From one to another: TransGen from CE Engineering Publishing Systems Inc. allows a user to define translations for text as it goes back and forth between a Tandem-based legacy system made by System Integrators Inc. and a Macintosh.

Extracting legacy-system data is newest trick for suppliers

You've been around this business for more than five years, and we know you whistle along with the database believer's anthem, "Anyway you want it, that's the way you'll get it."

Yet your newspaper's classic systems intone Frank Sinatra's mantra, "I did it my way."

Most emerging systems -- business, editorial, advertising, marketing, web publishing, archiving -- can retrieve and sort information, in the text, in the header, with as little as a word or two to start a search.

The grail is a system on which a user can, from one PC, search for information in the editorial, archive and wire receiving systems all at once. On the customer service side, the ultimate is a system that cross-references a customer name in marketing, billing and ad placement stashes.

The older systems -- well, old dogs can't sing new tunes.

A lucky few of your system manager compatriots can declare a Y2K crisis and justify complete system replacements. The rest can teach a few new tricks with tools that push in, drag out, slice and dice legacy content for database use.

What's at stake
More than integrating production systems these days, exchanging data can be critical to relations with business partners, such as ad agencies and advertisers, and your own emerging profit centers, such as marketing arms and on-line servers.

And as LANs, WANs and WWWs become commonplace, your customers' expectations may grow faster than your system conversions.

More tricky than interfacing foreign systems, which was headache enough, building bridges between legacy and database systems means overcoming obstacles that weren't even a thought in the lives of sequestered systems.

First, we now insist that all the meaningful text on the old screen appear on our new systems, even though many standard platforms don't manage as many printing characters or even use matching fonts. (If there's a cent symbol on Atex, we need a cent symbol on the Mac.)

Second, we want the old system's secret signs to disappear -- while useful formatting codes are converted to new-system formatting. Take my familiar favorite, System/55 from System Integrators Inc. of Sacramento, Calif., which drops into the text stream different invisible control symbols at the end of boldface lines than at the end of lightface lines -- just at the point where the lines wrap.

That single difference can derail a string substitution.

Third, we want story-organizing hints, usually stored in headers, to be useful on the new system. Will the old system disgorge header-field data in the order the new system requires?

Fourth, we need methods to reconcile differences in data handling, as there are several ways to express or store dates, telephone numbers, account numbers and billing identities.

Fifth, we expect old and new systems to converse, so these first four requests should happen in two directions. No whining anymore about system brands that once actively barred the doors to standard communications. Today, we want interaction.

Sixth, we'd like databases to be very useful, which means tidying up the content to be searched, sorted and compared. All words for a single item should be identical -- so, three-bedroom apartments should appear as "3BR flat," even though paid typists are inconsistent and on-line users are anarchists.

Even better, we would like text to be normalized or dimensionalized or granularized or deconstructed -- anything that digs meaningful words out of text blobs and nestles them into fields.

Really, it's all very reasonable, yes?

Well, in response, some companies have rolled out products since mid-year to jump the gap between the old and the new.

ATS WebBridge
If you have an Atex J11 system and want to feed a web site, a recently released joint product from Advanced Technical Solutions and FutureTense Inc. may appeal to you.

Announced Sept. 20 at the Association of Publishing Systems Users conference in Norfolk, Va., the system builds a translation and management component into ATS' existing Gateway and NewsDesk products, then packages it with web presentation software by FutureTense of Acton, Mass., and a web server.

The new interface from ATS -- the Wilmington, Mass., company whose Gateway already bridges Atex legacy software into Odbc-compliant database servers -- converts Atex proprietary format, modes and characters into Microsoft Word RTF (Rich Text Format), and from there into HTML and web-savvy file management.

Like previous Gateway versions, this all happens in the background, that mystical place where users never go.

In an Atex-devoted newsroom, this means the on-line editing team can use the same commands and access the same story base as print editors, as long as they use ATS' NewsDesk PCs. With this product, you get an entirely new, database-oriented front-end system, hardware, software and "site-tunable" translation files.

That may be more than you had hoped to install, but compared to building a home-grown interface, the intermediate system may be a worthwhile investment.

CE Engineering's TransGen
CE Engineering Publishing Systems Inc. has twisted its Decade/33 terminal emulation software into another SII interface, DecLink, this one feeding a network server with headers, text blocks and appearance attributes.

Monitoring a set of SII directories, known as baskets, the emulation product copies stories into like-named folders on a Mac server. Along the way, the SII text can be tailored for end-product software with TransGen. For example, SII Styls convert to Quark XPress tags, or a word can be isolated and parsed into fielded data.

Although many SII owners do wonders with their systems' Mapgen and Textgen utilities, it takes a programmer's mind, compiler-level access and lots of testing time to do so. TransGen places its translation controls on the Mac network side, using a simple type-point-click interface.

Properly configured, this pass-and-convert approach can work in two directions.

Michael Brown, advanced systems editor at the Detroit News, has been pummeling the first revisions of the software, although he hasn't finished installing it. Brown said he is satisfied that the approach works for the paper's SII-XPress pagination approach, and he gives the Loomis, Calif.-based CE team high marks for cooperation and responsiveness.

The El Paso (Texas) Times went live in July with its SII-to-XPress implementation.

"It's pretty easy to use once you know what you're translating from and to," said Kate Gannon, systems editor at the Times. She changes font attributes, inserts XPress style tags and even sends stories back from XPress to SII occasionally without redoing boldface and italic type manually. "It's really made a great difference for us."

MCC Business Services' TextCon
One approach to character conversion is MCC Business Services' TextCon converter.

Although conceived for integrating SII System/55 with CCI Europe LayoutChamp pagination, the developers have a universal character-set converter in mind. You supply the beginning character set and the results you want; TextCon will do your conversion invisibly as files pass between systems. Or drag-and-drop files one at a time.

The program works by matching character names or string substitutions, including case changes, wild card and structure matching (which helps with stripping HTML pages).

This program can be developed into a simple text parser; however, it doesn't use logic against header fields. Even with its multiple-pass feature, this is a stream-of-consciousness approach, without "ifs," "thens," "wait.for" and variables to redirect chunks of text.

Future developments depend on MCC customer requests. Now, it's best for feeding text from one system into another, where there are more tools for hashing out results.

SII Gateway
A team of former SII programmers has broken through the cloistering walls around those daunting production and image databases -- with SII's blessing.

In June, Grover Digital Technologies Inc. of El Dorado Hills, Calif., unveiled the SII Gateway, which replaces every communications path SII ever made with a socket directly into its bowels and a standard web-capable browser, customized to purpose or customer.

Intended to open SII systems to all comers, SII Gateway is immediately useful for cutting long-distance expenses out of remote reporter work and for off-loading data entry to advertisers, complete with H&J capabilities.

Cast on a larger stage, Gateway's open view into SII's relational databases -- including newspaper-created data structures -- releases its pent-up powers to manage workflow from anywhere. Or by any signalling device -- say, a Touch-Tone telephone.

The Arizona Republic of Phoenix, Ariz., is installing Gateway with new SpanLink IVR-ware (interactive voice recognition), enabling customers to kill or extend their classifieds by phone. The San Francisco Newspaper Agency expects to do the same later this year.

The Gateway portion was up and running in late September, confirmed Randy Pitcher, developer on Phoenix's IT advertising development team. The first live use will give telephone callers access to their classified ads and the ability to kill a selected ad with a tap of the phone pad.

"From what we can see, the SII Gateway seems to be working very well with the IVR," he said. SII loaded the software on Phoenix's server, and all came up smoothly.

Unisys Atex Ferry
Impressive for its seamless welding of standard platforms, Unisys, of Irving, Texas, starts with a Sybase database core. That permits lots of database advantages, such as full-text searches of everything on the system and automated alerts when the right stuff arrives from reporter or wire.

Welding older front-ends into that frame rather defeats those benefits, but, realist that it is, Unisys makes a go of locking Atex J11 systems into its mix. Called Hermes Atex Ferry, the two-way interface passes page geometry as well as text between the digital nations.

On Atex systems the story shape, or format, is in a separate file from the text, so Hermes pagination speaks directly to the format file. Whether the layout started its life on Hermes and was passed to Atex, or was added on Atex, the results are the same: conversion to Hermes dialect when the story heads to the pagination system.

Since page design changes several times in a day, the layout-language bridge is two-way. An Atex-based reporter or editor sees an alert when a Hermes user orders a shape change, and then sees the change applied.

Technically, the Atex Ferry monitor updates the format file with modified data, updates the header fields of the text file and clears the justified flag, on a single story or on a page of stories. It'll move stories from "justified and ready" to "working" queues as well.

Mission Critical Remote Ad Suite
Best known for systems that wrangle faxes into advertising text, Mission Critical Technologies of Concord, Mass., is leveraging its knowledge of interfaces to improve interactions between newspapers and their advertising customers.

These customers, especially ad agencies and real estate associations, can do a lot of their own ad placement and pricing work, given the right exchange tools.

So Mission Critical (which is being acquired by FutureTense) is shaping a "remote ad management product line" around its core interface and database server, called Adcommand, and the optical character-reading technology in AdFAX.

Adcommand does the basic "hello ... hook me up ... have some data ... give me answers" transactions between remote users and advertising systems. This is the muscle behind the AdFAX and Adfast systems that feed classified front-ends at the Boston Globe and 32 other newspapers.

Now, another layer of tools facilitates more ad component exchanging.

An extension of the Adfast remote ad entry client is Adfast.com, which allows direct classified entry over the Internet and returns pricing to the web user. Since advertising doth not live by text alone, Adfast/Images swallows graphic files attached to the text files, forwards them to the appropriate processing system and helps Adcommand keep an eye on them.

One permutation is AdPartner, an interface for real estate agents developed with Interealty Corp., a Big Player in Multiple Listing Service systems. MCT lists the key benefit as having MLS images arrive with the corresponding property text, as a pair. Status, pricing and other useful information are returned to the real estate office networks.

A third version helps advertising agencies having their own networked ad management systems swap data with newspaper front-end systems. Called AdExchange, it accepts ad files in the advertiser's format, converts them to the newspaper's format, transmits them, and returns information back to the advertiser's system, such as the final text, ad numbers, run date, classification numbers and prices.

Mission Critical's second initiative, AdCapture, addresses the most perplexing task in web advertising -- that of extracting text and images from camera-ready ads. Here MCT is reworking AdFAX technology to tear apart a TIFF and recover the ASCII text, which is oh, so very searchable.

The Boston Globe will beta test.

Other AdFAX features hook this web-worthy "ad" to your contract information and on-line database, so you can generate HTML calls to existing logos, advertiser web links and additional content.

The new tools are entirely advertising-oriented, but so very easy to adapt for editorial.

This year's development appears to be tools aimed at a single problem, and an observer can only hope they are individual parts that will contribute to a harmonious symphony of data dissection.

Meanwhile, other industries are contributing new tools for feeding databases (watch for references to the nimble Junglee Corp.). How those jump start -- or bypass -- newspaper suppliers will be worth watching, as always.

-- Marion J. Love

CE Engineering Publishing Systems Inc.,
(916) 652-5263; (800) 526-5752,
e-mail: cesales@ceengineering.com;
MCC Business Services,
(248) 269-0211, e-mail: mcc@aracnet.net;
Mission Critical Technologies,
(508) 287-0018,
e-mail: psorn@mct.mctinc.com;
System Integrators Inc.,
(916) 929-9481,
e-mail: sii@sii.com;
Unisys Corp.,
(972) 541-8059, e-mail: williamhstroud@unn.unisys.com.

From THE COLE PAPERS, October 1997, Copyright © 1997, All Rights Reserved.