Looking askance 5
Looking askance 5
February 2005
I apologise for the long delay since my last ramble among the swimlane lilies and behavioural nettles of Business Process Management. My time was eaten up in finishing the book on BPM that I had been writing. It is now with the publisher, for its editors to do their damnedest to its fine prose and intricate structure. Publication should be in the summer. (Yes, that is a long time. The publisher's own business processes do not bear thinking about.)
Despite the gap, Mark McGregor has kindly, if insanely, asked me to continue inflicting on you the mixture of acid drops and twisted logic that goes under the title of Looking Askance, so he's partly to blame for what follows.
In praise of negative feedback
According to James Boswell, Samuel Johnson once observed that, “The greatest part of a writer's time is spent in reading, in order to write: a man will turn over half a library to make one book.” I can vouch for the truth of that, as can any author.
One of the nuggets I discovered in the word quarry I dug was a paper called Negative Feedback: Encouraging the “Right Stuff”, by Chris Matts. In this, he describes “the pressures that Waterfall development puts on people to give positive - and wrong - feedback”.
Sometimes called the systems development life cycle (SDLC) model, the 'waterfall' method is linear, sequential and compartmentalised. It gets its name from the way discrete activities are completed before 'cascading' to the next level down, if square lumps can be said to cascade. The routine is nothing like so fluid as the name suggests. It might better be described as the box, seal, heave, thud and unwrap method (BSHTUM, because the least said about it the better).
One Web site I consulted said this about waterfall development: “It limits the amount of cross-team interaction that occurs during development… and it allows for greater ease in project management since plans aren't constantly being revised.” And these were among its good points.
There are two impulses obvious in methods like this. The first is the urge to neatness and the other is the rush to completeness. A good solution, it seems to say, leaves no loose ends and can be forgotten about once signed off - and the sooner the better.
Chris Matts' paper contrasts this anal-yst's dream with the drawn-out messiness of real-life systems design. He suggests several ways of dealing with untidiness other than pretending it does not exist. Matts lists these under such headings as “face-to-face”, “fuzzy versus hard” and “celebrate failure (of the model) / break the model”.
If you are familiar with General Systems Theory, you will recognise these stratagems as being aimed at variety reduction. By contrast, like many conventional design methods, waterfall development is instead mostly aimed at variety avoidance.
Except for the observations below, I shall go into no more detail about this paper but I do recommend you to read it. You can get it at http://www.pols.co.uk/business-coach/NegativeFeedback.pdf.
What caught my eye in the paper, under the heading “paper/lo-fi prototyping”, is the idea of giving users (that is, fellow designers) deliberately rough-looking prototypes to consider. As Matts says, “The nature of the prototype encourages feedback because it has an unfinished feel to it.”
He mentions a piece of Java software to help with this called Napkin Look & Feel (see http://napkinlaf.sourceforge.net ) This lets you produce mock-ups of user interfaces that look as though they were hand-drawn. Because they are data driven, the designs can be converted to a production-stage format once they have been agreed.
I like this idea and would like to see it extended. It would be wonderful if someone were doing something similar for flowcharting. This would allow the best of both worlds - data-driven and easily revised flowcharts that invite defacing and change by users.
Once at a certain level of complexity, IDEF, UML and similar diagrams inhibit rather than aid understanding and agreement. They unnerve users and make their brains hurt. The intricate nature and heavily coded symbology of these charts get in the way of the patterns and paths that people need to see and understand. They should be distributed in a brown wrapper and only to people with some sort of engineering or scientific background (including programmers and systems analysts).
For the average worker or manager, these complicated diagrams are little more than ornamental ransom notes - “Sign here, Mister, or you'll never see your processes again.” The average worker or manager does not have the time, inclination or training to make good use of them. They are usually pictures of the way analytocrats like you and I might see the world but not the way ordinary folk do.
Instead of Visio and the like, why not instead use solid models (Lego, children's toy blocks, modelling clay, etc.), videos, enactments and role-plays, pieces of string and anything else that comes to mind and hand? Years ago, Hewlett Packard produced a videotape of some production engineers demonstrating how pull-mode rather than push increased throughput rates on a production line. It was a version of Kanban working. They used cardboard boxes with large identifying letters, a roller line and a stopwatch. The voice-over explained what was happening; the results appeared in captions and hand-held signs. It was brilliantly persuasive and stayed in the minds of all who saw it.
At the end of your project, only convention stops you including hand-drawn diagrams, amateur photographs, cartoons, line drawings and similarly informal material in the report's main body. If the report is in electronic format, videos and voice-overs could go there, too. Avoid anything too flash (or Flash).
Detailed flowcharts can go at the end or in appendices, with other material that is tough to digest. They are necessary but not to an overall understanding of what has been going on and what will be.
What's in a name?
Cisco has a “Director of Customer Listening”. I think he is really its sales director. It's very nice that he should be free to portray himself as such; even better that his colleagues should accept it. In many organizations, people would just snigger at such self-labelling. I assume Cisco sees it as expressing an underlying truth about his job.
Truthful job titles would pose problems in some companies I know. There, any evidence that a Director of Customer Listening might pass on would mostly likely go to the Director of Not Wanting To Hear the Messages (formerly known as the product director) and his colleague, the Director of Doing Sod All About It Anyway (typically the engineering director).
Bizarre job titles are fine for a while, but their shock value fades. Some just invite ridicule. At a conference a few years ago, I had to keep a straight face while introducing a speaker who had given himself the title of “Corporate Seagull”. Heaven help the people under him, I thought.
Process modelling at the Prado
In the last Looking Askance, I promised to discuss what Hieronymus Bosch and Diego Velázquez can tell us about systems. These thoughts arose during a visit my wife and I paid to Madrid about 18 months ago. We are both art lovers, so naturally put the Prado museum near the top of our list of places to visit (see http://museoprado.mcu.es/ihome.html).
Among the many wonders and masterworks on display were two I was particularly keen to see. One was Hieronymus Bosch's The Garden of Earthly Delights; the other was Diego Velázquez' Las Meninas (“The Ladies in Waiting”).
Bosch's work is the earlier. He painted it in 1504 or so, probably in his home town of 's-Hertogenbosch, in the Netherlands, after which he named himself. (His given name was something like Jerome of Aachen, after the now German town of his birth.)
Bosch - Garden of Eden
The picture is a triptych, a hinged set of three wooden panels that can close to conceal its contents. These typically went above the altar in a church or private chapel, so were usually on Christian themes, as this one is. Bosch was a religious man, even by the standards of those times, and a member of an influential group called the Confraternity of Our Lady.
The triptych is a detailed and superbly executed visual lecture on what happens if you engage in lustful behaviour. On the left panel, shown here, is the Garden of Eden, as serene, bountiful and wholesome as any you could wish to see depicted. In the large central panel, Bosch shows the earthly realm, with not two but hundreds of people and all of them up to no good, sometimes with beasts. On the right panel, he shows us his vision of the hell that awaits these sinners. Machines appear for the first time, inflicting pain on the fallen in ingeniously horrific ways.
Bosch displays astonishing craftsmanship throughout. He also lets his imagination run amok, producing scenes and creatures of a surreal and hallucinogenic quality centuries before Surrealism and LSD were invented. The effect is astonishing and, metaphorically at least, transfixing.

I spent a long time looking at these panels, usually close-up; Bosch paid plenty of attention to the minutiae. You can get a (muted) idea of the panels' appearance here - http://www.ibiblio.org/wm/paint/auth/bosch/delight .
Unfortunately, the significance of many of the symbols Bosch used and the references he made are lost to us today. It is, so to speak, written in a partly lost code. The work's early history is now unknown, too, along with many details of Bosch's life.
We do know that about 90 years after its completion the triptych came into the possession of King Philip II of Spain. (He was then also ruler of much of the modern Netherlands.) It went into his palace, El Escorial, being acquired for the Prado in 1940.
The second painting, Las Meninas, was painted in 1656 for Philip II's grandson, Philip IV. Velázquez was his court painter from 1623 until he died in 1660, becoming the king's friend and confidant. This was extraordinary at a time when painters were seen as mere artisans. Mainly they were required to crank out formulaic paintings on religious themes. Velázquez overturned this convention, too, becoming a revolutionary in the way he depicted people, especially commoners.
Velázquez - Las Meninas
Las Meninas is an example of this. A large (roughly 10 feet square) painting of the royal household, it reduces the king and queen to smudged reflections in a distant mirror. Closest to you is a large, sleepy dog, normally the lowest member of the hierarchy. Then come two dwarfs, commonly found at that court. Centre stage is the young princess, the Infanta Margarita, flanked by serving maids.
To the child's right is the painter himself. Dressed to the nines, like a plumber in a dinner jacket, he looks back at you quizzically, appraisingly, challengingly, with brush in hand and head cocked to one side.

Could he be you, or are you instead one of the mirrored monarchs? Is the canvas he is working on this one? And why is he so far away from it?
I spent an even longer long time in front of this painting. Guidebook in hand, I went backwards and forwards, looking at this detail or that, usually seeing just random-looking blobs and streaks of colour when I did.
During one retreat I found myself at a point, some distance from the canvas, when these matters didn't matter and the painting disappeared. Instead, specifics were replaced by a sense of being part of the scene, of silently being absorbed into it. The poet, Théophile Gaultier, commenting on the effect, reportedly wondered aloud, “But where is the painting?” How Velázquez carried out this wonderful trick defies full analysis, and many have tried.
As with The Garden of Earthly Delights, I felt privileged to have been allowed to see this painting. Also as with the Bosch work, there is no room here to go into a comprehensive description of it. You can see an image of and commentary on it here - http://www.artchive.com/meninas.htm .
What has all this to do with Business Process Management, you might wonder? Thinking about these two masterpieces afterwards, it came to me how the differences between them symbolise the differences between looking at the mechanistic aspects of business processes and at the human aspects. I have summarised them here:
The Garden of Earthly Delights
Las Meninas
Finely detailed
Impressionistic
Needs close inspection to reveal all its information
Looking closer tells you little or nothing more
Tries to show you everything as you progress through
Tells you everything important in one look
Mostly imaginary
Based on reality
Involves the use of machines
Entirely human
Conforms to and exaggerates existing orthodoxies
Subversive, of social and technical conventions
Normative in intent; sermonising (and negative) in tone
Explanatory in intent; humane and embracing in tone
Explicit as to causes and effects
Contains subtle messages about relationships and power
The observer is detached (as in Newtonian science)
The observer is part of the situation and affects it (as in modern science)
Dependent on a particular, and largely obsolete, symbology
Universally understandable, even today

Finally, in case you are thinking this is a 'knock the techies' diatribe, I ended up by asking myself two questions:
1. Could we manage with just one or the other way of looking at life? Answer: no.
2. Would we be the poorer without one or the other. Answer: indubitably.
Petri versus pi… the search continues
Also in the last Askance, I mentioned the debate cum dispute over which of two methods is better for writing BPM software - Petri nets or pi calculus. Since then, some people have been debating this question and ways to resolve it, if it is resoluble. The result is a planned workshop to discuss the matter more formally (but not too formally). The organizing committee comprises Wil van der Aalst, Rob van Glabbeek, Keith Harrison-Broninski, Robin Milner and me.
Our idea is for speakers to present examples of how Petri nets or pi calculus can inform business processes, whether for analysis or implementation. We are keen to attract adherents of both camp - and people who support neither. Our intent is to make a bridge among BPM users, providers (software makers and systems integrators) and academics, so ideally we would have all three constituencies represented.
The event is planned for 6 to 8 June 2005 (lunchtime to lunchtime), at Eindhoven University of Technology, the Netherlands. For more details, please read the call for participation, which is at http://www.office-futures.com/PiPetri_CFP.htm [link now defunct].
Give it a try
Change happens when people who don't normally have a say talk to people who don't normally listen.
Anonymous source, from the Civic Forum in Banja Luka, Bosnia.

Roger Whitehead