Well hi Stewart - it’s been a long time! Hope you’re doing well and enjoying life in Canada.
We did make some beautiful books, didn’t we? I loved the challenges presented by yourselves at HarperCollins and all of the other publishing teams that we worked with.
The dictionary and thesaurus projects were a particular joy. If you are interested, here’s how the magic happened…
The data arrived as two files - the dictionary text and the thesaurus text. We extracted the headword lists for these and sorted the two files into alphabetic order, so that the dictionary headword was followed immediately by the thesaurus headword (if there was one - not all dictionary headwords had a corresponding thesaurus headword).
After reconstituting the data into one file the text was processes. Dictionary headwords went into one typesetting stream (diversion in troff parlance) and the thesaurus headwords went into the other. The remaining white space on the page was monitored and if either a dictionary entry or thesaurus entry exceeded the available space, it was split and the overflow moved to the next page. The page was then output, dictionary text first and split into columns, followed by the designed text that separated the dictionary entries from the thesaurus entries and then the thesaurus text split into columns. Finally, the page livery was output, including running heads, folios and any running foot text. The process started again for the next page, until the entire text was output.
Initially, when our typesetting system was running on Intel 80486 computers, we started a pagination run as we left the office on an evening, to find paginated text waiting for us the next morning (providing nothing went wrong overnight). As technology evolved and processors became faster, we could paginate in real-time. A 500 page book would take about ten minutes to half an hour. The last time I used our typesetting system (in 2010 approximately) it would paginate a 1000 page text in less than a minute. On today’s hardware, I reckon it would be virtually instantaneous.
And therein lay our problems - we made what we did look too easy and it was hard to justify (no pun intended!) the amount that we charged - which wasn’t a vast amount - with what we provided to the customer, which was usually a CD-ROM of PDF pages ready for the Computer-To-Plate process done by the printer. It came down to perceived value – a CD-ROM didn’t look as expensive or work-intensive as a stack of film, bromide or paper. We couldn’t actually reflect how much investment in equipment and resources we had made. For example, our first typesetter cost twice as much as my first house – you would be looking at around £400,000 in today’s money. Our first 28Mb (Megabyte!!) hard drive cost £20,000 second hand.
Still, happy days and exciting times. Everything needed to be invented and I learned a lot about software and hardware and how to bring them together, how to build systems that people could use and how to optimise processes to reduce cost and increase profitability.