Within the pages of European Design Engineer we have previously reported on desktop search tools and CAD search utilities, but this present article focuses on something that is neither of these, nor a conventional search engine for indexing material on the world-wide web and making it searchable. The people behind Wolfram-Alpha have called it a computational knowledge engine and, if engineers try it for themselves, they may well find that it is an invaluable - yet free to access - addition to their suite of computer-based engineering tools.
Wolfram-Alpha was launched in May 2009 with the aim of bringing expert-level knowledge and capabilities to the broadest possible range of people (Fig.1). Today, Wolfram-Alpha is claimed to contain over 10 trillion pieces of data, more than 50000 types of algorithms and models, and linguistic capabilities for over 1000 domains. And there is much more to come: rather than a finished product, this is just the first step in an ambitious, long-term project to make all systematic knowledge immediately computable.
The brainchild of British-born physicist Dr Stephen Wolfram, and available for use online at www.wolframalpha.com, Wolfram-Alpha is the first product from a project that, according to Dr Wolfram, will try to take all of the world's knowledge and make it computable (Fig.2). When Dr Wolfram started the project five years ago, he was not even sure it would be possible, but the world-wide web and computing power generally had both evolved to a point where the time to try seemed to be right.
"Fifty years ago," says Wolfram, the founder and chief executive of Wolfram Research, "when computers were young, people assumed that they'd be able to ask a computer any factual question and have it compute the answer. I'm happy to say that we've successfully built a system that delivers knowledge from a simple input field, giving access to a huge system, with trillions of pieces of curated data and millions of lines of algorithms. Wolfram-Alpha signals a new paradigm for using computers and the web."
Wolfram-Alpha is paradoxically both less than a search engine and more. Unlike traditional search engines, it does not scour the web for as many references as possible that look as though they might contain the answer to a given question. And, unlike a conventional online encyclopaedia, it does not simply deliver a set of facts. Instead, Wolfram-Alpha searches its own databases for answers to questions, putting that information into context and performing any related calculations as required.
Creating new knowledge
When Wolfram-Alpha receives a user query, it extracts the relevant facts from its stored computable data and then applies a collection of tens of thousands of algorithms, creating and synthesising new relevant knowledge.
Its drawback compared with typical search engines is that it really only pulls information from its own databases. So if a user questions its knowledge about day-to-day trivia, Wolfram-Alpha struggles. That might lead users to conclude that it is of limited practical use in the real world, and certainly nothing to challenge today's established search engines and online knowledge sources. But present it with a science-based question requiring a calculation and this computational knowledge engine begins to reveal its power.
Much of that power comes from the fact that Wolfram-Alpha is built on Wolfram Mathematica, which is well known across engineering, science and other technical fields as an all-in-one computation and visualisation system, development environment and deployment engine. Wolfram-Alpha's core code base now exceeds five million lines of symbolic Mathematica code. Running on supercomputer-class computer clusters, Wolfram-Alpha makes extensive use of the latest generation of web and parallel computing technologies, including webMathematica and gridMathematica.
Mathematica has three crucial roles in Wolfram-Alpha. First, its very general symbolic language provides the framework in which all the diverse knowledge of Wolfram-Alpha is represented, and in which all of its capabilities are implemented. Second, Mathematica's vast web of built-in algorithms provides the computational foundation that makes it practical to implement the methods and models of so many fields. Finally, the strength of Mathematica as a software engineering and deployment platform makes it possible to take the technical achievements of Wolfram-Alpha and deliver them broadly and robustly.
Comparisons between Wolfram-Alpha and Mathematica are interesting. While Mathematica is a broad, deep computing environment that lets users handle arbitrarily sophisticated problems, Wolfram-Alpha gives small, quick, one-off results on the web. But the developers at Wolfram Research say that extensions to both Wolfram-Alpha and Mathematica will steadily bring the two closer together.
A new kind of science
Beyond Mathematica, another key to Wolfram-Alpha is NKS - a New Kind of Science, as defined by Dr Wolfram. His preface in the development of NKS was that mathematical equations do not capture many of nature's most essential mechanisms. But thinking in terms of programs rather than equations opens up 'a new kind of science' because, while mathematical equations are specific to particular kinds of rules, computer programs can embody far more general rules.
Many specific ideas from NKS - particularly related to algorithms discovered by exploring the computational universe - are used in the implementation of Wolfram-Alpha. It makes both conceptual and practical use of the NKS idea of generating rich, complex behaviour from simple underlying rules. According to Dr Wolfram, in many ways Wolfram-Alpha is the first 'killer app' for NKS.
The goal for Wolfram-Alpha is that it should accept completely free-form input. Its ability to understand this sort of free-form input is based on algorithms that are informed by analysis of linguistic usage in large volumes of material on the web and elsewhere. But if there is one major limitation of Wolfram-Alpha currently, it is that it is only really able to understand questions that are very carefully phrased.
However, the developers say that as Wolfram-Alpha grows, a whole new level of linguistic data will be captured, thereby greatly enhancing its linguistic capabilities.
So, setting aside those limitations, how can Wolfram-Alpha supplement an engineer's suite of computer-aided design, engineering and analysis tools? Wolfram-Alpha is adept at handling complicated mathematical queries and answering knowledge-based scientific or engineering questions. It has strengths in mathematics, physics, various engineering disciplines, materials, chemistry, astronomy and other computational sciences.
uppose, for example, that you were working on the design of an airfoil section using an NACA4351 standard design. NACA airfoils are airfoil shapes for wings developed by the National Advisory Committee for Aeronautics, with the four digits describing the shape of the airfoil: the first is the maximum camber as a percentage of the chord; the second describes the distance of maximum camber from the airfoil leading edge in tens of per cents of the chord; and the final two digits describe the thickness of the airfoil as a percentage of the chord. So the NACA4351 airfoil has a maximum camber of 4 per cent located 30 per cent (0.3chords) from the leading edge, with a maximum thickness of 51 per cent.
Now suppose you wanted to know how that airfoil would perform at a 15-degree angle of attack. Typing 'NACA 4351 15 deg' into Wolfram-Alpha would offer data for incompressible potential flow (lift coefficient, leading edge pitching moment coefficient and estimated critical mach number) and airfoil properties in fractions of chord length - including the centre of pressure, maximum camber, maximum camber location from the leading edge and maximum thickness. It would also provide a diagram of incompressible potential flow (with the option to show or hide the velocity field) and a graph of pressure coefficient for incompressible potential flow.
Crucially, it would provide this information far more quickly than it could be sourced elsewhere - unless you happened to have specialist software available.
In another engineering-related example, suppose you were working on a product design and you needed to know the maximum spring force.
Although you could calculate this or glean it from the manufacturer's catalogue or spring calculation software, a simple query to Wolfram-Alpha with suitable parameters and input values delivers the answer without having to look beyond a single web-based interface.
If you were looking for a suitable material for a product design, Wolfram-Alpha can give not only the composition of different alloys or the formulations of different plastics, but also information about how they will perform in real-world conditions.
Saving time
In short, whatever scientific question you care to ask - and as long as you are prepared to take the time and effort to phrase it carefully - Wolfram-Alpha will search its databases and compute the answer for you. That makes it of immense value across the whole spectrum of scientific and engineering disciplines, potentially saving a lot of time in some of the most laborious aspects of design.
Wolfram-Alpha therefore represents a substantial technical and intellectual achievement, and the computational knowledge engine available online at www.wolframalpha.com certainly has impressive capabilities. But this is only the beginning of the journey. Of course there is the need to keep adding information to its databases, thereby boosting its capabilities as an online tool.
However, beyond that, the developers argue that the combination of Mathematica and NKS points to far-reaching opportunities for Wolfram-Alpha in the future, from the development of a radically new kind of programming to the systematic automation of invention and discovery.
Some pundits even suggest that Wolfram-Alpha may represent the model for a new approach to creating an 'intelligent machine' that is simpler than the traditional artificial intelligence approach and less complex than the open-standards Semantic Web approach.
Whatever the future might hold for it, Wolfram-Alpha's value as an engineering tool can only grow.
Within that context, the system's current linguistic failings can be forgiven, as can its ultimate knowledge limitations, and it deserves to be recognised for what it is: a contextual, complementary tool of undoubted value in the engineering design environment and beyond.
