Mad Dog 21/21: Craft Nouveau
September 7, 2010 Hesh Wiener
Somebody at Google knows about Bing. Not Bing the Microsoft googloid. Bing as in Siegfried (Samuel) Bing, the German expatriate whose Paris gallery, L’Art Nouveau La Maison Bing, founded in 1895, gave its name to a worldwide movement in art, craft, and design. A Google logo celebrated Art Nouveau on July 24, the 150th birthday of Czech illustrator and temporary Parisian Alphonse Mucha. Today, one can find that je-ne-sais-quoi of turn of the Twentieth Century Paris in cyberspace. Its message: Computing, to succeed, must, like the Eiffel Tower, marry art and science.
The Eiffel Tower, opened in 1889, is an inspiration to every computer scientist who believes in doing the most with the least. It is made of iron, not steel. But the engineering of the thousand-foot tower uses very little metal. According to the Wikipedia piece, if the entire tower were melted into a block covering its base, the block would be less than two and a half inches thick. The original elevator mechanism was based on huge water-powered hydraulic cylinders.
A decade after the Eiffel Tower was completed, Paris opened its Metro system; travelers using the new transportation technology were treated to graceful Art Nouveau station entrances.
During that era, metallurgy and electric grids were changing the world with irresistible force, much the way networking and information technologies are redefining societies today. Art Nouveau reminded people living in the emerging steel framed environment to honor nature and keep mankind’s accomplishments in perspective. In some places, such as Riga, Latvia, the city center was rebuilt in honor of the Art Nouveau aesthetic.
Engineers who undoubtedly appreciate the engineering of the Eiffel Tower pack the capacity of a recent mainframe into a laptop, fold elaborate software into iPads and Kindles, and jam a stunning amount of digital magic into iPhones and their ilk. And, when the craftsfolk are at their best, they provide an interface that celebrates the natural world even as it opens doors to an artificial one, in the manner of Guimard’s Metro entrances.
The best programmers try to build large and strong but light structures into their software. At the computing culture’s best, there is a never-ending race to make operating systems, Web browsers, and application programs more friendly, using finesse rather than brute force to minimize the size and weight of software packages. The most skilled coders have learned to craft document pages for the Web that use the best available display technologies to keep a lid on bandwidth requirements. The results not only reflect skill and a love of elegance, they also have a practical payoff as they reduce the computing and electrical power needed to serve, transmit, and render messages.
Computing technologies are now past the stage that corresponds to that of the French centennial celebration of 1889, when iron gave way to steel all around the world. A dramatic example of the change in structural architecture that created skyscrapers can be found in Chicago, at John Root’s Monadnock Building. The north half the building has curves that are a tip of the designer’s hat to Art Nouveau. It is one of the last big buildings in the USA that was held up by its walls (which are something like six feet thick at ground level). The south half used a steel frame and it is much lighter. It didn’t sink into the ground the way the north half did; it didn’t need entrances dug down to where the north half finally settled.
Of course, computing encompasses much more than technologies a person can easily carry; servers and cloud systems can also be beautiful when they are well thought out. Server technology that more or less corresponds to cast iron architecture emerged during the late 1960s in the form of Unix. Since then, Unix has evolved and inspired variations, such as Linux, and functional imitators, such as Windows Server. IBM lost its way, or at least its mainframe group did, and stuck with descendants of batch mode operating systems that don’t offer much when it comes to reaching out of the glass house with charm or beauty. But IBM has another, somewhat neglected, software environment that may be the opposite, software that was created to bring more natural services to the user’s end of the wire.
The proprietary IBM technology with the most aesthetic potential was born as the System/38, grew into the AS/400 and continued to evolve through its hardware i5/OS stage and into the IBM i implementation on Power. It represents an extensible and hopeful concept of computing that stands in stark contrast to the closed and mordantly practical IBM mainframe.
The unusual concepts that defined the S/38 and its offspring are so attractive that IBM has never been able to abandon them, although users probably have recurring nightmares in which Big Blue ditches their platform. On the other hand, the i ideas are hard to implement in an affordable system, or at least that has been the case right from the beginning, when the System/38 was announced with prices that made other midrange customers hug their relatively economical System/34s or flee to non-IBM minicomputers.
While it often looks like stodgy old IBM is foolishly chasing the children of Unix like an aging arthritic hound going after a fox, a group of zealots in the company keep trying to complete the ambitious project begun by Frank Soltis and his colleagues long, long ago.
The basic idea behind the System/38 was to build a machine that really does what users want in situations where what users want is old-fashioned basic information processing. The starting point is that the users have data and they want it turned into answers. The original end users envisioned by the S/38’s developers were not tuned into GUIs and wireless networks. They were green screen drudges, except that they wanted their bean counting appliance to be friendly, easy to use, and natural enough to make asking a question different ways an easy thing to do. The operating environment that become IBM i is not Unix and never well be. It’s pretty clear that IBM i is not going to support Facebook or serve up chopsocky videos. No, to somebody who just wants to understand a business with a zillion factoids, the System/38 concept is a cathedral and Linux is the bizarre.
In Barcelona, which a map says is in Spain and Spain says is in Catalonia, there is an Art Nouveau cathedral called La Sagrada Família, the Sacred Family, designed by Antoni Gaudi. It has been under construction for more than a hundred years. The cathedral might be done by 2026, the hundredth anniversary of the Catalan architect’s death, and then again it might not. It is impossible to describe this building in a few sentences or even with a few photographs because each of the building’s sides is unique. The interior is almost beyond description, too, with columns that branch out like a tree, supporting the roof on thin fingers of stone and concrete. There’s nothing quite like it anywhere else, except in other Gaudi buildings. It is as unusual as the Art Nouveau center of old Riga.
It there is a place to pray for IBM i developers to be given resources and freedom, Gaudi’s cathedral is it. The software needs more than top-notch coders. It needs developers who can create GUI technology that expresses a welcome to what is left of the original S/38 dream, before the whole mission is thrashed into a bunch of historically interesting APIs talking through their iron masks at ugly old AIX. You don’t have to be an iPhone app developer to see that AIX is not OS X.
IBM probably has a choice. If it takes one path it might send its system architects and any marketing executives it might have who are not made of wood to look at the aesthetic that Bing brought to Paris, the European and American ideas that may been inspired by Japanese woodcuts, and wound up in not only in Paris but also in Barcelona, in Chicago, in Prague, in London’s Liberty department store, in the Yellow Book, and in the arts and crafts furniture of upstate New York. If it takes the other it is Carnegie Steel, a great piece of history near the end of its chapter that ought to be building public monuments while it still has the dough.
IBM needs to figure out what the most important thousand things are that its computer ought to know how to do. And then it needs to figure out how to do these things with beauty and style in response to requests that come to its servers from phones smart and not-so-smart, from tablets, from e-books, from thin clients, from thick clients, from cash machines and satnav appliances.
IBM cannot afford to wait until its customers ask for all this stuff, because by that time the customers’ questions will be about how to get the answer fastest and cheapest. The methods IBM’s customers envision will be the ones taught to them by Apple or Google or Amazon, possibly Microsoft (although I have some doubts about Microsoft’s ability to lead), and very possibly by other companies IBM hasn’t heard of yet and wouldn’t understand if it had.