finitude: Quentin Meillassoux and Speculative Realism

2010-12-25T10:53:00.000-08:00

I profess to a significant interest in mathematics and set theory, along with their applications within large-scale cloud computing, ontology, and gerenative-procedural art. Recently, I have become intrigued by the writings of Alain Badiou and his student, Quentin Meillassoux. Badiou's intriguing view of ontology-as-mathematics, with a deep excursion into Georg Cantor's transfinite set theory, paves the way for After Finitude by Meillassoux. A superb summary of After Finitude is quoted below. I urge those interested in speculative realism to read the book, as well as Badiou's Being and Event and Number and Numbers.

In his book After Finitude: An Essay on the Necessity of Contingency, Meillassoux argues that post-Kantian philosophy is dominated by what he calls “correlationism,” the often unstated theory that humans cannot exist without the world nor the world without humans. In Meillassoux’s view, this is a dishonest maneuver that allows philosophy to sidestep the problem of how to describe the world as it really is prior to all human access. He terms this pre-human reality the “ancestral” realm. In keeping with the mathematical interests of his mentor Alain Badiou, Meillassoux claims that mathematics is what reaches the primary qualities of things as opposed to their secondary qualities as manifested in perception.

Meillassoux tries to show that the agnostic scepticism of those who doubt the reality of cause and effect must be transformed into a radical certainty that there is no such thing as causal necessity at all. This leads Meillassoux to proclaim that it is absolutely necessary that the laws of nature be contingent. The world is a kind of hyper-chaos in which the principle of sufficient reason is abandoned even while the principle of non-contradiction must be retained.

For these reasons, Meillassoux rejects Kant’s so-called Copernican Revolution in philosophy. Since Kant makes the world dependent on the conditions by which humans observe it, Meillassoux accuses Kant of a “Ptolemaic Counter-Revolution.”

Shogun Toolbox for Machine Learning

2010-12-23T06:50:00.001-08:00

The Shogun Toolbox is an extremely impressive meta-framework for incorporating support vector machine and kernel method-based supervised machine learning into various exploratory data analysis environments. It focuses on large scale machine learning, something critical for online and Web-scale applications of these techniques. The integration with environments such as MATLAB and R is also vital to streamline the workflow of machine learning libraries (engines). The introduction from the main site is worth citing: (Shogun's) focus is on large scale kernel methods and especially on Support Vector Machines (SVM) [1]. It provides a generic SVM object interfacing to several different SVM implementations, among them the state of the art OCAS [21], Liblinear [20], LibSVM [2], SVMLight, [3] SVMLin [4] and GPDT [5]. Each of the SVMs can be combined with a variety of kernels. The toolbox not only provides efficient implementations of the most common kernels, like the Linear, Polynomial, Gaussian and Sigmoid Kernel but also comes with a number of recent string kernels as e.g. the Locality Improved [6], Fischer [7], TOP [8], Spectrum [9], Weighted Degree Kernel (with shifts) [10] [11] [12]. For the latter the efficient LINADD [12] optimizations are implemented. For linear SVMs the COFFIN framework [22][23] allows for on-demand computing feature spaces on-the-fly, even allowing to mix sparse, dense and other data types. Furthermore, SHOGUN offers the freedom of working with custom pre-computed kernels. One of its key features is the combined kernel which can be constructed by a weighted linear combination of a number of sub-kernels, each of which not necessarily working on the same domain. An optimal sub-kernel weighting can be learned using Multiple Kernel Learning [13] [14] [18] [19]. Currently SVM one-class, 2-class and multiclass classification and regression problems can be dealt with. However SHOGUN also implements a number of linear methods like Linear Discriminant Analysis (LDA), Linear Programming Machine (LPM), (Kernel) Perceptrons and features algorithms to train hidden markov models. The input feature-objects can be dense, sparse or strings and of type int/short/double/char and can be converted into different feature types. Chains of preprocessors (e.g. substracting the mean) can be attached to each feature object allowing for on-the-fly pre-processing.

Wikileaks and Foucault's Revenge

2010-12-22T00:55:00.000-08:00

"The epistemological mutation of history is not yet complete."
(Michel Foucault, Introduction to the Archaeology of Knowledge)

Calculus of Form in Greg Lynn's Design Language

2010-03-21T10:43:00.000-07:00

I am an admirer of Greg Lynn, a noted architect and theorist. His firm, FORM, has completed a number of remarkable projects. The Korean Church Project in Queens is an excellent example of his infusion of design with computer engineering, calculus, and an understanding of flow and form in sculpting space. Lynn is also a theorist: he wants to understand and articulate the underlying mathematics of change (calculus). He advocates its use as an overt toolset for design at all scales - from small artifacts to the most public spaces. In the linked video, Lynn makes an insightful observation about the shift in viewpoint and approach as one moves from the classical ideal forms (lines) and fractional geometries of design to a sophisticated languages of curvature that is infused with modern computer-aided design. This is a convergence of procedural and differential techniques. The results are quite stunning. His monograph, On Animate Form, is highly recommended.

Some people refer to this trend as blobotecture - referencing the biomorphic and overt variational curvature of materials that are a hallmark of using models such as heat diffusion to model interior/exterior spaces. I think it is fascinating and noteworthy. But, take a look at Christopher Wren's work: I think you will see the precog echo of thinkers like Lynn working in the early 21st century. Force and form are always in dynamic equilibrium and amply in display in innovative architecture such as Lynn's.

Search, Discovery, and Insight via Emergent Semantics

2009-12-07T11:57:00.001-08:00

I am very pleased to be able to announce that Microsoft Semantic Engine, a group that I founded and led up through the current PDC 2009, has finally been made visible to the general public. The blog linked above called it the "hidden gem" of PDC 2009. That was very kind of him. In the two years that we have been building this technology, our goals have remained the same: to make semantically-enhanced search, discovery, and organization autonomic, scale invariant, and immediately useable by anyone used to search engines.

The system is built from three basic ideas: streams, properties, and concepts. The latter are the emergent categories and clusters that are discovered by the Semantic Engine and used for indexing. This is the core of the "semantic" in semantic search and discovery. It is also the basis for finding the hidden (i.e. latent) relationships in information automatically and making these relationships readily available to users (consumers or business workers).

Approachable, semantic, and self-organizing. No one wants to have to deal with the tera- and peta-bytes of information that they are exposed to daily. People need the right kinds of filters to find and use the information that is most relevant to them. Learning what those filters should be is part of the Semantic Engine's role. This is where a background service integrated deeply with the power of SQL Server and Cloud services like SQL Azure makes all of the difference.

It is definitely a tall order and one which Microsoft is investing in heavily. Meaning-driven indexing, automatic classification, and scalable storage for all shapes and sizes of information (documents, videos, audio, business forms, web pages, RSS/ATOM feeds, etc.) is the vision of "emergent semantics and recombinant information processing". It is also the basis for many value-added services: from predictive analytics and sentiment analysis to more personalized and tailored search and discovery services. Another great blog describing the PDC session is at Harnessing Unstructured Data .

Finally, the direct link to the PDC session video is at PDC 2009 Video. This video covers all of the main technical and business aspects of semantic engineering at large scale and the unique hybrid approach that Semantic Engine has taken to accomplish its goals. It is not a W3C SemanticWeb(tm) approach but one which melds the unique capabilities of unsupervised machine learning (hierarchical clustering), information retrieval models (higher-dimensional vector spaces), pluggable and trainable classifiers (SVMs, Naive Bayesian, Maximum Entropy, Decision Tree, etc.), and personalized filtering and ranking. Your search, discovery and organization finally become based on your preferences. This enables sharing and access of your enhanced and conceptually indexed information artifacts with social and business cliques. The stuff you care about is always "handy and nearby".

It all starts with three building blocks (streams, extracted properties, and concepts) that allow for emergent semantics to become a reality. Emergent mechanisms are the way to go for ensuring scale invariance and adaptive systems that are robust to change and autonomic evolution over time. Biological systems are and information-processing systems need to get that way too. A new order of Deleuzian 'machinic phyla' that can handle the rhizomatic nature of ideas and things.

Meaning-driven computing has arrived. The next decade should be very interesting for developers, consumers, and business people. The Microsoft Semantic Engine is just one technology on the road to omnipresent and unified search, discovery and insight.

The Geometry of Interaction

2009-11-24T11:53:00.000-08:00

Linear Logic, proposed by Jean-Yves Girard, remains one of the most important and startling developments in the study of logic and human cognition. I would have to place it within the same eschelon as Aristotle's and Frege's works. It is the first major, and successful, move to displace the concept of truth (preservation) and replace it with the notion of computational resources and their consumption in proof, reasoning, etc. This may not seem like much, but consider the practical implications to computer science, epistemology, quantum mechanics, and modeling of actual human cognition. Linear Logic combines some of the best aspects of constructivist approaches with a system that eschews the facile semantics often used to describe logic. The Stanford University site is an excellent resource. About the only thing even more interesting that this modified sequent logic (doing without weaking and contraction in proofs and emphasizing interaction and duality) is Girard's subsequent work on Ludics. More on that later.

From Vermilion Sands to Eden-Olympia: The Dystopian Catastasis of JG Ballard

2009-10-04T20:11:00.000-07:00

With the passing of JG Ballard, it somehow seems appropriate to revisit the gallery of his hallucinatory micro-utopias gone awry. Stretching from the gantries of abandoned Cape Canaveral Saturn V launch sites and deserted motels with sand filled pools and broken deck chairs, to the aesthetic and psychotic permutations of Vermilion Sands, and finally passing by to the terminal dystopias of secure leisure along the Mediterranean and what they unleash in their inhabitants in Super-Cannes, Cocaine Nights, and Kingdom Come. The interesting thing about Ballard's fiction is its progression backwards from the 'future', that staid cliche of science fiction, to the 'eye blink' of time in front of our perpetual 'present'. All science fiction takes place five minutes in the future. The present trying to see itself in the rear view mirror is what we call the 'future'. Gone in that final gallery are the crystallized forests and nature overcoming homo sapiens through biological evolution and catastrophe. Everything has been replaced by the human inability to cope with the very fabric of its deceptive, utopian desire: endless leisure, infantile lack of responsibilities, and the ready-to-order brutality of any sort of 'game' the residents of these claustrophobic micro-enclaves might require to pass the time. It is a literature of involution and catastasis. Even the ritualized murders, incest, adultery, parricides, and all-pervasive corruption are just prelude to a final act that simply never arrives. The future can never arrive at this present closed off from time. Ballard's literature is a brilliant reflecting mirror for this 'monadic' prison- one imagines it as a religious allegory recounted by descendants of ours as they scurry through the broken skylines, looted shopping malls, and disused airports of the near future. There has been a frequent criticism of Ballard's later fiction: the repetition of tropes without resolution is somehow a tepid 'sampling' of his own best work. This misses the point entirely. I think Ballard captured in a performative gesture the profound enclosure of culture and psychology in a world longed for yet utterly lethal to our species. One does not require a new, benevolent psychopathology to liberate the human animal from the 'mind forged manacles' of civilization, as Ballard's masterpieces of the 70s - Crash, The Unlimited Dream Company, and The Atrocity Exhibition - suggest. Instead, one must come to accept the melancholy recognition of watching a species at a dead end in rich, exclusive resorts: a world turned into a theme park and zoological preserve for the affluent and their entourages. Exhibit: homo sapiens. In a way, I miss the surrealistic joy of Vermilion Sands, the most ecstatic and playful of Ballard's imagined psycho-topographies (one really cannot call them geographies) - where cloud sculptures and singing machines vie for our attention without the shadow of senescence and thoughtless depravity seeping into the edges of the scene. The imagined micro-utopias (dystopias) of Eden-Olympia are far colder climes - populated by all the misery of those who received exactly what they wanted and have nothing else to look forward to now. They are time travelers trapped in the perpetual present. What a brilliant metaphor. In the end, the unconscious (or 'id'), time loose, pre-individuated, and primordial, is the only force liberated in these cul-de-sac enclaves - to run amok in a series of scenes that look like mosaics in a Byzantine church: visions of Purgatory. Ballard never lost the facility for language and description even as his literary desires shaped atrocities that the atavistic scientists of his earlier fiction would have found cold and sterile. As the wealthy gather around the burning ruins of a villa high on the cliffs that they have set fire to simply to rekindle a sense of the atavistic and magical in their lives, we experience the next five minutes of our collective future. Only the greatest artists can capture that reflection. Ballard's brilliance lies in capturing the marooned time travelers under a breath-taking, Mediterranean sky.

The Real is not Visible

2009-08-11T19:21:00.000-07:00

We often mistake the visual world around us as the "Real": the actual. The is a combination of several factors: habituation, the embedded anticipatory epistemology by which we try to make sense of things, and a multi-layered cultural semiotic that pre-filters our experiences. I think this is the underlying intuition of Jean Baudrillard in his (in)famous Theory of Simulacra. We experience a copy of the Real mediated by models, preconceptions, and generous quantities of cultural reinforcement. We experience the copy (simulacrum). Is there any reason to believe something that, at first glance, appears so ridiculous that it barely deserves our attention? I think it does. For me, the experience of the 9/11 terrorist attacks on the Twin Towers drove Baudrillard's basic point home: I watched the second plane in "real time" (televised) hit. My initial reaction, prior to understanding the full horror of what was happening, was to think: this doesn't seem very realistic. Where was the massive, cinematic explosion and sound effects? The "Hollywood Effect"? Was this staged? Only after that brief moment did I realize that insidious manner in which my perception of real events was being colored, quite unconsciously, by the expectation of a full Hollywood/Special FX production. The real was, somehow, judged insufficiently "true", a shadow of itself. This was extremely dangerous, since I was substituting the Real with the Imaginary - the realm of cinematic, sculpted experience for the horror of what was happening to those poor people in the towers and elsewhere. In that second, though, I experienced reality copied into a cinematic language that actually kept me from connecting with the Real. The language of occlusion and diversion: self-referential and self-replicating.

I found another example, but with a much different result, in watching the brilliant film Tokyo Story by Japanese filmmaker Yasujiro Ozu. In eschewing all of the devices of standard Hollywood cinema language, Ozu's film transcends the hieratic and deeply ingrained "syntax" by which most of us have been raised to watch and perceive film. Gone is the moving camera and quick cut, the viewer is "fixed" in place by Ozu. Gone is the exposition and visualization of key elements in the story: we only see the echoes and effects of life changing events (like the mother's sudden illness and death) on human faces. The Real reemerges: the emotional, psychological and experiential dimensions of our being in the world and being with others (to use two of Heiddegger's terms from Sein und Zeit). We reassemble the totality from partial information - just like we do as human beings in our normal lives. There is no safe, omnipresent and comforting illusion that weaves everything together. We are simply there in the constructed moment but one whose artifice we can appreciate and sense. It does not keep us from "seeing" and understanding the human drama of Ozu's characters. It amplifies the human scale. It takes a displacement of our habituated expectations about what we will see, feel, and experience to move us past the simulacrum to the Real. Tokyo Story is one of the most moving films ever filmed: it defies while at the same time embodies the traditional aesthetics of Japanese art (wabi/sabi, mono no aware, etc.). This is the power of transcendence through art. You can be taught a new "Way of Seeing" as the critic John Berger called his landmark BBC series on art. Ozu manages it beautifully.

On the one hand, we have the the deeply invidious perception warping of the 24 hour news channel's simulation of reality (complete with stirring theme music for Tomahawk missile strikes against remote and unseen "enemies"). Reality as televised video game: rendered totally invisible by the audio-visual semiotic. On the other hand is the humanizing and poignant view of Ozu revealed in his film. I think the only counterforce to the power of the simulacrum is art. Only art can reveal the Real - the one thing that has truly been made invisible by our global systems of mass communication and entertainment. Art is the process of seeing the invisible. The Real is invisible to us unless we use a tool like art to draw it back into the open. Globalized entertainment has simply meant the displacement of most forms of "seeing" by one syntax and framing narrative. The implications of this are vast. The copy replicating everywhere and passing itself off as the Real.

Perhaps this is why the Aboriginals of Australia place such a value on the Dreamtime? The co-presence of the Real beyond our fragmentary and partial perceptions of time. It is one of the most original and powerful theological concepts that I have ever encountered. They must be watching a great film sub specie aeternitatis.

Large-Scale Kernel Machines

2009-05-10T16:19:00.001-07:00

Support vector machines, kernel methods and optimization algorithms working at large-scale is a fascinating convergence of mathematical intuitions and engines to deal with the huge volumes of the data being harvested from the Web by search engines (and their successors). This volume offers an excellent overview and deep dives into the techniques of scaling SVMs, KMs and algorithms from mathematical programming to the problem of online, stochastic and large-scale (data-wise) machine learning. A superb collection that is definitely worth the investment of time to study thoroughly.

Dreaming of Tilda Swinton

2009-02-27T01:01:00.001-08:00

I seldom remember my dreams these days but last night was an exception. I found myself co-starring (or was it a walk-on?) with Tilda Swinton in a horribly dubbed indie political thriller called "The Curse of the Conquistador's Gold". I am not sure which language it was originally done in but the strange argot that we were speaking to one another "on screen" was a marvelous pile-up of Old Church Slavonic, Mayan and Spanish. The dubbing to our mouth/lip movements was incredibly poor - we are talking Gigantor or Ultraman appalling. It felt very direct to video/DVD/Netflix. The primary action in the scene seemed to involve endless time spent spent loitering Godot-like at a remote bus stop high in the Andes waiting for a bus to pick us up. The bus arrived, the top half some atavism from a 1960s commune and the bottom a shimmering, blue-tinged, maglev system that seemed to run on the collective frustration of the people inside it. We wound up chasing it for 20-30 meters desperately trying to remember the combination to the bus's door. As Koji Suzuki said in a recent novel, "When you know that you are dreaming, then you are about to awake." Sure enough, my iPhone alarm woke me. I never did see any gold or Conquistadors but Tilda looked awesome.

RjDj: Reactive Sensory Music

2009-02-20T07:48:00.001-08:00

RjDj is a genre-defining iPhone application that adds the critical reactive and environmental component to generative music and sound design. There are various scenes that can be loaded which amount to different DSP processors for sampling the ambient environment and creating everything from beats to hyper-modulated soundscapes. The analogy to Electroplankton is closest, in that you choose a style of algorithm (I mean plankton) and start reacting and interacting with it as it performs its functions and behaviors. RjDj's scenes are an interesting combination of vocal amplification and filtered distortion to more soothing and blissful audioscapes. Once again, the iPhone proves to be a great genetrix of creativity. The mobile reactive sound factory and improvisational audio lab - quite an achievement.

BTW, RjDj uses Pure Data (Pd) as its DSP engine. It shows that you can take the core engine and wrap it in something far more attractive than Pd's native - ahem - UX. Pd's native UX makes MAX/MSP's look approachable and easy. I have never been a fan of visual patching languages and the only one's which work require a lot of thoughtful UX design: Reason and Korg's DS-10 being good examples.

Laeterna Machina: Body-Machine-Music

2009-02-18T09:37:00.000-08:00

The use of the body augmented by sensors and tied into generative music and sound design technology is actually quite intriguing. Although results can be varied, the idea of turning the 'poetry of motion' into electronica and soundscapes is a powerful framework for self-expression. I like the fact that they utilize MAX/MSP and pD (Pure Data) - which really are the basic building blocks for this style of bio-generative art. The resulting musical performances are available on iTunes. I applaud this kind of effort to meld traditional art forms with the latest technologies. Art has to remove our artificially imposed barriers of perception and cognition.

Solver Foundation 1.1 and Solver Plug-in SDK Ship

2009-02-17T13:21:00.001-08:00

It seemed just 2-3 months ago that we launched the full v1.0 release of Solver Foundation. This marks the second commercial release of Solver Foundation. This release primarily delivers some significant improvements to the Excel Add-in Designer and our algebraic modeling language (OML) as well as the Solver Plug-in SDK. We have also introduced extensions to the Linear Programming Solver to support Special Ordered Sets (SOS2) modeling.

The Solver Plug-in SDK is what allows any third party solver manufacturer, whether writing in native code or managed, to quickly and efficiently integrate their solver into the foundation. We actually used this plug-in architecture to integrate Gurobi’s mixed integer programming (MIP) solver in v1.1. Most importantly, we have also developed in-house and released reference plug-ins for CPLEX, XPRESS-MP and LPSOLVE. The feedback on the SDK and Solver Foundation overall have been phenomenal.

Look for even more cool features towards the end of 2009 in our v2.0 release.

Tenori-On: Interview with the Creators

2009-01-31T16:50:00.000-08:00

The Tenori-on, as I have somewhat effusively gushed several times, is the example of a new kind of creative artifact. It exists in the space between toy and instrument and shifts the perspective of the artistic process: away from hard-coded models of musicianship and technique to pure, reflective play and discovery. It is literally hypnotizing to use and is extensible with sounds (samples). Any age group can immediately use it as a platform for self-expression and playful experimentation. In addition to Toshio Iwai's brilliant ElectroPlankton and Korg's DS-10 (complete with integrated KaossPad and WiFi sync on Nintendo DS), it is indispensable. The cost of such a unique object is completely besides the point.

How does one value pure joy? It is a "Gateway to Wonder". This elegant phrases from Daoism describes the ultimate reality of being (Sein) as pure and eternal delight: unnameable, all-encompassing, intuited and infinitely generative.

Gunkanjima Island: Reality Mythologized

2008-12-26T23:55:00.000-08:00

The images of Gunkanjima (Japanese: Battleship) Island in this online gallery are haunting - in the same equivalence class as the aerial images taken 20 years later in Chernobyl. These modern Catal Huyuks suggest some kind of allegory of late 20th Century industrial civilization. Gunkanjima was an artificial island constructed over a bare reef for purposes of housing a complete mining colony by the Japanese.

Black and white really suits the desolation of the abandoned town as well as its final year in 1974. JG Ballard wrote so eloquently about the "pure" objects of technological society in the stories collected in Memoirs of the Space Age. As they are abandoned by humanity, they revert to a mute, insular purity that casts them adrift in time. They become our visual mythology. Who will understand the purpose of rusting gantries for a Saturn V rocket 50 years from now? No one.

They will be as enigmatic as the statues of Easter Island.

Pd (Pure Data) Interactive Digital Media Composition

2008-12-24T12:52:00.000-08:00

In researching the SuperCollider and MAX/MSP environments I also ran across some other systems for interactive digital synthesis and performance. Siren is one built entirely within Smalltalk and done out of UC Santa Barbara. Pd is another. Miller Puckette's successor to MAX, Pd (Pure Data), is an interesting evolution of the dataflow- and patch-oriented editor/assembler paradigm.

I especially like the 'data structure' concept that lies at the heart of the system. Structures can be used as complex "assemblers"/DNA and "catalytic agents"/enzymes over digital signal data (read: muRNA - Musical RNA) for digital synthesis, composition and performance tuning. The dataflow programming model in Pd reminds me a lot of Prograph CPX, the visual OOPL that really was a dataflow programming model in OO clothing from the University of Nova Scotia.

Still, I think a different paradigm for performance- and real-time manipulation is required. All of the MAX-derived systems seem to end of lifecycle in complex patchboards of processors and filters. The "code" is not appreciably easier to read than C# or Smalltalk.

Generative Composition and Interactive Performance Devices

2008-12-21T20:29:00.001-08:00

I admit to being fascinated by environments such as Csound, MAX/MSP, and Supercollider. The interesting thing to note is that all provide a very procedural-generative mechanism for sound design, composition and exploration of electronica in the broadest sense: from digital VCO and filtering to microsound engineering as one might encounter in Alva Noto and Ryuichi Sakamoto's Insen. Of the lot, Supercollider strikes me as the interesting one: it is within the same complexity class for use as the others but has the added benefit of being free and possessing decent tools. There is even a rather third rate Tracker-style interface for it. More should be done to exploit the potential algorithmic-reactive sound composition. Csound has 'blue' and other environments but it still seems like programming electronica in assembler (opcodes anyone?). Now, the really interesting opportunity seems to be to combine the combinatorial power of the algorithmic and DSP modules of systems like Supercollider with truly performance oriented and interactive musical environments such as Bloom, Electronplankton, and devices such as the Tenori-on, Jazz Mutant Chameleon, etc. A WiFi device would be ideal - say along the size of a NintendoDS or iPhone/iTouch - which would obviate the need for tons of equipment and cabling - just ushering a new era of mobile/global/real-time DJing and on-the-fly composition. I think the mash-up of performance-oriented, fun, and interactive with generative-procedural digital synthesis will produce some brilliant results. Stay tuned. How did Kraftwerk put it, "By pressing down a special key/it plays a little melody"? Exactly.

Solver Foundation v1.0 Ships!

2008-11-14T15:33:00.001-08:00

Today we RTM-ed Solver Foundation. The feedback during the CTP was tremendously positive and helpful. I was quite amazed at the amount of blogosphere activity around Microsoft's newest franmework for mathematical programming and optimization. If you haven't tried it out, the link leads you to the Express Edition download. It includes our Excel2007 Add-in for Solver Foundation. I have noticed it is a very popular way for people to try out the system and many people have already built SuDoKu and ported models from other modeling languages into the Optimization Modeling Language (OML).

I feel like a proud father whose child has just graduated from college. Elated actually. I would enjoy getting feedback on the Solver Foundation - either here or on the regular discussion boards on the MSDN Code Gallery site. For MSDN and TechNET subscribers, Solver Foundation (Standard Edition) will be included in the February 2009 DVD distribution. Enterprise licensing is available via Microsoft IP Ventures group. But try out the Express Edition: it is exactly the same as the others but with some limitations on model sizes.

Best of all, Solver Foundation is free for academic and small-scale commercial usage. Our license terms are specified in the About... dialog within the Excel Add-in and within the readme.txt and installers available for immediate download.

Me so happy.

Indeterminate and Ineffable

2008-10-19T14:33:00.000-07:00

These are a series of essays and ruminations on one of my favorite contemporary artists, Olafur Eliasson. He is a constant visual inspiration and communicates more in a single visual piece than most writers or philosophers can accomplish in an entire book. The mind understands visually and emotionally. His work is the slow cinema of light falling on our place in the surrounding arena we call life. In many of these essays, the perceptive critics calls out one of the most important aspects of the artwork (German: Kunstwerk): the freeing of the indeterminate and not-to-be-captured-in-language nature of reality when prismed via art experience.

Beyond makeshift ideas of totality, a common synonym for reason and discursive rationality, lies the open and ineffable. We feel and sense it in our everyday lives but we cannot articulate it. Sadly, we are trained to reflexively reject such things as pre-rational or unimportant. But human nature is subtle and the "unsayable" actually opens a backdoor to the Real (the "Open"). We appreciate the Open precisely because we first try to capture it in a surrounding or context. But it slips out of our grasp and in that momentary event, for which I shall use the German term Ereignis used by Heidegger, we experience the unspeakably vast and beautiful world. What I term the Real. Language naturally fails to capture the experience ex post facto. At best, it approximates or paraphrases, but never does justice to the experience. So, in a sense, art is an experience machine that never produces the same result twice. The observer is entangled but the observation does not ever succeed in reducing the event (Ereignis) to a thing, concept, or label. It merely demarcates the hazy boundary where our sovreignity dissolves and the Open begins. The Open can be the emotions of other people, the profoundly inhuman forces of Nature (what Kant would have termed the sublime), and even the exquisite and temporary beauty of a moment in time shared with friends in a special location (note the metaphysical paintings of Caspar David Friedrich). Some of Eliasson's pieces remind me of Paul Bowles's Sheltering Sky and the passages about what that sky shelters us from. Unlike Bowles, I think the other side of the sky is exactly what human beings crave rather than what they need to be sheltered from.

Emergent Semantics: Start Simple and Small

2008-10-16T21:25:00.001-07:00

Although the article starts with a bit of a provocation, I have to admit to concurring with the author. Semantic Web(tm) really is a version of "Intelligent Design" for semantic engineering. Emergent Semantics(tm) takes a far more bottom-up, evolutionary approach. This article uses Flickr automatic tagging as an example of creating meaning via tagging (descriptors), microformats, etc. You gain the benefits of lightweight semantic markup without the hassles, compromises, and blindspots of full-on ontology engineering. Emergent Semantics also scales witht he web - stream-oriented processing of media and tagging and microformat injection/extraction creates enough binding sites (to use a biological metaphor) for human beings and learning systems to be able to extract the basic contours/shape/significance of a stream.

Incidental Music for Imaginary Videogames

2008-10-05T22:56:00.001-07:00

2008 nearly draws to a close with the latest electronica from Pansentient having just been released. Entitled 'Incidental Music for Imaginary 8bit Video Games', it is a purposeful return to minimalist-8bit era music for machines designed to run at 1Mhz. One imagines a SID programmer coaxing 'Eternals versus Brutals' out of a Commodore 64 with a Tracker program - while also tracing out sprites on graph paper to be later painstakingly entered via a 6502 macro assembler in hex. Speed and memory efficiency were everything (I guess things haven't changed all that much).

Sticking to an overall theme of space carnival music meets slightly detuned Kraftwerk analog synths, it is a homage rather than a mockery of the era when a single person wrote the entire game: AI, music, sprites, game logic, etc. It mixes references to Alfred Bester and Zardoz with an equal portion of 70s-era Goblin synth work as hallucinated by William Blake (the first Ancient Astronaut).

The entire release was created and realized on a Tenori-on at Studio Castalia. The Tenori-on is the remarkable performance instrument designed by Toshio Iwai as a follow-up to his brilliant Electroplankton software for the Nintendo DS. Long live the artist-creator in all eras working with the technology at hand. Energy is eternal delight.

Microsoft Solver Foundation CTP

2008-10-01T21:43:00.000-07:00

The team has finally released the full Solver Foundation for community preview. This is complete with primers, API documentation, tools for loading in existing models (via MPS and QPS), a rich Excel2007 Designer Add-in for developing optimization and mathematical programming models, and enough samples to start anyone off - whether they are C#, F#, Python, C++, VB.Net, or Excel modelers and developers.

The CTP does have some limitations on the model sizes that it will accept. These are worth calling out: 50K for LP models, 1K for MIP models, and 50K for CSP (Constraint Programming) models. These are just limitations placed on the evaluation edition of the solvers and not a reflection on the upper bounds of the solvers or framework and services. Solve Foundation has scaled to several million non-zero coefficients in LP models without any issues.

So, please try out the CTP. The full commercial release of Solver Foundation is set for mid-November 2008. The team and I will be attending the INFORMS Conference in January to speak about and demonstrate the Solver Foundation and its rich offering of solvers, services, and programming tools and APIs.

Simplicity by Chopin

2008-09-19T22:34:00.000-07:00

“Simplicity is the final achievement. After one has played a vast quantity of notes and more notes, it is simplicity that emerges as the crowning reward of art.”
-- Frederic Chopin

The /Gestell/ of Technology and Bernini's Sculptures

2008-09-19T22:10:00.000-07:00

My thoughts recently returned to an essay by Martin Heidegger, "The Question Concerning Technology", and its central concept of the /Gestell/ ("enframing") of instrumental rationality. The essay can be easily dismissed as a dense, elliptical, and meaningless concatenation of philosophical terms of art applied to a phenomenon that Heidegger did not "really" understand. I think this is a mistake. Writing style aside, Heidegger makes a vital point about the worldview generated by technology - the instrumental perspective that places all of the natural world in a relationship of tool or reserved-for-use to humanity. This relationship is quite invisible since it is part of what a philosopher like Searle would call our "Background" - the tacit posits about the world and our relationship to it that we cannot question it unless confronted in a jarring or thoughtful way. Like fish drowsily swimming in water we don't notice but rely on, we enframe the world with an attitude that seeks to control, reduce, and utilize it. It takes experiences, like those found in art, to break us out of the "frame" - and look at something without the foreclosure of the art object's being into a thing (mere presence in our visual field). Usually, we ascribe the beautiful or sublime to such moments - where we experience the world or something in it as special or beautiful. This essay of Heidegger's is a complement to his other essays on art and poetic language. For me, an occassion where the Gestell was "broken" for a moment was seeing the sculptures of Bernini in the Villa Medici in Rome in 2004. The beautiful is a weapon against the subtle tyranny of the visible (and obvious). I thought of Heidegger after that visit and really had a deeper appreciation of his aim in this essay on technology.

Electronica by Pansentient

2008-09-08T21:04:00.000-07:00

This is a link to electronica that I record under the alias of Pansentient. This happens to be the Snowflake EP from 2006 containing trip-hop and old school electro mixed with found sounds from Japanese language tutorials and 1970s apocalyptic cinema. All of the compositions were done entirely within Reason 3.0 and mixed directly to hard disk. No natural instruments were harmed in the process.

Look for the upcoming 'conceptual' soundtrack to the film 'Fahrenheit' later this year.

Microsoft Solver Foundation: It's Here!

2008-08-29T18:27:00.000-07:00

I am very proud to announce the release of Microsoft Solver Foundation today on MSDN. Since founding the project at Microsoft, the team and I have been busy working to get it ready for release. This is the first public outing: our "Express Edition" for scientists, researchers, students, etc.

Solver Foundation is a framework of solvers and modeling services enabling planning, risk modeling, and scheduling for .NET developers. It is integrated with the full power of the NETfx 3.5+ platform including LINQ for declarative binding to enterprise databases. Solver Foundation is delivered in a single, compact, CLS-compliant library. This managed code library may be used from any modern CLS-compatible language (F#, C#, C++, IronPython, etc.). It aids quantitative analysts (quants), business modelers, and line of business programmers in making feasible, near-optimal, and optimal decisions in business critical settings via applications or Office-based solutions. Solver Foundation ships with several production grade solvers and provides easy third party solver integration. A Solver Foundation program is a declarative model embedded in familiar NETfx design patterns and development environments (Visual Studio 2008). The model is solved by application of numerical and symbolic solvers, meta-heuristics, constraint processing algorithms, and advanced local search techniques. Included in this release are model pre-solve and validity checking. These Solver Foundation services may be leveraged by any of the solvers and provide a rich set of tools to aide to modeling, solving, and post-optimality analysis. Solver Foundation provides these scalable and performance-driven solvers and services while supporting integration with industry standard modeling and serialization formats. This permits users to leverage existing modeling investments directly within Solver Foundation-based solutions.

From Office Excel2007 workbooks to NETfx 3.5+ enabled applications, Solver Foundation provides a new class of services to users developing complex planning, risk, and optimization models.

Electroplankton (Music Through Interaction)

2008-08-27T09:51:00.000-07:00

I cannot believe that I never blogged about my favorite musical composing and sonic discovery tool. Toshio Iwai has created the most brilliant, interactive software for the Nintendo DS. Both Mina (age 6) and I are both major users. The ability to create music through drawing and experimentation cause a level of deep immersion that exceeds that (for me) of most videogames. Hours can be spent sending the various species of musical electroplankton around the screen. I shall be using this tool along with Reason and Max/MSP (v5) in my new electronica project. Mina will be credited as co-creator on a track or two - armed with her Electroplankton DS "synth". Luminaria and Tracy are my favorites. Childlike and deep exploration of the aesthetic realm co-exist happily. This once again underscores my contention that the Nintendo DS is *the* current device for supplanting the PC and mobile (iPhone) form factors with a new platform for play, competition, learning, social communication, and discovery of new "modes of creative expression". 75M sold world-wide and doing well. How many iPhones have been sold? Hmmm....

Social Connection via Dynamic Visualization (MoMA)

2008-08-27T09:11:00.000-07:00

I Want You to Want Me by Jonathan Harris and Sep Kamvar represents a clever and dynamic visualization platform for social "hook-up". The conjoining of baloons generates thought balloons and causes dangling pictures of the people appear - allowing for a strange and alluring mixture of Electroplankton meets Speed Dating. I think it is a harbinger of the clever and real-time possibilities of graphics rich interactive environments to address human needs for connection and discovery of one another. This work was commissioned by MoMA for their Design and the Elastic Mind show. Now, what if Olafur Eliasson put his mind to building such devices and teamed with Tadao Ando to construct a "living space" of aesthetic and social "encountering"? I think the results would be phenomenal.

Jack Kirby: Mythologist of the Fourth World

2008-07-26T19:26:00.000-07:00

I just read through the collected story arc of what Jack Kirby called the "Fourth World". It is difficult to contemplate the scope of storytelling he undertook in a medium that was (and still largely is) replete with lurid, hyperventilated tales of Spandex(tm)-ed superheroes squaring off against a rolodex of one-dimensional world conquerors - all endlessly recycled through "retcons" that reset the whole sorry mess every decade or so.

Kirby was different. He created a mythology that is still repackaged and reused to this day. Although skewed towards a Manichean world view (but hey, so are most sagas and myth cycles) he managed to extract pathos and wonder from four-color offset printing and the strictures of the comic industry of his time. Orion, Darkseid, MotherBox, Boom Tubes, etc. The whole closed universe of mythic storytelling.

Even in rubbish like "Jimmy Olson - Superman's Super Pal" (it must be seen to be believed since words cannot adequately capture the puerile nature of those books) one sees the feint thread of his saga - pitting Father against Son, Choice against Fate, and Man against "the New Gods". Mindblowing.

There are only two great comic artists and storytellers from the 60s: Jack Kirby and Steve Ditko. Both were overshadowed by He-Who-Shall-Not-Be-Named commercially and saw little return on their efforts. But it does not matter, they remain eternals.

Super Mario Galaxy and the Evolutionary Biology of "Play"

2008-07-26T19:16:00.000-07:00

This is a short and interesting monograph. It tries to establish a framework for interpreting gaming and refreshingly does not try to reduce it done to a lame post-structuralist "reading" trope. (i.e. "reading the inter-textuality of World of Warcraft" - wtf?!?) Instead it provides a ludics-centered model. I wish to applaud that even if I am not entirely convinced of all of the conclusions drawn by the author Galloway.

Ludics is engrained in the evolutionary biology of mammals - as both a teaching and socializing mechanism - so that having a modicum of critical analysis of it in the context of something like Super Mario Galaxy (Wii) is entirely appropriate.

Surface and Symbol

2008-04-12T19:34:00.000-07:00

My wife and I are launching a new site for our design work. She is a fashion designer concentrating on children's clothing while I primarily dabble in electronica compositions and the odd monograph. Anything from hermeneutics on Gadamer of his writings on the /agathon/ (Beautiful) in Platonic Dialogs to the writings of Michel Houellebecq to the cinema of Yasujiro Ozu.

In fact, a short monograph on Ozu's late cinematic efforts is close to completion. Check out the site. The online catalog should be available soon.

Nintendo DS and the OLPC

2008-01-03T09:40:00.000-08:00

This Christmas we purchased a One Laptop Per Child device (OLPC) and a Nintendo DSlite for our five year old daughter. The Nintendo DS (NDS) is the clear winner as an interactive device and learning platform. It truly is a paradigm shift for UI/UX. Our daughter was able to figure out how to configure two DSs for PictoChat (via WiFi) and now constantly wants to "IM" pictures and text with me. It's an eye-openning experience.

The OLPC is the utter disappointment though. Everything about it is sluggish, unresponsive, cryptic and just sub-optimal. It's a device I really wanted to love given the integration of so much multimedia software and the new UI called "Sugar". But it is irredeemably slow - applications take upwards of 30-40 seconds to launch. The device itself is far from the DS's instant on - it takes about 40-50 seconds to bring up the Sugar "main" stage. The applications themselves are poorly designed and distinctly amateurish in feel and usability. The Web Browser is almost unusable even over fast WiFi. I have a faster experience running my iPhone over AT&T's GPRS network. Even the python editor is lame. I hate to say it but a Commodore 64 had a more intuitive UI out of the box. More importantly, the device costs nearly $180 dollars (well, $425 since I had to donate one)
. WiFi access and configuration was a pure "press random buttons until you figure it out" affair and the antenna and WiFi performance is trounced by the NDS.

The Nintendo DS on the other hand is really very close to an ideal device for the same market targeted by the OLPC. Its BOM cannot be more than $30-40. the WiFi support in the device and most applications (ahem, games) are fabulous. This includes downloading via WiFi games to other NDSs - so that multiplayer support is well-conceived and executed. The viral marketing is justa bonus. The NDS also boosts a throughly touch and gesture-based UX - far more radical than Sugar's - which is really a simple modal menu-ing system with radial layout of buttons/menu items. In a game like Zelda: The Phantom Hourglass - you get gestural UI, voice-based UI, and touch-oriented interaction with the addition of multiplayer mode. The relatively low-power processor and sprite hardware are really put to good use - there is a night and day difference between the NDS and OLPC. The NDS delivers a cinematic experience with true interactivity. teh OLPC delivers a very - how shall I put this - "academic" idea of what people (children) will want and like. The NDS was clearly tested and usability done on each aspect. The OLPC looks like all the decisions were hardcoded early on by a brain trust of "experts" without any thought of the actual experience of using, maintaining or enhancing the device.

As we try to propagate the technology/IT revolution to all markets it strikes me that the NDS is a better baseline/reference design than the OLPC. it has a built-in virtual keyboard, support for English and Japanese (while, Katankana and Hiragana), voice/IM/chat, WiFi sharing and play and more than sufficient "experience engine" (XE) to generate educational as well as entertainment solutions. The upgrading model, cards, is also a great way to produce self-contained applications without "DLL Hell" or the other problems of doing general purpose legacy-oriented OSs. NDS's OS, I am certain, is a minimalist affair - as it should be!

NDS is thw clear winner with the five year old and a great tool for me to teach some aspects of UX and computing to her. it's also a great medium for father and daughter to share experiences and remove all "technology phobias". For $125 retail that is a true bargain.

Watching Bird Flocks: Turbulence and Emergent Order

2007-05-22T10:17:00.000-07:00

My daughter Mina and I were driving this morning to school and work when she noticed a flock of birds circling over her school. It was quite jarring to see the state transition from turbulent, semi-ordered motion (circles of epi-circles of birds flying at various elevations and tryign to /stabilize/) to a nicely ordered /line/ of them start off in one direction. The line is approximate but definitely visible. An order was present that was missing just seconds before. It really drove home the point of emergence visually. Of course, it took my daughter noticing them out of the car window to draw my attention. Chrildren simply see more because they experience more purely their immediate environment. It takes years of epistemic accretion to bury the sense of beauty beneath cognitive stereotypes and mundane /models/. This, of course, got me thinking about the mathematics of describing that dynamic process and Craig Reynold's groundbreaking work on Boids and flocking algorithms. He identified three main rules in the flocking systems - which are broadly applicable beyond just bird flocks (think fish, insects, etc.).

Basic flocking is controlled by three simple rules:

Separation - avoid crowding neighbours
Alignment - steer towards average heading of neighbours
Cohesion - steer towards average position of neighbours

I really amazed repeatedly by the beauty of such elegantly simple systems to extract order out of a dynamic system without any /master controller/. So much fothe software and IT systems we build require the /ghost in the machine/ figure coordinating everything. This isn't how Nature works. Why are we surprised then when our IT systems hit scalability walls? This morning it took a child's interest in birds flying over our traffic-jammed car to remind me of the rich world outside the scheduled commute and overly ordered routine of adulthood.

Patterns and Generative Algorithms of Architectural Design

2007-05-08T12:32:00.000-07:00

This short work is another example of the confluence of algorithmics, aesthetics and engineering - in this case architectural engineering of space. The authors deploy several forms/patterns and define them algorithmically as well as illustrating them visually while providing a design context. It is a delightful treatise/monograph and somwhat along the lines of the Atlas of Tectonics. The description of the text provided by the publisher includes: "In Tooling, the latest installment in our renowned Pamphlet Architecture series, the technologically progressive young firm Aranda/Lasch illustrates how advanced computational methods and algorithmic codes can be used to foster architectural design. Tooling explores patterns generated by computer codes that in turn create an organizational template assembling projects."

Mathematics is beautiful.

Logic Programming: Why has it Failed?

2007-02-13T13:29:00.000-08:00

Someone recently asked me about Concurrent Prolog and whether it would make sense to do a version for CLR/.Net. Although I am a serious partisan of goal-directed and data-directed programming, I have to admit that the pure logic and functional programming schools have always left me feeling suspicious of their claims. Restricting my comments to Prolog and its derivatives, it seems that there are some insurmountables to their adoption as mainstream languages.

1. Horn Clause syntax is just too unfamilar and cumbersome for "really simple tasks" and many people get confused by aspects such as the treatment of negation, clausal-ordering effects on programs termination (this last one is a major flaw and show-stopper), use of "cut", etc.
2. Pure logical variables (pattern matching) is non-deterministic in ways that frighten many programmers and require a different way of thinking about value assignment, scope, and binding
3. The pervasive use of recursion and search in program execution for Prolog and derivates is a steep learning curve for many trained in languages of the C family with its emphasis on iteration and programmer controlled statespace "exploration"
4. Debugging Concurrent Prolog is a significant skill and requires quite a bit of training (costly)
5. The majority of programmers are queasy about determinism (don't care and don't know varieties) and only in a few cases (such as SQL query processors and optimizers) have programmers been convinced to "let the runtime do its magic" and control execution semantics, operator application, etc.

There are educational projects such as PSharp and PJava out there but they remain academic curiosities. The real action, in my opinion, is in goal-directed programming informed by mathematical programming techniques as well as constraint programming. Constraint programming can be seen as a major evolution over the fixed execution semantics and proof strategy of Prolog (DFS, simple backtracking, cut, etc. of the standard WAM). Constraint processing when combined with the appropriate "phrase structure grammar" for a domain can make all of the difference between academic curiosity and widespread adoption (look at Outlook inbox rules).

Structure and Interpretation of Computer Programs

2007-02-05T15:34:00.000-08:00

A link to the full online version of the best theoretical computer science book published for undergraduates. Section 2.4.3 on data-directed programming is worth careful study. Nothing in the text is especially Scheme-specific, the techniques transcend the programming language with reasonable modification. The effects on the reader's/programmer's "way of seeing" (to paraphrase John Berger) are profound.

Generative Grammar of the Immune System

2007-01-22T12:43:00.000-08:00

This is fascinating application of Chomskian generative grammar concepts to the study of immunological system behavior and function. This article is the 1985 Nobel Prize speech by NK Jerne. It might be interesting to take recent work on principles and parameters by Chomsky, Baker, Pinker and others and update this interesting "cross-pollination" of medicine and formal linguistics.

Post-Fregean/Tarskian Semantics based on Speech Acts

2007-01-08T13:00:00.001-08:00

This is an absolutely fascinating study - one which seeks to reground semantics (Fregean Semantics) on a purely speech act foudnation. It is something that I have often considered given the various intractable areas of natural language processing (semantics, not just syntax) that Fregean theories wind up in - even Tarskian theories of meaning like Davidvson's Convention-T. I applaud the author in this line of research. The book is called "Renewing Meaning: A Speech-ACT Theoretic Approach" A description follows: "The framework he offers is a speech-act-based approach to meaning in which semantics is entirely subsumed by pragmatics. In this framework: meaning resides in syntax and pragmatics; sentence-meanings are not propositions but speech-act types; word-meanings are not objects, functions, or properties, but again speech-act types; pragmatic phenomena one would expect not to figure in semantics, such as pretence, enter into the logical form of sentences; a compositional semantics is provided by showing how speech-act types combine together to form complex speech-act types; the syntactic structures invoked are not those of quantifiers, open sentences, variables, variable-binding, etc., rather they are structures specific to speech-act forms, which link logical form and surface grammar very closely. According to Barker, a natural language - a system of thought - is an emergent entity that arises from the combination of simple intentional structures, and certain non-representational cognitive states. "

Dimensionality for Kernels Methods

2006-12-26T10:04:00.000-08:00

This is a good collection of talks and presentations for Kernel Methods and classifier development specifically focused around the topic of grammar induction (learning) and classification. Also worth noting is an interesting confluence of several of my interests - namely statistical learning methods, natural language processing, and convex optimization. The latter may seem a tad unintuitive but it really is there when you peer closely at the "maximal separating hyperplane" aspect of the learning kernel. How does one calculate the optimal separating hyperplane? So once again, learning and mathematical programming (convex optimization) meet in a semantico-linguistic domain. The first presentation is a good overview and pro/con of kernel methods. Other papers cover Hidden Markov Models (HMMs) - the old standbyes of grammar induction.

Online Stochastic Combinatorial Optimization

2006-11-10T15:28:00.000-08:00

This is a recent publication from Pascal van Hentenryck of Brown University. It attempts to define a framework for online (aka real-time, temporally constrained) decision making with exogenous uncertainty (versus the plain-old-endogenous kind modeled in MDPs and multi-stage stochastic linear programming). Worth a read, as are the author's other texts on COMET, OPL, Numerica,a nd Constraint Logic Programming (all by MIT Press).

Dynamic, Generative Art from Email Spam

2006-07-26T08:52:00.000-07:00

alex dragulescu - dynamic for the people This is a marvelous example of the confluence of aesthetics and information visualization through applied algorithmics. The Romanian artist has developed visual rules for pseudo-plant growth and morphology based on email spam traffic. Could this be a glimpse of the future of "glanceable" information (ubiqitous computing) where passive monitoring of systems happens through simple visual metaphors (like a strange plant growing in your electronic garden) rather than the state of play today - lists and readouts of ports, traffic levels, etc? This would permit a qualitatively more natural simulation of human information processing since the categoris of normal visual perception (identifying odd or anomalous forms) is fully harnessed but in a subtle way. This works reminds me of the pioneering work on Lindermeyer Grammars (L-Grammars) developed in The Algorithmic Beauty fo Plants. Here, the L-Grammars are being used to project information DNA rather than plant DNA. Beautiful work.

Haptek PeoplePutty(Tm)

2006-07-25T11:17:00.000-07:00

www.haptek.com - Haptek, Inc.This is an interesting model for the UI for small form factor devices (mobile phones) for SecondLife(tm)-like environments. Rather than forcing a UX model of moving tiny sprites around a realm, move the majority of interaction to locations/stages that are populated by Haptek-like figures. The emotiveness simulated (and I must say fairly reasonably) is a strong anchor for people. Imagine in-world/realm chat and so forth with these Haptek-like interfaces. I think this is quite clever.

Algorithmic City Generation and Pseudo-Randomness

2006-07-25T08:57:00.000-07:00

monoform.org | Vis 141b - CityScape: An OpenGL City Simulation This is another self-generating cityscape project. Again, algorithmics are applied to define the layout, area popualtion with artifacts, and simulate a natural (though manufactured) environment. What about algorithms for non-natural (non-terrestrial) environments? What would the L-grammar look like for a self-replicating and self-modifying 'city' (enclave, arcology, whathaveyou) orbiting Earh at a Lagrange point? How would the physics of the simulated environment be used to mould the self-generating cityscape? Defining the minimal grammar for such environments would be quite intriguing and good step towards a self-generating 'playspace' with lower/minimal costs for content generation and maintenance. Because of the pseudo-random number generation involved in many of these projects - the storage of 'seeds' for the Random() function could be persisted and accessed as part of the "growth" process- so that a large online environment could be regenerated from pseudo-random "first principles"?!? Now that's skirting some of Chaitin's work, infinite and periodic spaces like the Library of Babel in Borges's eponymous short story.

Thematic-based Search with Vivisimo

2006-06-29T14:03:00.000-07:00

Vivisimo founder interview This appears to be an interesting combination of semantic tagging and clustering via "themes". If the themes are extractable in a rich way that preserves sub-clusters then this may in fact prove to be quite useful. Simple clustering for web-based search is too coarse grained. Mtc only.

Blurring the Order of Simulacra: Second Life and Flickr Mashups

2006-06-29T13:33:00.000-07:00

Wonderland: Second Life & Flickr mashups First order simulacra meet third order simulacra (to use Baudrillard's taxonomy). The line between RealSpace(tm) and Overworld(tm) is blurred as the integration tools and protocols allow for simple cross-over from real to Internet/WWW to virtual spaces. This relative ease of integration of content presages the larger integration of sensors and actuators in RealSpace that can effect, be manifested in, and effect the Overworld (virtual space). This seamless blurring and eventual continuity of experience is important to produce the deeply emotive (affective) relationship of the user to the simulated environment. Now imagine auto-generating virtual spaces (ala Scalable City) incorporating real-time data from sensors and physical devices carried around by tens of millions of users. The cell phone might well become the"Eye of the Overworld" to paraphrase Jack Vance.

AllegroGraph 64bit RDF/SemanticWeb Store

2006-06-23T14:31:00.000-07:00

The Semantic Web RDF standard from W3C is making increasing inroads into government, the enterprise, bioinformatics, telecommunications, and other demanding arenas. Its simplicity and flexibility make it suitable to represent any structured or unstructured knowledge (be it call data networks from telecom providers or intelligence data from Homeland Security). However, for real-world applications the number of RDF triples can easily grow into millions or even billions, making it difficult to process efficiently with traditional tools. Even working knowledge (data) sets often contain millions of triples. Systems that must load, manipulate, and query such enormous triple data stores require the best possible performance. AllegroGraph efficiently performs the three most important tasks for a triple store: to load, store, and query data.AllegroGraph can load gigabytes of RDF data in minutes.Storage is persistent, including between application launches in on-disk binary trees. There is no additional serialization or deserialization overhead.
Querying is both flexible and performant. Multiple indices support fast access through a simple triple-level API, Allegro Prolog, or SPARQL (the emerging W3C standard RDF query language). When querying for a particular subject with ten triples, AllegroGraph can retrieve about 40,000 triples per second, from disk.

S-CREAM: Semi-Automatic Creation of Metadata

2006-06-19T12:45:00.000-07:00

This is an article describing a system of semi-automated metadata induction based on training for a domain. It appears to be a good potential adjunct to other taxonomy and semantic web projects. Again, being able to throw silicon at the problem is always advisable given the sheer volumes of data (and extractable metadata) being autogenerated today. Scalable tools for metadata and semantic analysis are needed. This is one technique.

Scalable City

2006-06-19T12:38:00.000-07:00

Scalable City: "The Scalable City is a set of projects that explore the externalization of algorithmic approaches to urbanization which intersect with geographic, political, economic and aesthetic zones of conflict. Version 0.8a of the Scalable City is a multi-media exhibition of landscape demarcation, personal embodiment and domicile transformations. Procedures governing the arrangement and operations of these discrete areas are interchanged across domains - moving them from a familiar basis to distorted and exaggerated extremes. Through these processes, the procedural basis underlying the development of cultural forms are revealed and the mechanisms of social formation are highlighted."

Kleiner Perkins 7 Rules for Software Start-ups

2006-06-19T12:36:00.000-07:00

I found this to be a good distillation of "rules" for almost any enterprise - start-up or otherwise. Instant, viral, low-IT, simple, personalizable, templated/configurable, and context-aware.

Computational Applications of Linear Logic

2006-05-11T15:38:00.000-07:00

This is a cogent overview of computational aspects (and implications) of linear logic. It starts with a solid overview of LL, then moves into concrete applications of LL to functional programming, logic programming, (dynamic) process modeling, and finally stateful programming. Definitely worth a read. LL's insistence on 'materializing' the operations of logical proof make it a great vehicle for thinking about how logic and proof are computations first and foremost - computations that consume resources. This is a "realist" viewpoint on logic - logic as a process within a dynamic environment/context. Substructural logics have much to recommend them - not the least of which is the jettisoning on Platonist assumptions about invarient truths and the static nature of "Being". (Sein) Girard's a dude - what can I say?!?

Concepts, Techniques, and Models of Computer Programming

2006-03-28T22:26:00.000-08:00

This is a superb book covering the compuational and programming paradigms prevalent in the industry. The approach of the authors is to build kernel languages for each paradigm - a unique technique. The declarative model is emphasized - with extensive coverage of term rewriting techniques and the operational semantics of declarative languages.

Amorphous Computing: Emergent Order

2006-03-06T15:38:00.000-08:00

This is the basic pitch of amorphous computing - which I first encountered in Michael Resnick's book Turtles, Termites and Traffic Jams (StarLogo):

"A colony of cells cooperates to form a multicellular organism under the direction of a genetic program shared by the members of the colony. A swarm of bees cooperates to construct a hive. Humans group together to build towns, cities, and nations. These examples raise fundamental questions for the organization of computing systems: How do we obtain coherent behavior from the cooperation of large numbers of unreliable parts that are interconnected in unknown, irregular, and time-varying ways? What are the methods for instructing myriads of programmable entities to cooperate to achieve particular goals? These questions have been recognized as fundamental for generations. Now is an opportune time to tackle the engineering of emergent order: to identify the engineering principles and languages that can be used to observe, control, organize, and exploit the behavior of programmable multitudes. We call this effort the study of amorphous computing."

Hidden Markov Models

2006-01-30T13:12:00.000-08:00

Hidden Markov Models

The value of Hidden Markov Models (HMMs) has long been understood. As learning machines for statical linguistics, they have been heavily studied by Cherniak and others. Essentially, the HMM is a learning machine with (dynamically) weighted inner state transitions -- states which are not necessarily visible to the outside observer. I think the most interesting application that I have run across is the use of SLT (and HMMs) in grammar induction of natural languages. With a corpus input - HMM-based learning has allowed for 'continuous recognizers' to be built. This ability to induct grammars from instances has some far ranging potential benefits - especiallya s speech becomes a more dominant (and non SF) aspect of HCI. HMMs, likes SVMs, are an interesting basis of learning and knowledge engineering.

Elegance is not Simplicity

2005-10-31T12:15:00.000-08:00

This is a slight tangent from my usual postings. It does touch an important topic for me - the need for simple, human-scale models and paradigms for "coping" with complexity, diversity, and heterogeneity and for putting the "delight" back into the heart of the development experience. Here is a nice quote that I admit to sympathizing with:

"The major delusion I want to shatter is that you can successfully use the same types and tools for solving the biggest problems in the world and have a pleasant time applying that to solving the normal problems of the world," Hansson said.

I remember Donald Knuth’s great line: “Take any problem and multiple it by ten – and you have a new problem.” I guess the inverse applies too: take any problem and divide by ten – and you have a new problem.

It seems to me that AJAX represents a very emergent phenomenon – a need-driven one. It is a revolt of the masses if you will – and a handy (but far from perfect) way of dealing with heterogeneity/versioning/configuration. People have to solve solutions today – so they will things like RoR, AJAX, etc. into existence. They desire it. There is nothing more powerful than desire.

People want the simplicity and approachability - they will trade power for comprehensibility and iterative value/feedback.

Perhaps the big insight of App Frameworks of the 1990s/early 2000s is that one-size-fits-all doesn’t? That a range of approaches can target the continuum of developers with genuinely different offerings? Remember how Openstep and Taligent were going to change the way people built apps? They didn’t. But REST/HTTP certainly did. Big time. We saw major wealth creation during the time the REST/HTTP/Web “disruption”. We’ve experienced nothing like it since (from a wealth generation POV).

Elegant does not equal simple.

My belief: People really want simple – even at the loss of fidelity or “power” (since they seldom understand how to use the powerful features anyway).

People like automatic transmissions (esp. in North America).

We’ve seen bifurcation within (Enterprise) Java Community (J2EE/EJB vs. “Lightweight”/IoC Styles), we’re seeing it in the PHP/Python/Ruby vs. C#/Java discussions, and even (to a smaller extent) in different approaches to workflow and process automation – lang. extensions (like what Collaxa did to Java) and simple state machines vs. full-on process algebras-in-XML flavors (BPEL, JPD, BPSSL, etc.).

I think the services-platform-in-sky push will require a commensurate push towards the very simple – Cockroach DNA vs. homo sapiens DNA. Ever wonder why cockroaches can survive so much radioactivity (doses that would kill human beings many times over)? Redundant, simple DNA – they can survive major genetic damage (mutation) and self-repair (they have many copies of the same coding sequences).

The future belongs to the insects.

I think we can/should apply the Simple DNA test at all layers of the stack – from storage models to application models to dynamic UI rendering pipelines. The easier it is to explain the easier it is to get people over the hump of using it.

PHP’s success (mtc) comes down to the fact that it is a syncretic but strangely “balanced” scripting language that embeds the basic tools required to do rinky-dink web/ecommerce programming by non-PHDs as primitives of the system – e.g. I dislike the syntax personally – but I have to give it props. Even Andreessen’s singing its praises now. PHP DB access is built-into the language (6-7 simple commands).

Simplicity is agility.

Simplicity is scalability.

Simplicity is approachability.

Simplicity is a design technique.

Simplicity is an aesthetic worldview.

Aesthetics lies at the heart of many of our “architectural” concepts – mostly implicit in our judgments and seldom noticed. Ideas like: essentiality, beauty, desirability, purpose, form, clarity, coherence, symmetry, mimesis – these are borrowed from our visual/perceptual – aka aesthetic – judgements. But David Gelertner has written a nice book on the subject and I recommend it to others.

I think it’s also the biggest weapon we have against the exponential space of distributed computing models and realities – if we keep the models and tools within a human frame of reference. People have to be able to follow each step of the system (even if only theoretically) – to feel like they understand what’s-going-on. Apache’s REST based model plays on that “feeling” of in-control for the (web)dev. Apache is not a panacea (since my argument really is that the market has largely differentiated poles (Simple eComm to Global/Federated WS-*-aggedon) that require different tools and mindsets. Two different customer segments – not just two different surface syntaxes for the same system and order of complexity model. Both major 3GLs today (you know who you are...) are within the same complexity class - and the marketplace still finds room for Ruby, Python, Lua, Io, Groovy, etc.

Are we making scripting a truly first class experience? Is scripting there mostly to mollify the “non-pros” – whom we seem to not take very seriously any more? Or is it the basis of a constellation of lightweight containers, edit-and-go interactive programming, and one-button deployment/packaging (closer to things like Sharepoint)? Can we return to a world where - like the 10 year old with a Commodore64 in 1982 - where computing can be fun, immediate, simple, and very beguiling? I think so.

Q Term Rewriting System

2005-10-11T22:18:00.000-07:00

This is an interesting language - which could form the basis of an interesting runtime for implementing several classes of "declarative" programming/specification systems. From the FAQ:

The Q interpreter is neither a computer algebra system nor an equational theorem prover (although you could use it for such purposes). Like other functional language interpreters, it is just an expression evaluator. However, using term rewriting as the basic model of computation is very powerful, and it allows you to formulate your programs in a high-level, concise and declarative style.

Business Logic with Symbolic Computations (aka Term Rewriting)

2005-10-04T14:49:00.000-07:00

From the introduction - a description of the embeddable term-rewriting system with lots of "extras". Worth a look - especially if you read Russian. ;)

TermWare system [RS02] is a term-rewriting extensible framework consisted of two main parts: 1) Java library which containing basic data structures and algorithms for rewriting techniques like terms, term rewriting rules , unification, rewriting strategies and others; and 2) Java framework, containing facilities for adding parsers, programming Languages and rewriting rules with actions which can be embedded into Java application and can be extended via Java framework. The main difference of TermWare from conventional term rewriting systems like Maude and APS consists in that term rewriting system is not 'closed' formal system in the sense that it does not intend to provide complete programming environment; and it is packaged not as interpreter or translator, but as a library for embedding. Unlike traditional set of rewriting rules it extends functionality for interaction with external world, represented [via a] deductive term database.

Adaptive Planning - Budgeting Forecasting Applications and Financial Reporting Analysis

2005-10-03T10:43:00.000-07:00

There is a good case to be made that a combination of adaptive, model-driven, continuous planning is possible - both from an architectural perspective as well as human collaborative one. Both sides are needed. The key differentiator - I wouold argue - is core algorithmics and feedback mechanisms embedded within the continuously optimizing system - and the degree of autonomics possible when the burden is shifted to the modeling system to "rejig"/tune itself based on data variations, new pattern induction, and "learning"-driven optimization. This requires a good business justification - but in looking at the Dells, Walmarts, and JetBlues of the world - it means the difference between being trapped in a hypercommoditized market and inching toward extinction one open source technology at a time or thriving and moving the game up oen level - towards a truly continuous, autonomic computing model that can hone and improve the declarative models (spreadsheets, policies, multidimensional datacubes) and intensions of business people.

Linear and Integer Programming Software

2005-09-26T10:50:00.000-07:00

This is an excellent reference point for various linear and integer programming software suites. Everything from Simplex variants to Interior Point methods. Also included are packages for Integer Programming (branch and bound, etc.). The convergence of Operations Research (OR) and Constraint-Logic Programming (CLP) paradigms can result in an incredibly flexible and powerful "modeling" and "solver" framework for industrial applications - from production planning, process optimization, supply chain management to distribution/transportation problems.The point of business process automation is not simply in automation and visualization - but in the economic tuning and optimization of core additive "value chains" to make businesses (and their partner ecosystems) more competitive and agile (this use a now cliche BizSpeak(tm) term).

Folksonomies: Classification and Communication Through Shared Metadata

2005-08-10T13:56:00.000-07:00

Folksonomies - Cooperative Classification and Communication Through Shared Metadata

Tagging for me is a core concept. (That and hypersets…) Tagging is the easy-to-explain version of semantic markup and encoding. “Folksonomies” is another popular term – often used (quixotically) by people who claim knowledge engineering (ontologies) is impossible to do. They don’t fully grasp that they are ontology engineering – in fact in a way remarkably similar to the cognitive development process of children. (Category induction and association rules inferred from attribute/feature tagging and recognition). Remarkably enough knowledge engineering for machine systems is deeply influenced by successful cognitive modeling of human beings. Symbolic processing is way up the stack (at OWL-level and above). This was part of the strange email I sent out the other day – talking about “objects” as attribute/feature bundles (or more precisely: as the n-dimensional intersection of property vectors) – the “thing”-ness of an object is first and foremost an interpretation of its perceived features/attributes/behaviors. You start with differentiated clumps of features (that you have glommed on/tagged) – and build the type systems up from there. We build up theories (models) of things as classes, kinds, types, etc. from there – using either classical or non-well-founded set theories. I like the latter for its power in computing systems but it’s just a preference. I was rereading a commentary on Wittgenstein’s Tractatus on Victoria this weekend – and it underscored the development of type theory in mathematical logic(s) from some basic intuitions around concept formation in children.

Introduction to Machine Learning

2005-07-06T12:24:00.000-07:00

Introduction to Machine Learning

I recently discovered Dr. Nilsson's brilliant book Learning Machines. I do not have a reference to it - so I gave a link to this draft book on machine learning techniques by the same author. Learning Machines is an exemplar in in concise, thorough, and approachable monographs for pattern classification and statistical learning. It was published in 1965 - and I recommend it as a first book on the subject. No partially debugged pseudo-Java or C# -- the algorithms are in clear structured English and mathematical notation. If you ever wanted to really understand the differences between hyeprplanes, hyperspheres, general quadratic surafces, and rth-order polynomial surfaces - this is the text.

It is a wonderful book. Dr. Nilsson is currently a Professor Emeritus at Stanford University.

Building Problem Solvers with Common Lisp

2005-07-06T12:16:00.000-07:00

Building Problem Solvers - The MIT Press

This is an intriguing approach to synthesizing constraint processing, forward and backward "chaining" rules, and symbolic algebra. Essentially all are implemented as "little languages" or engines within Common Lisp. The ramifications of this are profound - the richness of the runtime and languages (CL) radically increase the productivity of developers/solver designers to concentrate on the domain-specific aspects of their solver - and rely on the reflective, meta-programming intrinsic features of CL for their runtime. Designing and implementing a similar framework in a more -ahem- available language (such as Python) might be very interesting.

Futures and Eventual Values

2005-05-16T09:29:00.000-07:00

David Kemp's Blog: Futures and Eventual Values Part 1

The complexities of distributed, concurrent programming are usually solved in one of three ways: (a) exposure of multi-threading primitives and concurrent resource controls (semaphors, monitors, synchronizers, etc.), (b) hiding of concurrency and distribution behind a seamingly synchronous/local 'facade' (RPC), or (c) elaboration of a higher-level abstraction which focuses on structuring how the programmer thinks about the availability of data and services (aka futures, promises, delayed-evaluation). This latter approach was first advocated in the context of a concurrent variation of Lisp - called Multi-Lisp. Futures are an elegant way to separate out the 'declaration' of a distributed resource from the 'consumption' of that resource. This is critical in distributed programming environments - where global standards of synchrony, resource availability, and liveness cannot be guaranteed. So - the smartest thing to do is to greate negotiate a "contract" of sorts - when I need the value of 'x' - I either get it or I wait for it (timeouts are a part of the model too naturally). Since the vast majority of programmers write sequential algorithms - this is a great marriage of convenience - sequential algorithms slightly peppered with futures to deal with the distributed portions of the processing - without the wholesale rewrite of the entire system to support - say - asynchrony and parallelism to a fine grain.

Constraint Logic Programming Systems

2005-03-14T11:00:00.000-08:00

The ECLiPSe Constraint Logic Programming System Constraint optimization programming systems are a novel and unique way of solving problems in the interrelated fields of planning, scheduling, optimization, and prediction. Important systems/implementations include: OPL, CHIP, CLP(r), Numerica, Prolog IV, and term-rewriting systems such as Mathematica and MATLAB.

Programmable Matter and Amorphous Computing

2005-03-10T18:04:00.000-08:00

Imagine a flexible substrate, consisting of millions of tiny interwoven programmable fibers, that can be programmed to assume a large variety of global shapes. Rather than build precisely engineered mechanical structures, a programmable material would allow one to program precise complicated structures starting from a single flexible mechanical base. Not only could one design many complex static structures by programming a single substrate, but also create dynamic structures that react to, and affect, the environment. Programmable materials would make possible a host of novel applications that blur the boundary between computation and the environment.

Io Programming Language - Futures, MOP, and ACTORS

2005-02-23T15:56:00.000-08:00

Io
I have to say it - this language has just about every feature I like - prototype-based, future-supporting, metaprogramming capabilities ala LISP, Actor-based concurrency/distributed messaging, and dynamic typing. Steve Dekorte has done a great job - it runs beautifully under MacOS X.3. Rock on.

Continuations in WebApps

2005-02-23T15:48:00.000-08:00

Use continuations to develop complex Web applications
A excellent account of the limitations of MVC-based architectures and their inherent maintainability problems. It proposes (quite sensibly) to use a continuation-based model as a viable alternative. This article also contains some good programming examples and scenarios (based on Cacoon).

Carl Hewitt's Actors Model and Event Algebra

2005-02-23T15:19:00.000-08:00

LCS Publication - MIT-LCS-TR-194

This is an excellent paper on the Actor model of distributed, parallel computing. Actors represent a significantly more sophisticated approach to 'object-oriented' views of distributed programming. This paper discusses many of the 'laws' (axioms) for these distirbuted, event-driven entities.

Continuations in Programming

2005-02-23T15:14:00.000-08:00

Ruby: ContinuationExplanation

Continuations represent a *very* powerful idea for programming. In fact, in languages of the LISP family (and LISP-inspired ones such as Ruby), they become first class constructs. Why are they so cool? Becuase they allow for a high degree of abstraction of control flow - essentially allowing a program to 'reason' or 'compute' with its execution context. They allow for a natural encoding of event-driven applications without forcing the user to program a complex state machine to manage the transitions. The Ruby explanation is a very concise and practical one. More formal and rigorous treatments are possible - check out Citeseer.

Linear Logic as Basis for Resource-Aware Computing

2004-12-03T11:36:00.000-08:00

Linear Logic, proposed by JF Girard, is a resource-aware logic. By resource aware, I mean that every step in a proof is viewed as a consumption of resources (axioms, terms, etc.). This has a number of interesting aspects, not the least of which is that complex distributed compuations (read: Web Services) with complex choreographies might in fact be viewed as physical (in the loosest sense) entities. The "cost" of one choreography vs. another could in fact be modeled - openning up pathways for research into process optimization, distributed correctness, etc. Samson Abramsky's work on operationalizing LL is worth checking out, as are Phil Wadler's writings.

Kernel Methods

2004-12-02T14:30:00.000-08:00

An interesting jumping off point for research and tutorials on kernel methods. Support Vector Machines (SVMs) are an interesting case of machine learning techniques with broad applicability to many domains (classification, category induction, etc.). Vladimir Vapnik (now of Google) may be viewed as the theoretical founder. Worth the learning curve.

Imitation of Life: Biologically-Inspired Computing

2004-11-12T10:19:00.000-08:00

I can highly recommend this new book by Nancy Forbes for those interested in a rapid-immersion in biologically-inspired computing. I especially enjoyed the sections on Amorphous Computing and DNA Computing. It is a readable text and good launchpad for delving into individual research areas. I was a little disappointed that the DNA Computing section once again described the "Traveling Salesman" problem - it would be nice to see other applications of massive parallelism and search ... like cryptography?!? A brief excerpt from the MIT Press blurb follows:

"Nancy Forbes identifies three strains of biologically inspired computing: the use of biology as a metaphor or inspiration for the development of algorithms; the construction of information processing systems that use biological materials or are modeled on biological processes, or both; and the effort to understand how biological organisms "compute," or process information. Forbes then shows us how current researchers are using these approaches. In successive chapters, she looks at artificial neural networks; evolutionary and genetic algorithms, which search for the "fittest" among a generation of solutions; cellular automata; artificial life -- not just a simulation, but "alive" in the internal ecosystem of the computer; DNA computation, which uses the encoding capability of DNA to devise algorithms; self-assembly and its potential use in nanotechnology; amorphous computing, modeled on the kind of cooperation seen in a colony of cells or a swarm of bees; computer immune systems; bio-hardware and how bioelectronics compares to silicon; and the "computational" properties of cells."

Speech Act Theory for Business Actions

2004-11-10T13:34:00.000-08:00

Speech Act Theory

This page discusses a fairly simple business actions encoding of Speech Act Theory. Speech Acts (illocutionary and perlocutionary) were studied by philosophers such as John Searle and John Austin - and have been used as the foundation for theories of meaning, multi-agent system programming languages, and (now) business activities. This is an important aspect of transformational UX - moving beyond what I'll term the "lambda caclulus reduction" of natural languages and human communication to programming languages well suited for CFG encoding and execution. Human speech and communication forms an action system (as Donald Davidson has noted many times) - you use words to accomplish things. The old paradigm of language as purely a vehicle of propositional assertion ( It is the case that 'x') is a residue of a philosophical traditional that really ended with publication of Austin's "How to do Things with Words".

Emergent Semantics

2004-11-10T09:52:00.000-08:00

An interesting overview of the emergent semantics project. Hey, the guy starts off with discussing early and late Wittgenstein theories of meaning formation ("sense", Sinn) - so you know it has to be good. Anyone who begins with the Tractatus Logico-Philosophicus (TLP) has my attention. (As you might guess, it's my all-time favorite hieratic early 20th century ontology-as-logic book). The key to ES is to generate "meaning" from interaction observation and inter-play rather than basing it on modeling an invariant structure "out there". That's the emergent part... and where machine learning and data mining techniques can play a big role.

Sony Labs and Emergent Semantics

2004-11-10T09:15:00.000-08:00

Sony is taking an interesting approach to the 'semantic web': rather than assuming a flood of semantic markup via RDF, OWL, DAML, SHOE, etc., they are using machine learning and statistical techniques. This is th right way to go.