Morpho butterfly in a rainforest…
(taken by Wikipedia user Asturnut)
Like me, I expect you have wondered why a modern commercially published book that is to all appearances superbly produced can neverthless have typographical garbage and weird other phenomena in it, or why odd entries in its Index are consistently a page out.
One of the forms of garbage that has always intrigued me most is foreign fonts — letters with accents or diacritics etc — that MUST have been correct in the final proof seen by the author, but have neverthless come out in the printing as percentage signs or something. How could this and other weirdnesses be the fault of the author, I reasoned to myself; they must have crept in at the printing stage. Yet this does not make much sense either, since what the printer receives from the typesetter is a PDF file corrected by the author and typesetter, and how could the printer’s software alter a PDF file? (In fairness, two computer-literate people have suggested to me that it can.)
Well, as a result of Sam1 and Sam2’s experience of seeing George Calderon: Edwardian Genius through the press, I think I can say that we have got to the bottom of it. The phenomenon is the fault neither of the author nor the printer. It is the fault of the complexity of modern computer typesetting software itself.
Take the following example. From the first typescript input by the author in Libre Office in about 2013 through three sets of proofs up to and including the fourth, which was in PDF, two book titles in chapter 8 of my book have always been italicised. But, for goodness’ sake, in the published book they are in roman! How? Why? A quick check with the hard copy proof 4 shows that they were still in italics then, and that was the form in which the text was submitted to the printers…
Except that, unbeknown to us, it wasn’t. When we received the final, fifth, proof from the printers, we were checking for typos, changes in exotic fonts, and in particular top and bottom justification of the spreads of pages (this was because we had noticed in another book printed by Clays, Ruth Scurr’s John Aubrey: My Own Life, that the bottoms of the printed text on right and left pages were rarely level). Two days was not enough to proofread the whole thing word by word, and we certainly did not think that, for example, italics that had been there for five years could have been dropped between proof 4 and proof 5. But they had been dropped in that interval.
Months before, we had noticed whilst we were editing the typeset version that strange effects could appear from nowhere, for example sudden single-line spacing, or a break in right justification preceding an image (see, unfortunately, the last line on page 334). These phenomena seemed random and incomprehensible, but at least they stood out on the page. They were easy to spot. It did not occur to me, at least, that other, less conspicuous phenomena were popping up at random elsewhere.
At this point, it is worth realising that ‘random and incomprehensible’ phenomena have always occurred in printing. Here, for example, is George complaining to the publisher Grant Richards in 1912 that the word ‘name’ at one point in his Two Plays by Anton Tchekhof has been printed as ‘game’, despite the fact that it was accurately printed as ‘name’ in the final proof sent to Richards:
On referring to my proofs I find that it was correctly printed in everything that passed me, and only when it had gone forth with the final blessing, ‘bon à tirer’, did some freakish Puckish mischievous intelligence play this absurd prank with my text, and make Tchekhof talk nonsense.
For heaven’s sake tell me by what rites this impish spirit is to be propitiated in future, or by what prophylactic charms disarmed, tell me where the creature resides, that I may avoid his habitation. One hardly feels safe, if such things can occur, perhaps even under your very nose. I dare not read the rest of the book lest I find more instances of such sacriligious sabotage.
Of course, there was a reason for the wrong character in George’s case, whether it was an anarchic compositor getting his fix or the force of gravity causing the right character to drop out, but still the problem was genuinely random. I fail to see, however, that a change that one hasn’t noticed in a computer typeset text can be random. It seems to me that the text is a deterministic digital system of such complexity that it meets the chaos theory requirement of being ‘highly sensitive to initial conditions’. You flutter your space bar on one page and it produces a cyclone of reformatting on another one without you knowing.
In fact, on page 237 of proof 4 we had to move up the inset quotation with which it starts, because another glitch earlier on had dropped the top of the text too far beneath the running head (PERFORMING FOR THE TLS)… Although there is nothing that I would describe as a logical connection between doing this and losing the italics from book titles in the next paragraph, Sam2 assures me that there is a computer reason, a causal link in a chain reaction. The italics were automatically lost when we moved the inset quotation up, but we didn’t notice that. One can easily see, then, how the bottom line of an indexed page could jump to the next page unobserved.
…and glitches on the printed page
But it is my very great pleasure to hand over now to Sam2 to discuss features of our experience of typesetting the book, in two guest posts. Being a typesetter, computer programmer and mathematician, he knows what he is talking about in this subject-area!
Related
Frontis piece by Frank Calderon
Jean Ryckaert, May 1916.
A happy memory of Trinity College: undergraduates breakfast together in Downy V. Green, p. 77.
George and Kittie Calderon 1905
Christmas 1901 Cartoon by George Calderon
George Calderon, 1912
Michel and Vera Fokine in “Scheherazade”, 1914 (The Music And Theatre Library of Sweden, creative commons)
Colonel Sir Walter Coote Hedley c. 1920
Monarch Butterfly Madeira
Captain of the 11
Book sent to George Calderon in Russia by his mother, Christmas 1895.
The entrance to Fort Brockhurst (picture from Britain’s Past)
Sir Richard Vincent Sutton c. 1913
Figure from Gauguin’s 1899 painting ‘Tahitian Women with Mango Blossoms’ (left); George Calderon’s pencil sketch ‘Manu’ (right)
Basic form of the Calderon coat of arms
Unpublished watercolour by George Calderon, Tahiti 1906
K.N. Das Gupta and Margaret Mitchell in an Indian Dramatic Society production of The Maharani of Arakan, 1913?
Patrick Miles in his summer house, 11 December 2016
George Calderon c.1901
Does computer typesetting produce a ‘chaotic system’?
Morpho butterfly in a rainforest…
(taken by Wikipedia user Asturnut)
Like me, I expect you have wondered why a modern commercially published book that is to all appearances superbly produced can neverthless have typographical garbage and weird other phenomena in it, or why odd entries in its Index are consistently a page out.
One of the forms of garbage that has always intrigued me most is foreign fonts — letters with accents or diacritics etc — that MUST have been correct in the final proof seen by the author, but have neverthless come out in the printing as percentage signs or something. How could this and other weirdnesses be the fault of the author, I reasoned to myself; they must have crept in at the printing stage. Yet this does not make much sense either, since what the printer receives from the typesetter is a PDF file corrected by the author and typesetter, and how could the printer’s software alter a PDF file? (In fairness, two computer-literate people have suggested to me that it can.)
Well, as a result of Sam1 and Sam2’s experience of seeing George Calderon: Edwardian Genius through the press, I think I can say that we have got to the bottom of it. The phenomenon is the fault neither of the author nor the printer. It is the fault of the complexity of modern computer typesetting software itself.
Take the following example. From the first typescript input by the author in Libre Office in about 2013 through three sets of proofs up to and including the fourth, which was in PDF, two book titles in chapter 8 of my book have always been italicised. But, for goodness’ sake, in the published book they are in roman! How? Why? A quick check with the hard copy proof 4 shows that they were still in italics then, and that was the form in which the text was submitted to the printers…
Except that, unbeknown to us, it wasn’t. When we received the final, fifth, proof from the printers, we were checking for typos, changes in exotic fonts, and in particular top and bottom justification of the spreads of pages (this was because we had noticed in another book printed by Clays, Ruth Scurr’s John Aubrey: My Own Life, that the bottoms of the printed text on right and left pages were rarely level). Two days was not enough to proofread the whole thing word by word, and we certainly did not think that, for example, italics that had been there for five years could have been dropped between proof 4 and proof 5. But they had been dropped in that interval.
Months before, we had noticed whilst we were editing the typeset version that strange effects could appear from nowhere, for example sudden single-line spacing, or a break in right justification preceding an image (see, unfortunately, the last line on page 334). These phenomena seemed random and incomprehensible, but at least they stood out on the page. They were easy to spot. It did not occur to me, at least, that other, less conspicuous phenomena were popping up at random elsewhere.
At this point, it is worth realising that ‘random and incomprehensible’ phenomena have always occurred in printing. Here, for example, is George complaining to the publisher Grant Richards in 1912 that the word ‘name’ at one point in his Two Plays by Anton Tchekhof has been printed as ‘game’, despite the fact that it was accurately printed as ‘name’ in the final proof sent to Richards:
Of course, there was a reason for the wrong character in George’s case, whether it was an anarchic compositor getting his fix or the force of gravity causing the right character to drop out, but still the problem was genuinely random. I fail to see, however, that a change that one hasn’t noticed in a computer typeset text can be random. It seems to me that the text is a deterministic digital system of such complexity that it meets the chaos theory requirement of being ‘highly sensitive to initial conditions’. You flutter your space bar on one page and it produces a cyclone of reformatting on another one without you knowing.
In fact, on page 237 of proof 4 we had to move up the inset quotation with which it starts, because another glitch earlier on had dropped the top of the text too far beneath the running head (PERFORMING FOR THE TLS)… Although there is nothing that I would describe as a logical connection between doing this and losing the italics from book titles in the next paragraph, Sam2 assures me that there is a computer reason, a causal link in a chain reaction. The italics were automatically lost when we moved the inset quotation up, but we didn’t notice that. One can easily see, then, how the bottom line of an indexed page could jump to the next page unobserved.
…and glitches on the printed page
But it is my very great pleasure to hand over now to Sam2 to discuss features of our experience of typesetting the book, in two guest posts. Being a typesetter, computer programmer and mathematician, he knows what he is talking about in this subject-area!
Share this:
Related