In book design, the combined inner margins of the recto and verso pages should add up to the outer margin of a single page (of course there are different designs). When I open a book, I see:
left margin | left page text | center margin | right page text | right margin
left margin is equal to
right margin and should also be equal
center margin (which uses space of both pages).
If I were to place the single pages smoothly next to each other on a table, I could achieve a correct relationship between the outer and inner margins by using half the value of the outer margins for the inner margins.
When a book is bound, its pages are connected using one of several binding techniques (glue binding, stitchbinding, spiral binding…). As a result of this process, a part of the pages is curved or disappears in the binding and is therefore lost as usable space.
To take this into account, a binding correction is applied. It reserves space (millimeters up to centimeters) on the inner margins on book pages. This space is subtracted (and therefore: excluded) from the dimensions of the paper which are used to calculate the type area and the margins.
Single-sided (but multiple-page) documents may be different in that respect, as only one printed page is visible at a time. For that reason it is argued (article »Ausgleich des Bindeverlusts bei Büchern« by Markus Kohm on p. 31ff, specifically p. 37f; in German) that single-sided documents do not need binding correction, as the reader does not look at a double page (the article discusses loose-leaf binders, ring binders and application portfolios: all documents, that do not typically connect the pages in their binding). It further suggests that with these types of documents, the margin should be adapted to reflect the punch holes.
The problem I see with this is that the margin makes part of the paper dimensions that are used to calculate the type area, whereas a binding correction is not. However, there are other forms of single-sided documents to which paper loss due to binding technique definitely applies: e.g. papers in academic journals and theses.
Should binding correction be applied to single-sided documents which are subject to a rigid binding method (like a book)?
This could be asked at the LaTeX stack exchange (like this question asking for specific binding correction values), but since my question is about document design and not about its implementation in LaTeX, I ask it here.