Production: The Paste-Up Era, 1989–1994
The Pipeline
In 1989, a paragraph traveled a considerable distance before it appeared on a printed page. A writer finished a piece and handed it — literally handed it, or sent it by messenger — to a typesetter. At a large magazine the typesetting might be done in-house, on a Compugraphic or Mergenthaler phototypesetting machine; at smaller titles it was farmed out to a dedicated typesetting house, many of which operated out of rural New England or Pennsylvania and accepted manuscripts by overnight mail and returned output within a day or two. The machine exposed type onto resin-coated photosensitive paper — the RC galley — producing a strip of glossy, faintly plastic-feeling original that was bright white and slightly warm to the touch. The typesetting house proofed this output, corrected it, and shipped it back. The original RC strip went to the production department. Xerox copies — standard 20-pound bond, warm off-white, toner sitting slightly raised on the surface — went to editorial for markup. The markup came back. The typesetter incorporated the corrections and produced a corrected strip. That strip went to paste-up.
Paste-up was the stage at which the physical pieces of a magazine issue were assembled into pages. It happened in a room — usually a floor below or above the editorial floor, often kept several degrees cooler because the wax machinery ran warm and the workers needed to concentrate — staffed by compositors whose job was to position every element of every page on a board, in sequence, in register, until the issue was done. The completed boards were photographed to make printing plates, and the plates went to press. This pipeline — manuscript, typesetter, RC galley, editorial markup, correction, paste-up, plate camera, press — was standard across the American magazine industry by 1989 and had been, in its basic form, for twenty or thirty years. It was also, in 1989, already being shortened from both ends.
The Mac was on the art director’s desk by 1989, running QuarkXPress 2.0 or 2.1 or Aldus PageMaker, producing layout compositions on screen that were output to a Linotype L300 imagesetter — a piece of machinery roughly the size of a large kitchen appliance — which produced RC paper or film for paste-up. The Mac had not replaced paste-up. It had inserted itself into the pipeline between the typesetter and the board, taking over some of the typesetting function, but the page assembly still happened by hand. One industry veteran, quoted in a contemporary trade account of Condé Nast production culture, described the exact 1989 workflow this way: “I would set and output graphics from my super fast Mac IIci with PageMaker or Illustrator 88 to a Linotype L300 imagesetter on RC paper and paste-up (wax) everything up to a board.”1 The man in the cold room with the X-Acto knife still held the page together. The Mac made his raw material.
The Board
The central piece of furniture in a paste-up room was the parallel-motion drafting table — a large, adjustable drawing surface with a horizontal straightedge that moved in parallel tracks on both sides of the board and could be locked at any height, allowing precise alignment of horizontal strips without a separate ruler. The board surface was a sheet of heavy cardstock or illustration board printed in non-photo-blue ink: a grid of lines, column guides, margin marks, all rendered in a color the platemaker’s copy camera could not detect and would not reproduce. The compositor worked against this invisible grid.
To his right — almost always his; the compositing trade in 1989 remained predominantly male, and the union rolls reflected Italian and Irish surnames at rates that varied by city — stood the waxer. The Lectro-Stick beeswax waxer was a device roughly the size of a toaster, with a heated drum that took about thirty minutes to reach operating temperature. The wax inside was honeycombed beeswax, dyed pale yellow, melted to approximately 160 degrees Fahrenheit. To apply adhesive to a galley strip, the compositor fed the strip of RC paper through the waxer’s rubber rollers; the heated drum deposited a thin, even coat of wax on the back of the strip. The strip could then be pressed onto the board and repositioned as needed until it was burnished down with a rubber roller, after which it held. The wax was repositionable, which was essential: a compositor might place a headline, realize it was a pica off from register, lift it with a fingernail, and reset it without damage.
The room smelled of several things simultaneously. The dominant note was beeswax: slightly honey, slightly damp paper, warm. The second note was Bestine — a rubber-cement thinner whose active ingredient was heptane, a petroleum solvent, kept in a flat metal can on every paste-up table. Bestine was used to lift placed elements, to dissolve excess wax residue, to clean the blade of the X-Acto knife. When a compositor opened the Bestine and wiped down a board, the heptane vapor reached the nose immediately, producing a brief, involuntary head rush — not unpleasant, not quite dangerous, simply present, like the smell of a gasoline pump at a specific moment of the morning. The third note was tobacco smoke, because in 1989 smoking was still permitted in private offices on the editorial floor above, and editors came downstairs with cigarettes.
The X-Acto No. 11 blade — a fine-point scalpel blade mounted on a lightweight aluminum handle — was used for trimming. A compositor trimming the bottom rule of a galley strip used the knife in a single, controlled pass against a steel straight edge; if the pass was not controlled, the blade skipped and the strip had to be re-output. The blades were replaced often. The index finger of the left hand, which held the straight edge in place, accumulated nicks and calluses over the years. A compositor who had been doing this work since the late 1950s had an index finger that showed it.
Photos were handled separately. The photomechanical transfer — the PMT, produced by a stat camera, a large floor-standing copy camera — was a positive print on photographic paper, sized to fit the layout. The compositor positioned the PMT on the board and crop-marked it with a red grease pencil: diagonal lines at the corners indicating the desired print boundaries, a notation in the margin for the stripper who would handle the film. The stat camera occupied its own corner of the room or a dedicated closet; the chemical trays were nearby. The room was not entirely dry.
The Markup
The galley proof — the Xerox copy of the RC original — was the document on which editors, writers, fact-checkers, and copy editors communicated their changes. At a monthly magazine, each proof passed through several hands in sequence and sometimes simultaneously; a collating department or a single collator reconciled all the marks onto a single master before the corrections went back to the typesetter. At The New Yorker, where this process was most elaborate and most documented, the path ran from the assigning editor through copy editing through a dedicated grammarian’s proof through fact-checking through collation through a final “Page OK” proofreader — six distinct sets of marks, each in a different hand, often on different-colored paper, reconciled by a collator whose job was to copy every change without interpretation, then check her work three times per page.2
The color conventions were not universal but were widely observed. Red for the editor — the highest authority on the page. Blue for the writer. Green for the fact-checker. Black for copy. At The New Yorker, pencil rather than pen was the norm, because pencil was erasable and revision was expected. Eleanor Gould Packard, the query proofreader at The New Yorker, worked exclusively in pencil from her arrival at the magazine in 1951 through the Tina Brown years in the 1990s — decades across which the magazine moved from hot metal to cold type to, eventually, digital composition — and she never used a computer.2 Her marginal queries were specific: “How so?” “This clear? (not to me).” “NOT Grammar!” By the time she finished a proof, Mary Norris wrote, the pencil lines on it looked like dreadlocks.
The proofreading marks were standard across the industry: the caret for insertion, the dele for deletion, stet to restore a word that had been marked for deletion and then reconsidered, transposition loops, marginal notes keyed to in-text marks with a line. Chicago Manual of Style marks were the baseline. Each magazine had house conventions that overlaid the standard marks — house spellings, preferred hyphenations, the specific style decisions that accumulated over decades into the publication’s fingerprint. A collator at a major magazine had to know not just the marks but the conventions they encoded. This knowledge was institutional memory. It lived in the collator’s head, and it was not transferable to a software system in any direct way.
The work was exacting because it was the only thing standing between the manuscript and the printer. A compositor who set a line wrong, an editor who failed to catch the error, a collator who missed a correction — each failure compounded forward through the pipeline and arrived on the printed page for several hundred thousand readers. The collating department was not a glamorous floor. It was not the floor with the editorial lunches and the editorial assistants who had gone to the right schools. It was the floor where the issue got closed, correctly, or not.
The Hybrid Moment
By 1989, the Mac was on the art director’s desk. By 1990, QuarkXPress 3.0 had shipped with its measurement palette and library support, and by May 1992 QuarkXPress 3.1 had consolidated those features into a tool that could handle body type, captions, headlines, rules, and screens with sufficient reliability for production use.3 The standard machine by 1992 was the Mac Quadra 700 or Quadra 950 — faster processors, larger RAM, capable of running a full Quark workflow — with output to a Linotronic 330 imagesetter that produced RC galley of a quality comparable to dedicated phototypesetting equipment.
The transition proceeded element by element rather than all at once. Body copy moved to the Mac first, then captions, then headlines, then rules and screens. What stayed on the board longest was photo positioning, cropping, and last-minute patch work — the small wax-backed corrections that compositors called stickies, trimmed from fresh output and pressed onto the board over the element they replaced. A page in 1992 might have been composed almost entirely in QuarkXPress and output to RC galley, then assembled on the board with wax and trimmed with a knife, with three stickies on the last column where late editorial changes had come in after the initial output. The Mac and the parallel-motion table coexisted on the same floor, within ten feet of each other, doing consecutive steps of the same job.
The art director’s relationship to the technology shifted first and most visibly. Where an art director in 1988 had sketched layouts by hand on graph paper and handed them to the compositors as instructions, an art director in 1992 was composing on screen, moving elements with a mouse, seeing the page as it would approximately appear before a single physical strip had been cut. The Mac gave the art director a kind of authority over the finished page that had previously required the compositor as intermediary. The compositor still assembled the board, but the board now reproduced a page the art director had already approved on screen, rather than realizing a pencil sketch the art director had described. This was a change in the nature of the job, not just its tools.
The compositors adapted to this change to the degree they could. The skills of precision alignment and knife work remained necessary; the output they were working with had changed. Some compositors learned the basics of the Mac interface and could troubleshoot output problems at the Linotronic when they arose. Most did not. The Quark skills belonged to the art department, not the production floor.
The End of the Room
By 1993 the Quadra workflow to PostScript output had made the board redundant for most operations. Pages were built on screen in QuarkXPress, proofed by printing to a laser printer, corrected on screen, and output to the Linotronic as film rather than RC paper — film that went directly to the plate-maker without requiring a paste-up stage. The board was not decommissioned in a single moment; it persisted for the last-minute patch, for the occasional element the software could not handle cleanly, for the habits of a production floor that had operated one way for thirty years. But by 1994 at most monthly magazines, and by 1995 at virtually all of them, the board was a backup, not the primary surface of assembly.3 By 1996, the American magazine was essentially clean.
The compositors did not move to the Mac. The transition that replaced their jobs was not offered to them as a retraining opportunity in any systematic way. Some were offered package buyouts — lump-sum severance payments in exchange for departures that the publisher did not have to characterize as layoffs. Some were reassigned to “production coordinator” titles that involved scheduling, vendor liaison, and light administrative work and lasted one or two production cycles before the title was eliminated. Some were quietly let go when their contracts came up or when the magazine restructured the floor under a new title for the production department. The San Francisco Chronicle, the last major American newspaper to maintain a full compositing room, sent its last physical page through the waxer at the end of 2003; for magazines the clearing had happened years earlier.3
The room after the wax machine went away was a different kind of room. Gone was the warm beeswax smell. Gone was the chemical bite of Bestine, the heptane head rush at eight in the morning. Gone was the low sociability of the compositors talking across the light tables while designers stood at their elbows and an editor could walk in and say “kill the third graf, run the photo bigger” and watch it happen with a knife and a roller in ninety seconds. What replaced it was a row of monitors, a cluster of laser printers, the high-pitched whine of a SyQuest cartridge spinning up to speed, the smell of warm toner. Everyone was seated. Everyone was silent. Everyone was individual. By 1996 the composing room of an American magazine was, on most floors, a server closet.
The craft that had been retired was approximately thirty to forty years old at the floor level — not an ancient tradition, but old enough that the men who held it had built working lives around it, had trained in it, had done it every day until they did not. Most of them were in their fifties when the buyouts came. The magazine industry in 1994 and 1995 had no particular mechanism for helping them find new work. There was no trade association program for displaced compositors, no retraining initiative, no industry-wide accounting of how many people had held the job or what they did afterward. The New England typesetting houses that had serviced New York magazines through the 1980s closed one by one through the same period, for the same reasons. The rural keystroke operators who had set type for metropolitan magazines, invisible to the editorial floor, lost their jobs at roughly the same time as the compositors in the city, and with less notice.
The craft that ended between 1989 and 1995 was not ancient. It had been built up through the phototypesetting era — roughly from 1950 forward — on top of earlier hot-metal traditions, and it had employed, at its peak, a specific layer of skilled, mostly unionized labor in the production departments of American newspapers and magazines. The compositors at a monthly magazine in 1989 had mastered a precision manual skill: the blade work, the wax, the non-photo-blue grid, the three-pass collating check that was the only thing standing between a manuscript and a printing plate. The skill was not obsolete because it had stopped working. The pages it produced were indistinguishable, to a reader, from the pages that Quark and a Linotronic produced in 1993. It was obsolete because the new method was faster, required fewer bodies, and put the page directly in the hands of the art director without a translation step. That the people who held the old skill were in their fifties, that they had spent twenty or thirty years accumulating it, that most of them would not work in print production again — these facts were recorded nowhere in particular. The industry did not count them. It counted the issues it closed, on time, at the correct trim size, with the correct color separations. By 1996 it was closing them faster than before. The years across which this happened, Bush’s term into Clinton’s first, produced no policy that addressed what it meant to retire a trade. The trade retired. The room went quiet. The wax cooled and was thrown out.
A magazine like Meridian would not, in the ordinary run of its editorial calendar, have assigned a piece about its own composing room. The floor below was production; the floor above was editorial; the transaction between them was the closed proof and the corrected board, and the people who managed that transaction were not subjects, in the floor’s daily accounting of what counted as a story, for the kind of writing the magazine did. The compositors would have been as present as the maintenance staff and as invisible in the same way.
The tougher assignment — a profile of the men with the scarred index fingers, in the months before the buyouts — would not have made it past a Monday morning meeting in 1994. An articles editor might have raised it as a labor story, a story about what the digital transition was doing to a specific tier of skilled workers, people with union cards and thirty years of institutional knowledge and nowhere else to go. The proposal would have met not resistance but a kind of bewilderment. Which byline on the masthead would have taken it? No regular contributor would have worked that floor, would have known those men, would have understood what precision looked like when it lived in a blade and a wrist rather than a cursor. The piece would have required a writer willing to stand at the parallel-motion table for several weeks and learn the work well enough to describe it from inside. No name in the rotation would have fit. The proposal would have made one pass through a pitch meeting and died there.
A senior editor on the eighth floor would have walked past the composing room every morning for ten years on her way to the elevator. She would have walked past the man at the table — the one in the flannel shirt, the one with the reading glasses on a cord — without learning his name. His scarred index finger. The smell of beeswax in the corridor outside the production room, a smell she would have registered and not registered, a smell she would have associated with the closing of the issue and not with a person who produced it. The last pencil-marked proof she would have received from the compositing floor would have arrived in 1995, on a Tuesday in October, and she would not have registered that it was the last one. She would have marked it in red. She would have handed it back. She would not have thought about the man who had assembled the page it described, or about what he was going to do the following month, when the board went away and the server came in and the hallway stopped smelling like wax.