Hamilton--Bits and Bytes vs. Dots and Spots

Gurus of management love to publish books with titles like "Reinventing the Corporation." And while many of this genre provide useful information about breaking down corporate barriers—usually internal ones—they assume that the reader works in a large corporate environment.

Sadly, this is of little help to managers of the average commercial printing company or trade shop, whose employees are often numbered in two digits. Yet, the impact of technological change is just as large in a 25-person prepress firm as it is for a giant corporation such as Applied Graphics Technologies. Therefore, this month's pontification will try to address workflow re-engineering in our little world.

Like any manufacturing business, prepress consists of raw materials and processes that result in an end product. In the good ol' days, the raw materials consisted of type and layout boards, hand-drawn graphics and photographs. The processes included page assembly, scanning, retouching, trapping, masking, stripping, and film and proof outputs.

Today, however, type and layouts are almost always provided digitally and only the images are supplied as physical materials. And the day when all images will be supplied digitally isn't too far off, either. Of course, this transformation has eliminated many profit centers, and the content creator continues to encroach upon the prepress technician's territory.

Workflow Re-engineering
This is where the re-engineering blather comes into play. Given the inevitable technological trends, we've got to stop looking at the workflow from a process point of view and start looking at it from a data-centric one. That is, every job must be designed and processed with the idea that each component on the page is a discrete object made of bits and bytes. The essence of a data-centric workflow is that every job is modular and built out of digital files that do not inherently have anything to do with any of the other components that make up the catalog, book, magazine, etc.

I'd love to tell you that I thought of this and am patenting it, but, sadly, others beat me to the punch. Bill Ray, president of Group InfoTech (East Lansing, MI) and one of the brighter bulbs to light up our industry, has been working on this one for a while, and he's been joined by a number of others, including the vendors selling digital asset management systems.

There are several important consequences of data-centric thinking: First, many, if not all, components can be processed in parallel to increase throughput; and, the ability to complete the job no longer hinges on a single point of failure. That is, when the inevitable font is missing from an EPS graphic nested in a Quark page, you can still produce the magazine on time. But not by using a workflow based on processes. Given the conventional serial workflow, pages are created, assembled into flats and then the entire job is RIPed. That's when you get courier type, tempers flare, and money and time are lost.

Data-centric Workflow
The cornerstone of a data-centric workflow is the database. Far beyond helping you find digital "stuff," databases—such as FileMaker Pro, SQLServer and Oracle—provide the ability to relate multiple objects together without them having to be physically married. This is the quality that enables tasks to be performed simultaneously, such as color correction on images, while graphics are being built and pages are being laid out.

In addition, the relationship can be structured very tightly, so that only if a set of conditions are met will the elements be merged to form the magazine page. This is how we avoid the missing font issue, as the database provides the means by which the operator would know whether or not the EPS had passed a preflight; if it hadn't, or hadn't been preflighted at all, the database would not "release" the graphic for page assembly.

Being Flexible Enough
While the database gets to define the playing field and the rules, a data-centric workflow requires that existing tools be flexible enough to support a just-in-time approach to prepress. While the most sophisticated tools for this run into hundreds of thousands of dollars, there are cheaper alternatives. Among these is the Datamerge QuarkXTension, from Meadows Software, which ties in with virtually any database. This application enables users to build Quark pages with boxes for all of the elements, but without having to insert the contents until the last minute. While the immediate application is variable-data printing, it can be leveraged by any commercial printer or prepress operation that faces tight deadlines.

At the high end lie applications such as Banta Integrated Media's B-media and Group InfoTech's Page Level Automation. These programs are designed to handle significantly larger data streams and enable more flexible and more complex processing to occur.

Since everyone has received direct mail pieces with the wrong salutation or another lame error showing what can go wrong with variable-data printing, a key component of a data-centric workflow is validation. That is, you've got to know that you have the right stuff in the right place. Otherwise, you get a picture of women's skirts with text describing men's slacks and prices for kids' shoes.

Right Stuff, Right Place
Validation eliminates this problem by invoking rules that oversee the database. In other words, you can apply rules for each of the fields (pass/fail) and about the relationships between the fields. Of course, this places a premium at the data entry level, but much of this responsibility will reside with the creative side (read: the client).

Validation can also be applied at later stages of the workflow by automatically generating proofs that are then sent to the client. Whereas we routinely do this now using IRISes and Matchprints, this can be done far more efficiently using PDF.

Since few proofs come back without changes, the benefit of the data-centric workflow pays off when the customer makes changes to her catalog descriptions and prices on her system, and these are then automatically uploaded to the service provider's database; alternatively, the database could reside at the client and the new data downloads to the prepress site. Upon receipt of the corrected proof and new data, a new page is generated and the cycle continues until sign-off. Using scripts and rules, all of this can be scripted so it occurs seamlessly in the background.

While the issue of color and color matching is significantly more complicated, this part of a job can also be subjugated to the rule of rules to expedite production.

If you think all of this is a bunch of baloney, think about this: Valassis Communications produces about 60 million free-standing inserts every week; there are approximately 10,000 different versions, with each running about 48 pages—all in color. By switching over to a data-centric workflow, the company has reduced its labor costs tenfold and narrowed the time frame considerably.

Now that's what I call process re-engineering.

—Alex Hamilton

About the Author
Alex Hamilton, a former technical editor with Printing Impressions, is president of Computers & Communications Consulting, which specializes in digital technologies for printing and publishing. He can be reached at (215) 247-3461 or by e-mail at alexh@candcc.com.