Desktop publishing (DTP)
Coined in 1985 the phrase was intended to distinguish low-cost, home- or office-based publishing systems from the large, expensive equipment that served the hot-metal presses that dominated the newspaper industry at the time.
According to Wiki, the term was invented by Paul Brainerd, co-founder and co-owner of Aldus Corp. Aldus produced the first DTP software, the all-but forgotten Pagemaker. Like so much else in early computer development, Pagemaker and the other fundamental DTP technologies grew out of work done at Xerox' laboratories*.
Pagemaker 1.0 was released in July 1985, and I guess if we have to give a date for the start of the DTP revolution, that was when it all began. The revolution depended on Apple Mac computers, Adobe PostScript formats and Apple Laserwriter printers. It was Pagemaker that brought these things together to deliver the first true desktop publishing capability.
Xerox PARC was the birthplace of all these technologies. John Warnock and his boss, Chuck Geschke founded Adobe Systems, while the precursor of the PostScript language was also developed by Warnock and Martin Newell at PARC. Apple Computer decided to put PostScript into the Apple printer drivers. The final link in the chain was Linotype making its systems compatible with PostScript.
Twenty years later, in 2006, the heavy, cumbersome hot metal machines have gone the same route as mechanical typewriters, telex machines and analog computers. I guess they still exist somewhere in the world, but the professional publications industry no longer has the time or money to keep them going.
DTP is faster, better and cheaper than any other approach, so DTP has become the standard approach in modern, professional publishing. We might as well drop the "desktop" bit and just call it publishing. For that is the modern reality: DTP systems are one of the workhorses of the publications business. We use those tools to create our publications, advertisements and flyers with hardly a moment's thought to the out-dated processes of our parents' generation.
The transition from conventional printing and publishing to DTP-based systems took about ten years, from Rupert Murdoch's move to Wapping in 1986, until about 1996, when almost all magazines had gone digital.
I don't think there was a similar confrontation in the US. Print unions in the UK were very powerful at the time, and DTP provided the technology for pioneers like Eddie Shah and press barons like Murdoch to break the unions' grip on the newspaper industry. While DTP has changed newspapers, the biggest effect has been on small magazines where costs have plummeted, allowing many new start-ups to flourish in ever-smaller markets.
In the course of our work in magazine publishing, we write about many industries and activities. One of the recurring themes of our stories is the process of change and the transition from a local economy to a global one. We write about the problems companies face as they respond to competition from China and Sri Lanka, and the changes necessitated by that globalisation: downsizing, increased flexibility, faster times to market and so on.
Not a single one of those businesses has changed anywhere near as fast, or as completely as the publishing business.
In ten years, we removed two or three whole classes of employees from the production process and the journalists became much more involved in page make-up, production and story editing.
In ten years the news and publishing industry switched from a business where the cost of entry to global publishing was measured in millions, to a business where any college student can compete with Reuters or the BBC. The rise of weblogs has turned the potential of DTP into a very real threat to conventional publishers. There are plenty of people who prefer a personal weblog to the dry, somewhat anodyne story-telling of Reuters, CNN and the Beeb.
In those ten years, the business changed from operating on a timescale of hours or days, or even weeks, to one where five minutes might be the difference between being first with a story or being behind the times.
DTP was merely the first step along that road, but the step was a critical one, as it paved the way for news and analysis to go digital. DTP changed the timescale for small-volume publications from months to days. A decade later the web changed the timescale again, from days to minutes.
What DTP means, and what you need to do it
DTP is an ill-defined phrase that usually means the process of taking extant words and pictures, editing them on a relatively cheap home computer, before assembling them onto a page, and then preparing them for printing on either a local printer or for transmission to an industrial-scale printing plant. As the distribution of news by electronic means has grown, DTP has come to encompass web-based presentation of information in addition to paper-based publications.
The key tool in this process is a computer equipped with a range of different software. The most basic software tool is a page make-up program, which enables the operator to bring together text, images and other elements, edit them and adjust the relative sizes and positions on the page. Typical programmes are QuarkXPress, InDesign and (previously) PageMaker.
One key feature of the page make-up software is that you have to see on-screen exactly how the page will print, down to line-breaks, picture spacing and text size. Pagemaker was the first to bring this feature to the publishing business. In those days computers tended to have monochrome screens and most had only one monospaced font and defaulted to an 80-column display, to mimic the typewriters they would soon replace entirely. PageMaker and PostScript made a huge advance by introducing proportional fonts in different sizes and the ability to run text around illustrations on a computer screen.
At the time, they called it WYSIWYG (What you see is what you get). Computer gamers understand the need for dedicated graphics cards, but back in the late 1980s, business computers never needed them -- even Apple Macs. DTP was one of the first applications to demand large screens, and those large screens demanded special cards to display the output.
Beyond the page layout software, the prospective desktop publisher needs some image manipulation software. This is used partly to modify the content of images, but more often to convert RGB images to CMYK and to adjust the pixel count of the images to suit the size and resolution of the final vehicle, such as computer screen, PDF or paper print. Photoshop is the de facto standard for this task. It has all the muscle needed to handle any task related to image manipulation. However, many simple programs can convert RGB to CMYK and crop and re-size images, so Photoshop is useful, but not completely necessary.
Next there is a text editor. Most page make-up packages include some kind of text editor, though the Quark version is fairly rudimentary. Most newspaper and production offices I know of aim to get spelling, facts and grammar right in their preferred text editor, before the text sees any kind of DTP package. This is often Microsoft's Word, used with the absolute minimum of formatting -- and with that accursed paperclip animatron banished from the screen forever. Personally I use the TextEdit software that comes with Mac OS. Once the text is near finalised, the DTP jockey squirts the text into Quark or InDesign at the last moment, where he can apply appropriate formatting. The built-in text editors are then used for cutting, tweaking and final adjustments.
The publisher also needs access to a variety of fonts. In the 1990s it was font incompatibilities between Wintel-based systems and Mac-based systems that led to one of the biggest schisms in DTP culture. The PostScript printing standard of 1984 launched the DTP revolution, allowing anyone with a compatible laser printer to reproduce different typefaces accurately and consistently. Unfortunately, Microsoft was late in adding PostScript compatibility, and this excluded Windows machines from playing in the DTP sphere until the mid 1990s. Later the use of TrueType and other standards removed many of the cross-platform issues. In the far-off days of the early 1990s, there was no choice for DTP staff, except to run on Apple Macs. Even though the rest of the organisation might have been on Windows, the DTP department were always Mac-based mavericks.
The final set of tools is a pre-press system, which converts the page files into something the printer can use. Nowadays this is either a proprietary format, or, more commonly, a full PDF. The PDF removes any incompatibilities between the systems at DTP and printer, and when prepared correctly, contains all the font information and all the image information needed to correctly print the page.
Personal recollections of the transition to DTP
In 1988 I moved to a new magazine. To my horror, they still used manual typewriters and coloured pens. I was accustomed to a word processor to write and modify stories. But whether typewriter or word processor, the standard office technology of the day could only produce text in a single font. We had to tell the printer exactly how each piece of text -- headlines, captions, straplines, pull-out quotes, page numbers as well as the body text -- would look in terms of font, size, alignment, italics, bold, spacing, width and so on.
The production process comprised many steps and by today's standards was extremely time-consuming. The shortest interval I could manage between writing a story and seeing it in print was 21 days. Three weeks. More typically, the deadline for text was six weeks before publication date.
- Type the text
- Read it, mark it up, re-type it if necessary
- Mark up the final copy, indicating font type and size; left or right alignment, bold, italicisation, and so on.
- Select photographs, prepare illustrations, stick to backing sheet, indicate size required,
- Send marked-up text and marked-up images to printer.
- Wait four to five days for marked-up text to come back as a galley All the text had to be re-keyed at the printer.
- Read, mark corrections (should not be necessary) Return to printer, wait another two to three days
- Take galley, cover the back with solvent-based adhesive, cut up into blocks and stick the blocks to a mark-up base sheet.
- Re-arrange make-up, by lifting cut copy, pictures, and re-positioning them. (Repeat as necessary) write captions, specifiy page numbers
- Send to printer; wait two to three days for return in page proof form.
- Final proof reading; final marks. Add and subtract lines to make stories fit the page.
- Clear pages.
- Ozalids returned, which include adverts, pages and other elements. Check for errors, final clear
- Printing
My first step on the road to DTP was to buy computers and word processors for the staff. That was in 1990, when you could buy an Amstrad PC with black and white screen for about £700, or a couple of weeks' wages. A 32MB hard drive (we called them Winchester disks) was a luxury.
The move to word processors allowed us to alter stories at will, without the need to re-type. It saved a great deal of time, even if we had to store them on 51/4 -inch floppy disks.
The next step in time-reduction was to make the modems work. I remember upgrading from a 1200 modem to 2400, and then to 9600 and the difference it made to transferring text from our office to the pre-press company. The company we used, Dan-Set Graphics, was associated with the printing company we used for the magazine. My highest priority at the time was to remove the delays and mistakes introduced by re-keying all the text. So we got the modems to work and we sent text files electronically using direct modem-to-modem dial-up connections and weird and wonderful protocols like Xmodem and Zmodem.
Soon after, in 1989, Dan brought in a copy of QuarkXpress Version 2.1. He was one of the first bureaux to offer electronic page make-up. At that time, Quark was trying to muscle in on the market developed by Aldus' Pagemaker software, but they didn't do it very well and XPress v1 was too flakey while 2.0 was full of bugs (typical of Quark). 2.1 was the first version of Quark to behave itself in a production environment.
And I remember to this day the astonishment I felt when I sent the first late news page to Dan. We had agreed a simple two-column format including fonts, widths and headline styles, so that no mark-up was needed. I sent the stories away by modem and within an hour I had a page in front of me (by fax) that needed only a couple of minor corrections before it could be cleared for press.
In a stroke, that cut the minimum time from typing a story to finished magazine by 75 percent, to just five working days.
The next step was to convert all our pages over to Quark. Dan built the templates and this stored all our standard styles so that we only had to send the text up, and indicate that it was "headline1" or "body text" or "caption" style. Dan hit the right key, and it was converted in an instant. So we sent the text up by modem, with a marked-up copy by fax and it came back either next day, or a day later as galley proofs, printed out on his laser printer. More time-saving. More cost-down.
We still made up pages in the same way, using Evo-stik and mark-up sheets, but the turn-around had speeded up dramatically.
We carried on in this way for a year or so, but in 1990, just as Quark XPress 3.0 was published, we decided to spend out on a set of Apple Macs for the office. A couple of tiny Mac Plus units and a Mac IIci, all with black-and-white screens.
This was the point of no return for DTP. We were one of the first business magazines in the country to do full DTP. Our printer -- one of the biggest in the country -- had been using us as guinea pigs in how to deal with the new technology since we first went down that route, and they told us that no-one else was using DTP in the same way as we were.
I say full DTP. Not quite. We still didn't do colour pictures. We could scan and process black and white images, but colour took longer and needed more expensive equipment. While we did low-resolution scans of pictures in black-and-white, we still sent the originals out to a bureau for hi-res scanning. Modem transfers of pictures was too slow, even at 9600 baud, so we sent the physical pictures up to Dan who scanned them and adjusted the size and colour to suit our own layouts.
At the time, the big newspapers were in transition. Murdoch had moved out to Wapping in 1986, and suffered 13 months of strike action by the printing unions as they saw their jobs sliding away. Thanks to the new technology, the people who used to re-key text were no longer needed. The people who chose the different fonts and set the machines up had no role in the brave new world of DTP. A whole army of skilled and semi-skilled workers was made redundant in the space of a couple of years.
The big newspapers had huge, expensive, integrated publishing solutions that linked journalists, typesetters and layout staff. On the small trade journals, we may have ridden on their technological coat-tails, but we were the pioneers of true DTP technology.
The final step -- colour images -- came a couple of years later, as we upgraded the computers and screens. It no longer seemed anything special. But it was still another 10 years before our US-based parent switched over to DTP. The advertising business took a little longer than the publishing industry, possibly because there was less pressure to save money in advertising, but today most ads for print media are prepared digitally, the last part of the process to go that way.
Nowadays the production process runs more like this. It all happens in our small offices, so the lines of communications are much shorter. In terms of speed, we no longer think it a miracle that we can research a story in the morning, write it, get photographs and send the cleared page to the printer in the afternoon.
- Type text. read it, proof it and correct it before putting it up as web-based news (daily task)
- Take the bulk text and squirt it into XPress. Format stories. Print as galleys (or not) (once per issue)
- Select stories for final publication, source images.
- Assign stories to pages, use Quark-based grids to make them fit. Chop large chunks of overmatter
- Adjust picture size to make stories fit.
- Write heads, captions, straps etc (on screen, in Quark)
- Print and proof
- Final marks and clear
- Assemble pages with advertisements
- Convert pages to PDFs
- Send PDFs to printer (either as CD-rom or by FTP)
The interval between writing a story and seeing it in print is still about a week, but this is entirely due to the printing process. If we paid the printer more, we could reduce that to a day or two.
Some notes on equipment and software
The more screen space you have, the better. If you want to get into DTP, buy the biggest screen you can afford. Some video cards can drive multiple screens, and that is usually cheaper than buying a single huge screen.
When building a page, it is much easier to get it right when you can see the whole page -- and usually it is better to see the whole two-page spread. You also need to see the text at sufficient size to read it properly. Add to that the palettes, tools and different windows you need available and I find that a 1280 x 1024 screen is nowhere near large enough. Adding a second screen of 1024 x 768 makes it acceptable, but I still find myself juggling windows to get a decent view of the page I am working on. In an ideal world, two 1280 x 1024 screens (or one 2400 x 1000) should be enough to create and edit A4-sized pages fairly comfortably, but again, YMMV.
Quark is a pain. The XPress software is not bad, but Quark as an organisation is a nightmare. They insist on huge fees for each upgrade and if you deal with any external bureaux or designers, they tend to have the latest version of the software. XPress v4 won't open files that were created or modified in XPress v5.0, so you have to upgrade, or nag the bureau to use save as... instead of just save. And then XPress 5.0 won't run on Mac OS X, and the internal spellcheck crashes the software. Flakey. Just too damn' flakey.
Adobe, by contrast with Quark, are paragons of virtue. The software is pricey, but not unreasonable and they have the ancillary applications more or less wrapped up with Illustrator, Photoshop and Acrobat. Buy all these in one package (Adobe Creative suite) and the cost ($899) is less than a single full version of Quark XPress ($1045). If I were starting out now, I'd have no hesitation specifying Adobe InDesign in preference to Quark, and it has been on my agenda for about three years to switch over to Adobe. However, the last time I was making a decision on make-up software, my corporate parent converted from their God-Awful proprietary system (XyWrite) to Xpress and this meant I delayed the decision to move across for another couple of years. This year I have to decide whether to upgrade everyone to Quark 7, or to convert to Indesign with all the disruption and training that will involve. All my instincts tell me to switch over, but I still have to consider compatibility with our corporate parent.
Macs are still better than PCs for page make-up, but only just. If a PC is good enough for everything else you do, then it will be good enough for DTP. Almost all decent laser printers now support PostScript, while font control on PCs is almost as good as on Macs. Then we get into Macs -v- PC territory, and it's not somewhere I want to go in this piece.
Some readers will disagree, but my personal feeling is that the interface on the Mac is a bit cleaner and the key strokes are more intuitive and more consistent across applications. Macs don't seem to get clogged up so quickly as PCs: I don't know why it should be, but even two-year-old Macs seem to work faster than brand new WIntel PCs when running programmes like Photoshop or XPress. If I have a choice I prefer to work on a Mac, but it's not such a big deal as it was a decade ago. At work I currently (2006) use a standard G5 with dual 1.8GHz processors and 1 GB of RAM, running OS 10.3.9. It was a mid- to low-range machine a couple of years ago, but I have had no problems at all with speed or memory. At home I use a recent, but low-end laptop running Windows XP. The only limitation on the laptop is screen space. Personally, I think that's a much bigger issue than the operating system or anything else.
Changes in attitude
We now look at the magazine as much more a collection of text, rather than a series of individual stories. This reflects the increased speed of production. In the old days, the production process would take six weeks or more, so that we'd be sending galleys away for one magazine while doing final clearing on the previous issue, and handling each story up to ten times at different stages of the process. Because each story was at a different stage of production, we would have to look at each story as an individual project. Nowadays, we store up all the text and then process it in one big batch. It takes around a week to convert all the text to style, make up the pages and do all the ancillary stuff like pictures and captions and headlines. We handle each story far less often, but when we handle stories, we tend to process large numbers of them in a single session. So we tend to spend most of the production cycle writing, with the final week split between make-up, proofing and writing last-minute stories. It is not a huge difference, but it somehow changes the way we deal with stories.
And the move to DTP has created a new class of worker: the Quark-jock, who knows Quark (or InDesign) inside out and back to front, and can drive the software fast and with great skill, yet has no sense, or interest in the words themselves. These people can lay out a magazine in no time, but have no interest in whether the text is Lorem Ipsum, or finely-crafted words of a skilled and knowledgeable journalist. The text is something to be handled, and processed. It is not a series of words, in which the carefully tuned meaning is defined by subtle nuances in word order and the grammar, but a commodity to be moved, chopped, extended and fitted to the structure of the page.
This is not necessarily a bad thing, just different from before. In the old days, we were all involved in the text, and we all helped lay the pages out and the production team was there to coordinate all the different stories and liaise with the service bureaux. That function has now evaporated, and the production team now takes text, images and advertisements and fits the different elements on to a page and makes sure the final result looks good.
Our task as journalists, is to ensure the text works and to make good cuts and additions, in close cooperation with the production team. On the flip side, our deadlines are greatly extended; we get to discuss the page; check the pictures; write the captions and headlines and overall have more control over how the page looks than we did before.
There is also a much more vigorous attitude to getting the stories out on the web ahead of everyone else. The good side of this is that we write a lot more stories, and we have a bank of proofed, finished stories ready and waiting when production week comes around. The bad side is that in the drive to get the who-what-where-when stories out there, we sometimes miss out on the analytical questions, such as why, how and what will it mean for the future.
Costs and manpower
There is, of course, absolutely no comparison between DTP and traditional page make-up operations. The most basic issue is cost, but beyond that there is the time from text to final page. Beyond that, the in-house team has control over every aspect of the production process. There is no possibility of an outside bureau introducing mistakes to the text. Cheaper, faster, better. The ideal solution.
Then there is the flexibility. If you have a couple of lines of white space at the end of a column, you have many more options. Either you can do the traditional thing and add a dozen or so words to fill the space, or you can tweak the size of an image to make the text re-flow and fit the space, or you can increase the vertical spacing of the text (leading) by some tiny fraction of a point or you can increase the space between one letter and the next (kerning) or you can increase the horizontal scale of the characters, or increase the run-around on an image or a margin. Even if the font purists baulk at such suggestions, very few readers will notice a percentage point change in these parameters. Having a variety options available makes the process faster and easier. Usually we still do the traditional thing, adding and subtracting words, but we can use the other options when we are especially pushed for time.
DTP has eliminated all the costs of external bureaux that used to do our scanning and typesetting. And even bringing all that work in-house, we have reduced the workload on ourselves. I used to estimate that one person could produce 10 000 words per month of high quality copy, assuming a reasonable production workload as well. I now think 15 000 - 20 000 words is a better estimate. We can produce twice the amount of text, while doing more tasks. Productivity has increased by a factor of three or four. You may be sure that pay has not increased by the same factor.
Costs have also decreased substantially. You can now set up a magazine from your bedroom. And indeed, the rise of the Blogosphere pays testament to the ease and low cost of entry to the modern world of publishing.
Schools and amateur clubs can now easily produce flysheets up to the standards of publishing professionals, and the costs continue to fall. All that remains is to get the design right. Unfortunately, not everyone with access to cheap DTP capability is capable of designing an attractive, readable page. But that is a whole different subject -- see Wharfinger's rant above.
Sources and further reading