Copy (of which few surviving examples exist for the period in question) could be of various kinds: written by the author himself (autograph) or by a copyist (apograph), or printed. These can be recognised by the presence of ink stains and fingerprints and/or format marks: scores, slashes or letter-number combinations made by the compositor to indicate where he had to continue on a new page. Printed copy was used for the composition of reprints. In an unmodified reprint in particular, a compositor tried to follow his model as faithfully as possible resulting in a line-by-line reprint. Producing illustrated books was more complicated: etchings and engravings had to be printed on an engraving press, often in an altogether different workshop, as only very few printers had an engraving press alongside their printing press. Obviously, specialized printmakers (etchers, engravers) had to be involved too; they worked from models drawn or painted by artists.
Before the typesetter could start working, the format and type face had to be determined. For some genres, the layout was more or less set: Dutch travel books, for instance, were almost always printed in black letter in quarto, while plays translated from the French were usually published in duodecimo with Roman type. Why some printers chose different formats, such as quarto-in-eights or octavo-in-fours, is often unclear, although it has repercussions for typesetting, layout and binding. The typesetter had a great deal of freedom when it came to his copy. He could, perhaps, partially determine the spelling and use abbreviations as he saw fit. He was the one to place catchwords at the end of a page. The question remains to what extent he was free in the composition of running titles, whether or not to use pagination and the method of signature marking. It is also open to question whether the compositor himself redistributed his composition after use into the type-cases. If that was not done properly, the result could be printer's errors. The distribution of used type was necessary from time to time, especially in smaller printing houses, in order to have sufficient type in the cases for the composition of the next formes. In exceptional cases, standing type could be kept for books that seemed to require new impressions (schoolbooks, bibles). Reuse of composition occurred in the printing of a text in two formats: after printing the folio sheet in two columns, the composition, made up differently, could be used for an octavo.
Seriatim composition was normally used, certainly in first editions, even though more type was needed than in composition per forme. An accurate preliminary calculation of the copy was necessary for composition per forme. When copy was distributed to a number of compositors or printing offices, occasional accidents happened which had to be repaired somehow or other.
In a well-organised workshop the work of the compositor harmonised with that of the printer which implied that a compositor had to be able to make enough formes per day for the printer not to have to wait. A number of views about this can be found in the literature. An experienced compositor could, for example, compose two folio formes in a day including notes. This would, however, also mean that when preparing a duodecimo (which had more text than a folio) in a smaller print-run, the compositor could not keep up with the press. We may suppose that in that case more than one compositor worked on the same project.
Depending on the size of a workshop, all kinds of apprentices may have made themselves useful. The pages secured by the compositor - possibly wrapped in paper and marked with chalk - stood in a particular order under the rack and could be taken away in the right sequence by someone imposing the formes.
Layout errors were certainly made but were probably discovered and corrected in an early stage (proof print or printer's proof). They are not often found in books.
The actual printing was done by two people: one used the ink balls to ink the composition, the other operated the press. According to literature, they could together print 1500 copies in a day of a single sheet (first impression and backing up). That the speed was often high is shown by the errors which were not noticed in their haste such as the folding of a section of the sheet on the tympan. If the folded corner was not too large, there was no loss of text, at least as long as the binder did not straighten the corner before cutting the textblock.
Other printer's errors were not so noticeable during the printing process, such as broken type or the accidental removal of a letter from the composition by the ink ball. Erroneously placed engravings in the text appear more often. Obviously, a printed sheet with the wrong image is not discarted, but the right engraving is glued in its place. The exact procedures for producing illustrated books with the engraving press is not clear, even though contemporary manuals still exist. There were some experiments involving printing in colour in the fifteenth century, but colour printing using copperplates really got started in the last decades of the seventeenth century. Apart from that, there was a whole industry of colourists or illuminators who manually coloured book illustrations and maps.
