Welcome to NAGNews & Season's Greetings

Issue 50, 8 December 2005

In this issue:

• Top Story - NAG launches Mark 8 of the NAG C Library
• NAG in the News - Thinking clearly about the component approach to development
• People of NAG News - Peter Kemp - John Nelder
• Tips & Hints - Using the NAG add-in functions from within VBA
• Events - London & Birmingham
• Product News - New Implementations

NAG is delighted to announce the launch of Mark 8 of the globally renowned NAG C Library. Extensive new functionality added to the NAG C Library at the latest release provides programmers with over 1000 easily embedded numerical functions. The new developments at Mark 8 make it far and away the largest commercially available numerical library for the C language. In addition to the considerable new functionality, Mark 8 will be available for 64-bit and 32-bit applications alike and feature compatibility with the latest compilers.

With the increased strength in terms of size and functionality, the NAG C Library, Mark 8 will benefit anyone needing to add mathematical and statistical functionality to their applications. Of particular interest to those working within the finance industry will be the extended optimization chapter, which now incorporates enhanced sparse matrix techniques, critical for solving problems, which involve a very large proportion of zeros in the constraint or objective functions. A typical example is in a portfolio optimization problem where a constraint exists on the percentage of the portfolio invested in stocks within a given country. In such a problem there will be holdings specified for many stocks but most will be zero making the problem large but also predominantly sparse in nature.

If you are interested in the new release please email us at nagnews@nag.com with any questions. Supported customers of the NAG C Library will receive their upgrades when their implementation is available.

Imagine going to a Porsche dealer and asking him to sell you a new Carrera GT (mid-engine V-10, 605 HP, 0-100 kph in 3.9 seconds) but then informing them that you'd like them replace the engine with one you'd built in your home garage. Farfetched? The tools and the parts, if not the know-how, are a few clicks away on the web though few of us would seriously contemplate building a car this way.

From the point of view of specialists in mathematical, statistical, and data mining components, this is an apt analogy for many 'build or buy' decisions surrounding analytical application projects. For many complex applications such as these, the arguments for 'in-house' building are as difficult to justify as their equivalents for building the engine for your Porsche at home. Even if the talent to do so is available internally, that talent might be better applied to identifying the best available components to incorporate into the application. This could ultimately create more value and competitive advantage for the business.

To read the entire article, recently published in Ubiquity, by Rob Meyer CEO of NAG please click here http://www.acm.org/ubiquity/views/v6i44_meyer.html

NAG recognises service of Peter Kemp
NAG has been honoured by the long, loyal and effective service of many volunteers in its community. We wish to take special recognition of Dr Peter Kemp of The University of Stirling and his outstanding record of service to NAG over a 35 year period. Dr Kemp first became associated with NAG through his numerical software research interests and his Computer Centre user support role at the University of Cambridge. While at Cambridge, he worked with us on porting the early Marks of the NAG Library to IBM mainframes, which proved to be a decisive step in the evolution of the Library towards high quality, portable software.

In 1979, Dr Kemp was appointed as a director of NAG Inc., the North American arm of the group, shortly after its founding. He served in that role until his retirement in October 2005. In addition, he has served several terms as a director on the NAG Ltd Council of Management, governing the entire NAG group of companies.

The hallmark of Dr Kemp's service has always been the deep and abiding interest in the progress of the company. He will be remembered for his thorough preparation, incisive questions, collegial spirit and keen wit. He encouraged and enjoyed regular interaction with the entire staff of NAG Inc during his regular visits. In recognition of that special relationship, the staff of NAG Inc held a reception for Dr Kemp at his retirement from the Board of NAG Inc. At the reception Dr. Kemp was presented with mementos of his service, alongside recollections of stories that tested both his memory and sense of humour.

The Council of NAG Ltd, the Board of Directors of NAG Inc and NAG staff wish to thank Dr. Kemp for his tireless service and contributions to the organization over many, many years. Without his efforts, NAG would not be the successful company that it is today.

John Nelder Awarded William Guy Medal in Gold by RSS
We are delighted to report that Professor John Nelder, FRS, of Imperial College, London, has been awarded the Royal Statistical Society's William Guy Medal in Gold for 2005. This prestigious award, the Society's highest and made only once every three or more years, is given in recognition of major contributions to the theory and application of statistics.

John's contributions include the simplex (polytope) minimization algorithm, advances in balanced experimental designs, the concept of Generalized Linear Models, and more recently, their generalization to hierarchical GLMs. He is also the originator of the two widely used and highly respected statistical packages, Genstat and GLIM.

The William Guy Medal in Gold is named after William Guy, a 19th century pioneer of the use of statistics in medicine. Distinguished previous holders of the award include Fisher (1946), Yates (1960), Kendall (1968), Cox (1973) and Box (1993). NAG is proud of its long association with John, which dates back to the 1970s and has seen collaboration in various areas, including GLIM, Genstat and several funded research projects. John was elected to NAG's senior technical advisory body, the Technical Policy Committee, in 1985, and in appreciation of his expert advice on computational statistics over many years, was made a Member Emeritus of the Committee in 2002.

The NAG Statistical Add-Ins have a unique ability to update the computed solutions whenever the data changes. They achieve this goal by returning the solution to statistical problems as an array function. In normal usage a utility macro, NAGExtract, may then be used to re-position the output in other parts of the workbook. This approach is very flexible and appeals to many users. The price paid for this powerful capability is twofold:

• The user needs to know how to enter array functions (Control+Shift+Enter)
• A long chain of dependencies can result in too much unnecessary computation

In this article we describe how, with modest programming skills, these problems may be resolved.

To read the entire article please click here http://www.nag.com/techtips/techtip021.asp.

• DevWeek 2006: 20 - 24 February 2006, London
  NAG is delighted to be exhibiting at the UK's longest established annual event for professional software developers and IT architects, DevWeek. We will be available to discuss how NAG's mathematical and statistical components enhance application functionality.
  
  For further information on the event, visit: DevWeek 2006

• The Education Show: 9 - 11 March 2006, Birmingham
  NAG will also be exhibiting at the UK's leading educational resources event in March. The Show will be addressing key issues in education across all subject areas, with a particular focus on core secondary curriculum areas such as English, Maths and Science.
  
  The Education Show is designed to provide all levels of teachers from all specialisms with a unique opportunity to review the widest selection of resources required to deliver better education, from school through to university.
  
  For further information on the event, visit: The Education Show 2006

NAG is committed to offering new implementations of its broad range of numerical and statistical software, compilers and tools. Since the last edition of NAGNews platform availability has increased for the NAG Parallel Library.

The NAG Parallel Library, Release 3 is now also available for the following platform:

• AMD-64 Linux64 using the pgf77 v.6.0.5 Compiler

Please feel free to forward this newsletter to colleagues or to post it to your company's Intranet sites.

To subscribe, unsubscribe or learn more about managing your subscription, please see: http://www.nag.com/NAGNews/index.asp.