With Space In Mind

Estuary at Maroochydore, Qld.

About four years ago, I hacked together a web site so that my father could promote his art online. His operation is now beginning to get more serious, as galleries exhibit his work, he gets media attention, and other people start to manage various parts of what is becoming a business.

Hand-hacked HTML and Python scripts no longer work in such an environment, so I’ve had to migrate his site over to WordPress to permit third parties to manage its content. I’m astounded at how far WordPress, and the infrastructure it’s built on, has come in the last few years. Back in 2003, I wrote my own CMS for Supperware using acres of php that provided minimal functionality. Now I’m using textpattern here just because it’s easy, I don’t blog much anymore, and I can’t be bothered with anything else.

The point is this: please take a look. I’m not going to vouch for his website, but David Supper’s pretty good.

Academic stuff

I realised I didn’t have my academic papers up on the web site anymore. Here are the more significant ones.

Supper, B, 2010. Processing and improving a head-related impulse response database for auralization. Paper 8267. AES 129th Convention.

Supper, B, 2009. Characterising studio monitor loudspeakers for auralization. Paper 7994. AES 128th Convention.

Supper, B, 2005. An onset-guided spatial analyser for binaural audio. PhD thesis, University of Surrey.

Supper, B, 2000. Virtual reality presentation of loudspeaker stereo recordings. Undergraduate technical project, University of Surrey.

VR audio engine running on RISC OS in 2000
VR audio engine running on RISC OS in 2000

Book Depository: The Goal by Eliyahu Goldratt

In an appendix, Goldratt explains how the Toyota Production System [TPS] was created especially for Toyota, transforming the business into an international phenomenon. The system’s not for everyone, though. If you manufacture something other than cars, in a different country, with a different pattern of market demand, TPS will probably cause as much trouble as it solves. To run a factory successfully it may be unhelpful to import Toyota’s system wholesale. Copy Toyota’s methodology, not their methods. Think like them. Apply the same reasoning through which Taiichi Ohno conceived and created TPS over two decades.

The Goal is a didactic novel. It was 1984 when it was first published, and the book portrays a man’s world. People smoke in meetings, and nonchalantly down two or three beers to take the edge off a hard day before hitting the road. It jars these days, as it would if the protagonist casually took a detour from the journey home to burn down a convent.

In the land of 1984, it also takes hours to track somebody down when they’re away on business and you may never have a true picture of what they do for a living. Such details are merely curious today, but other truths about human nature and the physical world are more fundamental and pervasive.

Conclusions arrive in a series of lessons and revelations to Rogo, the production manager of a struggling factory in a Midwestern town. From the start, Rogo’s plant is in a state of permanent crisis, with stacks of late orders, unpredictable output, perpetual panic on the factory floor and union disputes, and his whole division is the same. His boss, Mr. Peach, is about to close the factory unless dramatic improvements occur within three months. Meanwhile, at home, Rogo’s marriage begins to break down as the pressures of his career bite into his personal life.

Corporate politics and the powerful grip of established practices hold Rogo back until a chance encounter at an airport with a former physics professor called Jonah Author-Surrogate begins to set him, his colleagues, and his factory on the path to redemption.

The lessons of this book are initially specific, dispelling fallacies common in manufacturing in the 1980s. First, an excess of unsold goods is bad because it impacts on cash-flow, representing money spent without income. It needs to be stored and eventually becomes obsolete, decreasing in value the longer it’s kept. The same is true of unfinished materials, because they cannot be sold but still incur a maintenance and storage cost, and are therefore a liability. The best way to manufacture is thus strictly to order, to satisfy immediate demand.

Limiting inventory and finished goods determines how a factory must work. The factory is now a system, and must be optimised as a whole rather than a series of stages. Optimising the efficiency of individual processes is not useful unless it improves the predictability and control of flow of finished goods. In fact, it is important to have some idle workers and machinery, as they can cope with statistical variations in other parts of the system. If people and machines must sometimes stand idle for the benefit of the whole, average labour per part is not an important measurement of the quality of factory management: only overall operational cost.

The process with the lowest capacity is a bottleneck, dictating the factory’s maximum throughput. Bottleneck processes can often be found by finding the places on the floor where work in progress accumulates. These can be pernicious, and it is sometimes economical to outsource particular stages of manufacture, or use more expensive techniques, to keep the rest of the factory productive and to prevent a pile-up of work in progress.

Upstream of the bottlenecks, considerable excess capacity may be required in order to allow a constant amount of work to be queued at the bottlenecks, and to allow this to be replenished quickly if it is ever depleted.

As Rogo applies these processes, his factory improves dramatically. Not every step is an improvement, though: some hit the diplomatic buffers or just aren’t appropriate. They are tested and rejected; they sometimes work initially and then become counterproductive later on. More dangerously, some changes work well but the company’s measurement techniques are wrong, and don’t show the actual improvement.

Rogo’s team starts to use computers to predict throughput so that lead times can be determined whatever the state of the current order book and work in progress. Then comes the revelation that marketing and sales are part of the factory system: demand is a component of the production process, and can be used to optimise the factory. In many cases, small batch sizes might often prove more economical than large ones in spite of the extra outlay in materials and machine setup times because they speed up the flow of goods between processes. This results in shorter lead times. Short lead times make the company more competitive and can be used to generate additional sales, whilst further reducing the necessary inventory of finished goods. This explains why Japanese factories take considerable pains to reduce their machine setup times.

The deductions that lead a factory to greater productivity unfold across more than two hundred pages, while the narrative keeps these pages turning quickly. Goldratt provides his protagonist with plenty of context outside the factory floor: he learns as he walks with the Scouts, plays invented games, and discusses abstracted versions of his factory problems with his colleagues and his young children. Certain points are laboured with a heavy hand, but the author hardly sets out to emulate Dickens.

The Goal concludes as Rogo is trying to establish a science of management: a set of general rules that will enable him and his team to tackle any future problem. Do new ideas proceed from inspiration alone, or from data? How much of these ideas are formulated by teams, and how much by individuals? What does philosophy say about how new theories and methods develop, and does this even help? As Rogo learns, he starts to apply these techniques successfully to his personal life, and rescues his marriage. The novel ends with Rogo’s promotion to divisional controller, widening his responsibilities and setting up Goldratt for a sequel.