WORKING WITH DATA CONVERSION LABORATORY

Virtual Engineering
A peek into the work of DCL's Project Engineering Group - the folks who map out and oversee the data conversion process...

ONCE A PROJECT has gone through the sales phase at Data Conversion Laboratory (DCL), it goes to the Project Engineering Group. They're the people who analyze incoming data and plan the project workflow down to the last detail. With the mission-critical projects DCL take on, nothing can be left to chance. For example, most conversion efforts need some level of human editing and so the Project Engineering Group analyzes exactly what needs to be done, down to how many seconds it should take per page, and how to ensure that it gets done the same way every time.

To find out more about what they do, I spoke with two members of the Project Engineering Group - Mikhail Vaysbukh and Lesly Gross - and asked them how the typical Project Engineering process begins?

"We often get involved early on to help determine scope and price the project," replied Mikhail. "But once a project actually starts up, the first stage is a kickoff meeting with the client. This could be in person or via a conference call. We basically go through the fine detail of the project and confirm the schedule and deadlines. We also make sure we have all the materials before starting a project - if any are missing, we reel them in! We make sure everyone's on the same page and that we have all the same expectations."

The Project Engineering team typically consists of a project manager, data analyst, and programmer. Essentially, the team is responsible for gaining a thorough understanding of how the client wants their material converted - what level of XML/SGML tagging is required, and so on. They also strive to get a clear vision of how the material should look at the client's end. They can then write up the specification document, which details all of the elements that will be encountered in the client's data and how they will be handled in the conversion process.

Hand-tagged sample
In the first instance, the Project Engineering Group prepare a project specification in conjunction with what they call a "hand-tagged sample," which involves taking a few pages of a file and tagging them by hand. The combination of these two documents allows the client to make sure everything is defined correctly and to get a first look at what the finished results will look like. And if anything needs altering, this is the time to do it.

Once the hand tagged sample and conversion specifications are ironed out, the Project Engineering Group develops software to automate the conversion process. They then submit a production sample to the client, which is a larger sample of data that has been tagged by the software.

Human input
On most projects, a degree of human input is required to clean up the data for conversion. "Software will get you 90% of the way there," says Lesly. "But there's always going to be a manual piece."

While the high-volume requirements of a project are dealt with by the Production Group, testing of the whole process on the initial small batches is done under the watchful eyes of the engineering group. An editor works with the engineering group to map out the editing process.

"The editor has to test it out here first," continues Lesly. "That way, before we put a thousand page book into production, we know all the editorial steps needed and that it should, for example, take two minutes per page to do. If in production, our control systems tell us that it is taking ten minutes per page, we quickly can tell that something's wrong and can diagnose the problem and put it right."

To keep costs down and reliability up, the manual effort is kept to a minimum. Usually all it takes is a few tweaks of the cleanup software to automate much of the conversion process. For DCL, automation, with intelligent input from humans, is a key selling point. It allows the company to convert very large volumes of material fast and efficiently - and at a cost that is attractive to most companies involved in large publishing efforts.

DCLnews Editorial

Read more data conversion related articles at DCL Library

Return to top