It looked like the UK had finally joined the long list of countries opting to sit on the fence in support of Open Source Software (OSS). So it was a pleasant surprise for the OSS community when Whitehall recently woke up to the benefits of OSS in public programs and civil procurement.

The UK’s powerful buying arm, the Office of Government Commerce (OGC), pronounced in a well-researched and clearly articulated study, that OSS is a “viable and credible alternative to proprietary software for infrastructure implementations, and for meeting the requirements of the majority of desktop users”. Reporting to the Secretary of the Treasury, the OGC helps guide and improve the efficiency and effectiveness of central civil government procurement in the UK.

The OGC study represents a milestone in the official recognition of OSS, not just in the UK, but on the larger European and global arena, because it examines recent public sector experiences across different countries. Carefully argued and with detailed supportive data, the report explores issues of potential interest to all countries faced with the costs, complexities and concerns of proprietary vs OSS solutions.

The OGC’s clear recommendation is that many pennies can be saved by those buyers wise enough to incorporate OSS solutions wherever appropriate. Dramatic savings can result from up-front elimination of license fees and long-term reductions in software and hardware refresh cycles. The report also recommends that public programs pay careful attention to, and evaluate the full life-cycle costs of continuing to use proprietary products, especially in terms of vendor lock-in, mandatory upgrades and exaggerated service and support requirements.

While concluding that many benefits of OSS are measurable and supported through empirical evidence, the OGC acknowledged that short time scales and the limited availability of certain financial data served to hamper full life-cycle cost comparisons between OSS and proprietary software. Fortunately, additional sources are available to illuminate at least some angles of the financial equation.

A few days after publication of the OGC report, another body of the UK government, the rich and powerful National Health Service (NHS), provided just such an opportunity to evaluate the cost of continuing to use proprietary software products. At the beginning of November, the NHS signed a £500M (US $920M) deal with Microsoft for desktop software.

“This is an exceptionally good deal for the taxpayer that genuinely reflects the buying power of the NHS,” said UK Health Minister, John Hutton. Indeed, if Mr. Hutton is taken literally, the deal reflects the shocking power that such a government body acquires to waste almost a billion taxpayer dollars. Are the desktop computing facilities of the NHS in such disarray that an almost unfathomable sum must be spent for its rescue and continued survival? The entire budget could have been saved by simply doing nothing and using the facilities which exist today - after all software, especially office automation software, is not perishable despite the insistence of vendors to the contrary. Furthermore, if it had been reported that Mr. Hutton wished to spend a billion dollars to convert the NHS to Open Source, the British taxpayer, and indeed the world, would have been justifiably outraged even if spending any amount less than a billion dollars to convert the agency would have been, prima facie, a net gain.

The costs needed to preserve the superstructure of proprietary software from a single vendor are painfully clear in examples such as the NHS’ contract. It is a classic example of an industry aligning with its customers to perpetuate itself and results in an apparent sanctity of IT budgets, especially in developed economies with discretionary wealth.

While the OGC carefully analyzes the potential for savings using OSS, its cousin organization, the NHS, spends profligately on its renewal of proprietary software licenses. If the public demanded a full accounting of the cost of NHS’s embrace of proprietary software, then its chief apologist, Mr. Hutton, might be looking for a new job. However, for some reason reflecting the infinite indulgence of the British taxpayer, Mr. Hutton’s continued employment seems assured. Nonetheless, he should be given some time off to read the advice of his colleagues over in the office of the Secretary of the Treasury.

Dr. Niklaus Wirth is best known in the world of computing as the creator of the Pascal family of programming languages. He has a long list of honors attached to his name - from the prestigious ACM Turing Award to the Leonardo da Vinci Medal - for contributions in fundamental research and design of algorithms, data structures, compiler design and software programming languages. Wirth is one of the pioneers of computer science who shares the stage of history with Donald Knuth, Richard Stallman and Bjarne Stroustrup.

In a recent lecture on programming languages at the Computer History Museum in Mountain View, California, Dr. Wirth talked about simple design and discipline being key to good software programming. He emphasized how good programming languages and software should respect efficiency and optimization even if, as it is especially true today, hardware resources are abundant. Even too much documentation can be a warning sign. Voluminous manuals may be an early symptom of a programming language’s failure. The essence of programming languages, according to Wirth, is to provide a level of abstraction from the details of computer hardware; to provide a suitable model of computation and to allow mathematical reasoning without recourse to knowledge about the underlying mechanisms.

As he walked through the history of programming languages, he described the progression of his work at Stanford University on Algol-W and, later at ETH Zurich, on Pascal, Modula-2 (modular programming), and Oberon (object oriented programming). He proudly described himself as part of the original group of five professors who created the Department of Computer Science at Stanford. In 1968, when changes in the US political climate during the Vietnam War increased anti-war sentiment on campus, he headed back to Switzerland where, as a Professor of Informatics, he continued his seminal work on Pascal, Modula-2, Lola and open source Oberon. Oberon is a single-user, multi-tasking operating system and language that is available under a BSD-style license and emphasizes openness and extensibility.

Wirth talked about basic concepts in computing, saying that not too many people today think about programming the way he did 30 years ago. Optimizing usage of hardware resources, improving algorithms and thinking about fundamental concepts are all part of creating simple designs. Languages have become complicated, unstructured and non-transparent. Most people today use libraries of subroutines to pick and choose the functionality they require, much like hardware designers choose circuits from catalogs and then glue them together. This results in bloated software and hardware. These ideas are similar to what Dr. Knuth, Wirth’s contemporary, also emphasizes. Simplicity and elegance in design are the essence of the best programming languages. Dr. Knuth, the father of Algorithmics, attended the lecture as part of the audience. He asked Wirth about recursion and why it was added as an after-thought to Pascal. Interestingly, Wirth responded that at the time Pascal was created, he was working on a CDC6000 system, a machine that did not handle recursion well. But as the hardware improved, Wirth was able to add recursion as a feature without sacrificing efficiency.

In the 80s, AT&T adopted and promoted C/C++ as the emerging programming standard and Pascal lost the star status it had once commanded, despite being a well-designed, efficient, type-safe and popular language. Software vendors such as Apple spent enormous resources to convert their code bases from Pascal to C/C++ as they rode the wave of C/C++ dominance in the US market. So when Dr. Wirth was asked about the rise of complex and non-type-safe languages such as C/C++, he replied that programmers may be starting to reinvent more disciplined, safer languages such as Java and C#. But, he was quick to point out that even these best exemplars of modern programming languages have a long way to go before they can “catch up” with Pascal or Oberon. So the evening concluded with the progressive Dr. Wirth recommending that we go back to basics - back to the future.