08 November 2006
Petr van Blokland
Petr van Blokland

CorporateIdMaker is web based server program that simulates the design process for corporate identities. It takes information from a customer, selects what it needs from a set of predefined elements and ready-made formats from a library (its design knowledge and experience) and creates solutions that will fit the customers' needs. The output can be downloaded from the site.

While this process is an iterative task (two steps ahead and one step back), the program works through a decision tree. The result of every cycle is presented to the customer as visual options to select. With this choice as input, combined with all the earlier decisions along the line, the program narrows its scope of possible options down to a design that will work. The process described above will lead to a set of corporate basic elements, e.g. a logo (or a whole family of related designs), a choice for typefaces, a colour scheme, grid, formats for layout.

Eventually the customer decides that the presented style fits his need. He then can go through a list of applications like stationary (letterhead, business card, envelopes, compliment card), models for forms, covers and internal pages for brochures, templates for folders, a core website and a corporate manual. If the customer selects one or more of these items, they go into his shopping-cart. The total amount will be charged from his credit card.
Subsequently, all documents can be downloaded in GIF, PNG, PDF or EPS file format.

Design as craft
Once all clothes were tailor-made. And all furniture was crafted by carpenters. Paintings were only commissioned by those who could afford it. Their materials and tools where carefully collected. All measurements were taken from the customers' body or the available room in his house. Choosing the right colour and materials was done by the experienced eyes of the craftsman. Determining the required style, strength, structure, smell, direction of the grain, the position of buttons, all had important influence on the final result. The needs of the customer, and more importantly, the depths of his pockets, formed the basic conditions from which the product was made.

Then came the industrial revolution, and an obvious division came into being: there are things that can be produced by automated processes and other things that cannot. The mass production of bottles, cars, clothes and food has made them commodities. But it is a general believe that products as legal advice, literature, art, bridges and corporate identities never can be structured to the extent that their creation can be done solely by computers. They all seem too divers to be encapsulated in descriptions and algorithms.
But development goes on. Nowadays the only time people talk to their tailors is during the preparation for graduation or wedding. The normal clothing comes ready-to-wear from fashion shops. Customers can choose from a fixed set of dessins, styles and sizes. Although the automation of production created a shear infinite amount of designs, the choice is still distinct. A selection from integers, with no in-betweens. After filling in some parameters on colour, model, the number of doors and the size of the spoilers, a customer can expect the ordered car appear in front of his house. The car was still tailor made, but the parametrized production process guided the creation of this apparent unique design.

Still, designers are essential. Only they can create a process like this, since design process needs to be carefully adjusted in order to separate creative component from the repeating production tasks. Every possible design solution, that in a more traditional process, would have been filled in along the way, must be prepared in advance. All possible values of these parameters must have a valid result during the production. At first sight this looks impossible, since the amount of possible option to predict grows to near infinity pretty soon.

But that does not need to be. In a well-defined system the available option may seem to grow massive, they still can be generated by a limited amount of parameters and options. If the parameters are chosen in such a way that they are not interdependent or overlapping, the can be selected separately without generating conflict at the production level. Just as the colour of the chairs in an ordered car does not influence the number of doors or amount of horse-powers of the engine. Only designers can decide whether parameter are likely to conflict or not, since sometimes they do as in the colour of the car and the colour of the chairs.

The design process
We need to structure the design process in such a way that it can be written as a set of rules and algorithms. Although this process normally is quite intuitive, most programming languages need a formal description of all possible options. Our automated design process has to working along the a tree of discreet decisions, always resulting in a simplified version of the process a real designer would go through. The aim of the system is to acquire enough resolution that the difference will not be noticed in most cases.

Every design process can be seen as a tree of micro decisions. The designer, or in our case the simulation program, walks through the tree from top to bottom. Just as with the limitations of the chess game, in theory all possible solution can be found, selected or rejected. But the exponential growth of choice with the increase of levels, makes it practically impossible to find all possibilities.

A chess player and a designer will intuitively reject whole branches in one time, if their value does not seems worthwhile from an overall perspective. Even if there would be a golden solution at the very bottom of that branch. The amount of available time, combined with the intuition and experience to recognise comparable situations of the designer/players, as well as a relative very little amount of luck, is the cocktail that will make the set of choice to walk through the option tree.

A design process always starts with the linkage of a problem and the need and facilities to solve it. The intermediate procedure has to be a design process, if there is no obvious method to go from top to bottom. Note that the scheme above is applicable for all problem sizes. It handles both questions like 'I need to cut this bread, will a knife do the work?' to large scale issues 'How do we solve the Ozon-layer problem?' In all cases the design process can be described a problem solving black box between the problem and the solution.

But in practice the design process is not a black box, although the outside world often thinks it is. In fact, the design process loops through the same procedures over and over again. Every time the output of a design cycle is taken as input into the next cycle. The disadvantage of a cycling process is that it will run infinitely if there is no defined condition where it will stop. If the design is 'good enough' (whatever that means) the cycle will be broken and the end result is presented.

Additionally, we have to recognise running out of 'time' is an ending condition too. If the result will never be good enough as in 'Solve the world energy problem in two days', this condition will stop the loop as well, without running into an adequate solution.

In a more generic approach of the previous scheme, we could also recognise that taking the input form one design cycle as input into the next, will slightly redefine the solution (as well as the problem). In fact the design process appears to span the whole structure, not just a black box in the middle.

The Tools
There are no standard tools available that are modelling the design process. Many application simply act as an aid to the designer, and in most cases they don't perform that role very well. Page layout programs and drawing programs merely computerise the traditional pencil and paint from the physical world. Not mentioning that they don't help the design process itself, nor do they automate the repeating tasks from the designer, unless he thorougly specifies what to do. The time where a program does act like a design companion to the design is still a long way off.

The simulation of design processes is not likely to be found in standard applications. Thus, a tool that will automatically generate a complete corporate identity, only using some answers and rough choices as input from the customer, needs careful preparation. If we don't, we will end up with just another drawing tools, where every update can be recognised from the new filters that pop up in design made with them.

Designs still need to be made. Sketches are the fluent of the a design process. The difference with CorporateIdMaker is that is now works the other way around: instead of drawing for a particular customer to solve a specific problem, every sketch may have a potential application for an unknown customer. Instead of selecting potential solutions from a set of sketches, the designer has to classify every option for utilisation in multiple brands and applications.

Trees, recursion and iteration
Tree-walking is a highly recursive process. This means that all processes can be described as a set of similar processes, but with a different set of parameters and in most cases on a different scale. This exactly matches our earlier definitions about the design process, where the choice from a known set of options leads to a new (more dedicated) tree with new options. This would run into infinitely if there wasn't a range of practical limitations to end the process.
In pseudo code the process described above would read:

aDesignProcess (setOfOptions):
for every option in the setOfOptions:
if this option is good enough
or we ran out of time
take this options as final result
stop the process

if this option looks promising enough
and we did not yet review this solution:

generate a new set of options from this option
start a new design process (newSetOfOptions)

The structure of a design process as pseudo code. This kind of syntax is often used to describe algorithms, without connecting it to a specific programming language. Some basic rules are needed to understand the working. The "setOfOptions", between the brackets, means that this set is taken into the design process as a parameter. Since the process is calling itself over and over again, this is referred to as a "recursive" process. The structure is simple, but in practice it can solve complex problems due to the enormous amount of levels and options that fit in the design tree.

The CorporateIdMaker Server
The program that is running somewhere on an Internet connected computer is called the "server". A request for an internet page from a browser will make the computer produce the page. Pages are not stored as static files of information, but created on-the-fly, as they are needed. This way the pages can and will have a different content for every customer, depending on the history of answered questions and choices along the design tree.

About this article
This article was first published in Dot dot dot 4 and is reprinted with permission.

About the Author

After VWO/Atheneum-school (pre-university education, grammarschool), Petr van Blokland (Gouda, February 29, 1956) studied at the Graphic and Typographic Design Department of the Royal Academy of Visual Arts in The Hague. In 1979 he graduated cum laude (with distiction) and worked as an intern at Total Design in Amsterdam and Studio Dumbar in The Hague. Since 1980 he is working as an independent designer. To specialize himself he studied for serveral years at the TU-Delft department Industrial Design. From 1984 untill 1989 he teached at the Academy for Visual Arts in Arnhem. Since 1988 he is a teacher at the Graphic and Typographic Design Department of the Royal Academy of Visual Arts in The Hague.

About Dot dot dot

Dot dot dot, graphic design / visual culture magazine
ISSN 1615-1968
two issues a year
Published and edited by Stuart Bailey, Peter Bilak

Dot dot dot is an independent, after-hours, graphic design magazine intended to fill a gap in current arts publishing. We are not interested in re-promoting established material or creating another 'portfolio' magazine. Instead, we offer inventive critical journalism on a variety of topics related both directly and indirectly to graphic design.