Fill 'er up on the fly: Digital Technology International uses Standard Generalized Markup Language (SGML) tags such as "headline" and "caption" to code text for use in print and on-line.

Sophisticated databases gain ground in web product lines

NEW ORLEANS -- With more than 600 of 1520 U.S. dailies appearing on the World-Wide Web, it's clear that lots of you already are equipped to get your best stuff on-line.

That has only encouraged web systems developers, who are itching to prove that, no matter how you publish on the Web now, they can go one better.

At NEXPO '96, suppliers weren't sure of newspapers' commitment to ever-morphing standards and practices. At NEXPO '97, suppliers presented a flock of products for all you webmasters outgrowing your Stride-Rite baby shoes and ready for Keds.

Notable was the number of packages that help you put your own content on your own web server. When ROI for web hardware was a daydream, plenty of publishers held onto their capital and instead trickle-charged hosting companies, such as AdOne, InfiNet, local ISPs and web malls.

Those outsiders risked the cost of equipment and upkeep, often in return for shares in the income rewards. Now that it's evident there are profits and markets to fight over, publishers are more sensitive about "growing a competitor" through outside partnerships.

When the time is right to bring your content home, suppliers are jockeying to equip you. NEXPO '97 debut solutions fell into three camps: web-feeding extensions from production storehouses, stand-alone subsystems dressed up for one advertising segment, and retrievalware enabled for 'Net browser access.

Stretching production systems
Entrants in the first category, notably Digital Technology International and CCI Europe, put a beat of hope into the breasts of Cole Group correspondents.

It has been downright demoralizing to hear about "repurposing information for new business opportunities" at the high-ideal workshops and then find mostly shovels on the show floor. HTML converters and filters that run against header data are necessary basic tools in a web workshop, but they still just slap markup codes around print-oriented text, which gushes out on an electronic imagesetter.

DT honors the concept of separating form from content. Make a web-page template that directs the look, then leave the space reservations blank until the reader requests the page, explained Don Oldham, president of the Orem, Utah-based company.

With this scheme, an editor or designer lists in each space in an HTML template the name of the desired file in the production system. At the moment of demand, DTWebClient heeds the web server's need and chauffeurs the content to fill in the blanks.

Think of it as just-in-time delivery, something like those FedEx trucks on TV filling empty stores and factories with goods just before opening. It means the reader can have the news de momento as well as the classified ad placed an hour ago.

It also means the on-line team will have the same tools and training as the groups creating today's news and ads; they'll just be managing a new edition.

Since DT's production systems use an SGML (Standard Generalized Markup Language) tag strategy for inserting typography into printed editions, converting them to HTML for web presentation is a short hop.

If you see problems with this approach, Oldham has answers.

"The objection always falls into one camp, and almost all of my biggest customers have objected," he said. "'We don't want outsiders to be in our production database!' How about a replicated database, a copy of the database?"

While that means extra server CPUs and hard drives, you can pare down the mass by selective replication, or only copying "Edition = Web." How much that costs depends partly on how brawny the hardware must be, which is a factor of how many concurrent requests you expect, but it still can exceed $15,000.

To keep up the speed, Oldham recommends keeping that replicated edition and web server close together -- on the same hardware even. While DTWebClient can respond to an offsite web host, cozy relationships still work best.

DT isn't the only old-guard supplier stretching its production system into web publishing. First glimpses of Electronic Media Server from Denmark's CCI Europe, whose U.S. office is in Marietta, Ga., hint that it will use its database and text tags in ways similar to DT (see The Cole Papers, May 1997).

Harris Publishing Systems Corp. of Melbourne, Fla., still seems shy about webbing its XP-21 database. At NEXPO, its new offerings were the Quest server for classified ads (see below) and an HTML-converter fed, static-page organizer called WebDelivery by New Horizons Team of Pottsville, Pa.

The latter is a good solution for baby-steppers; it won't thrill the Keds or Nike crowd.

A few other centralized systems suppliers, such as Baseview Products Inc. of Ann Arbor, Mich., are talking about web clients to come.

Sure, this tap into the print publishing system and automatic HTML conversion still smacks of shoveling -- if the editors simply toss the same stories unedited into the web server and classified ads remain blobs of unparsed text. If you are a database purist who believes that free-text and keyword searches are poor ways to sort classifieds on-line, you'll be interested in systems that start with fielded data.

(Oldham has a trick up his sleeve to help ad-takers insert facts into fields, so watch those space reservations on his web pages for details. DT is at http://www.dtiint.com/.)

Write once, sell many
There's no wait for fielded-data fans at the stand-alone server booths.

Disguised as automotive classified ad or real-estate agent or entertainment-guide systems, these servers are of one flesh, with the mission of enticing someone -- besides newspaper employees -- to type the copy right into database fields the first time. Then these servers have a field day driving info to readers.

You've seen the possibilities in the national real-estate web sites. Similarly, television listings suppliers harness database benefits in their editing tools for Mac-paginated grids and logs. Just look over the shoulders of your TV listings fine-tuners, or try the "personal listings" choices at Tribune Media Services, http://www.tmstv.com/. It's worth a thousand of my words.

Many NEXPO debutantes were auto classified systems, younger sisters of real estate systems, dressed to emphasize profit potential rather than technical pizzazz. The business aim is to get the dealers to enter the car characteristics and photos, then drive them into more printed products as well as on-line.

Take that, AutoTrader!

The systems have several common features. They start with getting all the inventory recorded with all the special attributes of each car in separate fields. Often the information is lurking in an automotive system from Reynolds & Reynolds, ADP or DCS, a business system or Kelley Blue Book reporting system.

On cue, which can be a scheduled automatic upload, the dealer's system sweeps the selected car records into the auto ad server, from where they are posted to the Web and forwarded to the ad production system, dragging photos and logos along while sold cars fall off the list.

The Web posts them for all to see, sort and cost-compare. The newspaper collects its classified liners, but often gets more, like copy and photos for those frenetic Sunday sale ads or ad-matching fax and e-mail services for readers.

A sample of the products available: