When did publishing move to the desktop?

Answer 1

In the mid-1990s. PageMaker shipped in 1985 and QuarkXPress in 1987, but all the little bits and pieces you needed to do truly pro-grade work on the desktop wasn't around until probably 1997.

You know those pretty pages in magazines that have full-bleed photos with white type knocked out of them? Here's how to do one on a 1989-vintage DTP system.

You first need a customer relationship with a Computerized Electronic Publishing System. There were several brands of these, but toward the end exactly two machines counted: the Scitex Response and the Hell Chromacom. They cost over a million dollars each - not counting the quarter-million-dollar scanner you also had to have - and the hardware was completely different, so no one had both. If your vendor ran Scitex you had to run QuarkXPress, and if they ran Hell you ran PageMaker. You would send them your transparencies - there weren't digital cameras then and "real" publications worked off 4x5 chromes, which are truly lovely things - and scaling data, and they prepared your scans. Then they'd send you a "for position only" file - a low-res image you could put in your layout.

Once you got the file done, you'd send it to the prepress house. Now the fun starts: A Scitex system (I never worked on Hell but I know the Scitex workflow very well) has two "layers" - CT and LW. CT means "continuous tone." This is your scans. LW means "line work." This is your type, vector graphics, boxes that you draw, and so on. The link between the world of Scitex and the world of Mac is called Visionary. It relies on a computer in your Scitex system called "Visionary Interpreter for PostScript," or VIP. These cost about $100,000 (that's in addition to the million you already spent on the Response 300 system) and they weren't perfect; QuarkXPress could do things the VIP system absolutely couldn't decipher, and when you sent it one of those things it would crash the VIP box. If you got it right, though, VIP would generate a LW file you'd then import into your Response imaging station. You could then manipulate the LW to your heart's content, and finally output the file onto your Scitex laser plotter.

Right now I know what you're saying: this is the craziest thing you ever heard of. Well yeah, it would be except that Scitex has far more control over the dots it lays down than a new system could ever dream of. On the desktop we're still stuck with 8 bits per channel, for all intents and purposes - RGB values run from 0 to 255, right? And a nice new fancy CTP rig will do 2540 dpi. Scitex uses 24 bits per channel - over 16 million steps per channel - and, IIRC, a plot resolution of over 5000 dpi. Multiply by four (because there's four colors on a press, ja?) and you see where you're at. Scitex gradients absolutely Do Not Have Banding. Unintentional jagged edges don't exist on Scitex film. It took a very long time for high-end customers to accept DTP.