WITH Microsoft’s venerable operating system, Windows XP, about to be cut loose by its creators, Babbage has been upgrading all his geriatric “wintel” computers to Windows 7. For those who have not yet done so, a word of warning: after April 8th, there will be no more security patches, bug fixes and free (or even paid) online assistance for Windows XP, as Microsoft ends its extended support for the ageing software. From then on, anyone who continues to use it will be at the mercy of hackers, who will doubtless find yet more devious ways of exploiting vulnerabilities in the 13-year-old operating system and applications that run on it.
In making the move, Babbage has avoided Microsoft’s latest operating system, Windows 8. It does things a bit too differently for his liking. Besides, having flopped badly, Windows 8 (now in version 8.1) is not expected to be with us much longer. Microsoft is now rushing out Windows 9. In due course, Babbage will leapfrog to that instead. Meanwhile, Windows 7 will do just nicely. Official support for it will last till at least 2020.
By and large, the upgrade process has gone without a hitch, even though migrating from Windows XP to Windows 7 is not as simple as upgrading direct from Windows Vista (XP’s loathed and short-lived successor) to Windows 7. Fortunately, there is an excellent tool on the market called PCmover that makes the migration a breeze (see “Say farewell to XP”, September 6th 2013).
Even so, it has not been simply a matter of swapping one piece of software for another. With ages ranging from two to a dozen years, some of the components in his fleet of workhorse computers have had to be replaced. Old computers tend to be limited by their skimpy amount of memory. A decade ago, a gigabyte of random-access memory (RAM) was considered adequate for most office chores. Few machines had slots for more than two gigabytes. But given the bloated size of today’s software, four gigabytes of RAM has become almost essential, and eight gigabytes preferable.
That has meant swapping out the motherboards of several legacy machines, and buying more memory as well as the odd multi-core processor. In addition, old power supplies have had to be replaced with beefier ones. While a computer’s video, audio and networking were once handled by separate circuit boards, such services are now built into the motherboard itself. Despite the integration, modern motherboards—thanks in large part to their extra memory and more powerful processors—use a good deal more juice.
At this point, readers may well ask why bother? The time and money spent bringing legacy machines up to scratch is hardly worth the effort. Thanks to Moore’s Law, computers—especially desktop machines—are so cheap and powerful nowadays that the obvious answer is to junk anything over five years old, buy modern replacements complete with up-to-date operating systems, and have done with it.
All true. But that misses one essential point: there is huge satisfaction to be had from salvaging superannuated gear and putting it back into productive use. That is particularly the case now the average computer offers way more power than most users generally need. Besides, there are interesting lessons to be learned concerning the price consumers pay for progress.
In this case, Babbage has been struck by the way the increased “value” of the latest computers—in terms of their better (ie, lower) price-performance ratio—has too often been paid for by a loss of quality. He doubts very much whether it will be possible to upgrade today’s computers in the way that older ones could so easily be made new again.
Out of curiosity, Babbage used a digital micrometer to compare the thickness of the metal chassis of his 12-year-old desktop with the innards of a two-year-old. The former employed 19-gauge (0.0359-inch) aluminium, while the latter was made of 35-gauge (0.0075-inch) mild steel. Taking the different stiffnesses of the two metals into account, the older case was twice as rugged as the newer one. The latter could be easily bent by inadvertently applying finger pressure in the wrong places. By contrast, Babbage wore out several Dremel cutting and grinding wheels modifying the 12-year-old old case to accommodate a modern motherboard.
To be fair, the older computer set Babbage back over $2,000 in today’s money, while the newer one cost around $650. Meanwhile, the later machine had seven times the processing power per dollar of the earlier one (before the upgrade). On price and performance, then, today’s computers offer tremendous value.
But customer satisfaction comes not merely from getting the best bang for the buck. Quality counts, too. However, while price and performance can be measured directly, quality is a far more illusive attribute. In a product, it can include durability, reliability, even serviceability. Does the product exceed the specifications expected? Are there more features than usual? Is it thoughtfully designed and pleasing in appearance? Apple understands this better than anyone else, and its customers are happy to pay a premium for products that frequently offer little more than average performance.
Paying an Apple premium of 35% or so is one thing. But few would pay 200% more (ie, $2,000 versus $650) for a computer just because it was built like a Sherman tank. That he accepts. But what bothers Babbage is whether—in the trade-off between quality-control and value-engineering—manufacturers have sacrificed too much in order to hit the price points they believe maximise returns. In short, would they be better off making more Mercedes machines and fewer Toyota ones?
Mention of the motor industry is apposite. Where the computer manufacturers have had the benefit of Moore’s Law to help them drive down the price-performance ratio of their products, carmakers have had to rely on value-engineering almost exclusively. And while today’s motor cars offer unprecedented value in terms of price, performance, safety and fuel economy, they have become more complicated than ever, more difficult and expensive to repair, and their economic lifespan severely shortened.
Value-engineering has much to answer for. The term was coined by General Electric during the second world war, to describe a way of finding substitutes for materials then in short supply. It turned out that, in some cases, the substitutes reduced manufacturing costs and actually made the products better. In theory, value-engineering seeks the optimal balance between lowered costs, increased productivity and improved quality. In practice, the quality aspect often gets short shrift. In such instances, value-engineering becomes little more than cost-cutting.
The guiding principle of value-engineering is an analysis of function—ie, what something does, not what it is. Once the cost drivers are identified, a product is redesigned so as to have a lower component count and to be made of cheaper material. A widget that was initially machined from a block of metal may finish up being injection-moulded from plastic. To see value-engineering in action, look under the bonnet (hood) of a modern car. The majority of unstressed parts are injection mouldings. Little more than a decade ago they were made of steel, brass or aluminium. Guess which last the longer, which are easier to repair?
Babbage has long thought the make-believe “one-hoss shay” memorialised by Oliver Wendell Holmes to be the most sublime piece of engineering ever conceived. The fictional deacon in the humorous poem crafted a two-wheeled chaise in such a logical fashion that, with each part made as strong as the other, it could never break down. The buggy worked flawlessly for a 100 years. Then, finally, all the parts wore out simultaneously. “[I]t went to pieces all at once, / All at once, and nothing first, / Just as bubbles do when they burst.” How Babbage wishes cars—and computers, too—were built as logically and as well.
Powered by Facebook Comments