If it’s not a codex why would you treat it as such? If you’re setting up single pages, your margins should (conventionally) be symmetrical (eg. L&R=1.5 : T=2 : B=3).

As for the text measure, the obvious answer is up the type size to 12, but if you’re anything like me 12pt type makes you feel ill. Aside from that, remember that Bringhurst’s suggestion is the “optimum” line length, and your format doesn’t particularly suit that “optimum,” therefore you’ll likely end up extending it another few picas.

The other option, as you mention, is to shift the content as though you were, indeed, working with a recto page and use the right margin for side-notes. This seems like a lot of work, however, for a downloadable, but if you have the time and inclination, you certainly have the page size to accomplish the task.

If you were working with a document that you felt would be mainly read on screen the landscape idea is a better solution, but as you say it would be printed and bound (presumably at the printer’s choice of binding), you would need a generous top margin (the binding edge).
For a portrait format you should probably budget for at least a 30mm inside margin to accommodate any form of binding, combined with your text measure, would give you a good 50mm column for footnotes, instead of having them aligned on the bottom, you could also align them with the reference in the text (although that becomes tricky when you have many footnotes to fit in).
Tim

You could try using two columns. A lot of medical journals are set in this layout. You can check www.nejm.org (New England Journal of Medicine) and download their PDF journals. (the journals are laid out in U.S. letter size, slightly wider than A4)

Although in order to have a great 2 column layout, you may have to reduce the measure to maybe 18 picas and type size to 10pt. The reduced type size may pose a problem though if majority of your target audience use inkjet printers.

Hope this helps

Joseph

1. Adobe Garamond looks good at 12 pt as it has a small x height.

2. If you are using proper page layout software a good solution is to have a large left margin into which you put sub-titles, large. Eg you can have them in regular italic Adobe Garamond 24 pt.

3. Another solution is to keep the portrait orientation but make a magazine style 3 column lay out. This will probably save you pages also, if that is any factor.

jason: As for the text measure, the obvious answer is up the type size to 12, but if you’re anything like me 12pt type makes you feel ill.
William Berkson: Adobe Garamond looks good at 12 pt as it has a small x height.

12 pt makes me feel somewhat retarded, but in a single column with a lot of whitespace it might be acceptable.

jason:… remember that Bringhurst’s suggestion is the “optimum” line length, and your format doesn’t particularly suit that “optimum,” therefore you’ll likely end up extending it another few picas.

Agreed, but my goal is to, although working with the A4 format*, stick as close as possible to the principles/guidelines of good typography.
* For Americans: close to your letter format.

jason: This seems like a lot of work, however, for a downloadable …
William Berkson: If you are using proper page layout software …

I’m using Indesign CS2. It is indeed a lot of work, but it might be a good opportunity to learn more about InDesign scripting. Can something like moving all footnotes in a document (originally imported from a Microsoft Word file) to a separate textframe (on the side of the page) be done with scripting?

timd: If you were working with a document that you felt would be mainly read on screen the landscape idea is a better solution …

I’m afraid of the landscape orientation because portrait seems to be the default on most printers. Having people print landscape PDFs seems to be asking for trouble, i.e. text falling of the page etc. In Acrobat Reader in Windows, ’auto rotate and center’ is not always on, if I recall correctly. I’m not sure what the situation is with Acrobat and Preview on the Mac.

timd: footnotes, instead of having them aligned on the bottom, you could also align them with the reference in the text (although that becomes tricky when you have many footnotes to fit in).

I’ve tried ’anchoring’ the sidenotes to the note reference symbols in the text, but that didn’t work for all notes. Would it be acceptable if sometimes the note is not right next to the reference but slightly lower on the page to make up for a longer footnote (or larger number of footnotes) above?

juki: The reduced type size may pose a problem though if majority of your target audience use inkjet printers.

Good point, which I hadn’t yet considered, really. Maybe my font choice is not optimal to begin with? Garamond Premier Pro (silly name, by the way) with its rather high contrast isn’t particularly suitable for bad printing conditions, is it? The reason I chose Garamond because of its Greek: I do need a good and complete polytonic/ancient Greek character set. Of course, Minion Pro has that too. Would that work better?

juki: You could try using two columns. A lot of medical journals are set in this layout. You can check www.nejm.org (New England Journal of Medicine) …
William Berkson: Another solution is to keep the portrait orientation but make a magazine style 3 column lay out.

juki, I’ve been able to find only three column layouts there. The papers I’m designing are on subject matter in the humanities, where you do find multiple (mostly two) column layouts, but not that often. I think setting the papers in multiple columns would convey too strong a ’dictionary’ or ’encylopaedic’ feeling. Even if the columns remind the reader of scientific journals, that’s not really what I want to achieve. The papers, most of which have been published in journals, should be presented here in a more personal and individual way.

Would it be acceptable if sometimes the note is not right next to the reference but slightly lower on the page to make up for a longer footnote (or larger number of footnotes) above?
Yes, although there might come a point when the footnote is completely bumped to the next page, which will be confusing, that really is only a decision you can make based on your knowledge of the frequency and length of footnotes. As for the portrait/landscape problem, if your readers are likely to make that kind of mistake then you should go for the default setting, there might be some document property you can assign that will allow you to specify the printing set-up.
Tim

timd: Yes …

But even if the sidenotes are moved slightly down while still sitting on the same page as their references, the reader’s eye has to scan a larger area (namely the side column) to find the note he’s looking for. If the notes were on the bottom of the page, where they would have more a more expansive line length, they would need less lines, i.e. the note numbers are closer together.

Tschichold says (Ausgewählte Aufsätze über Fragen der Gestalt des Buches und der Typographie, p. 148): “Fußnoten sind die späteste und höchstentwickelte Form der Noten. Marginalnoten (…) sind eine längst veraltete Form.” (My translation: “Footnotes are the oldest and most developed form of notes. Marginal notes are a form that has since long become obsolete.”)

For some contrast, see Bringhurst Elements, p. 68: “Sidenotes give more life and variety to the page and are the easiest of all to find and read.”

Apart from those few statements I’ve not found a systematic account of sidenotes and these being put to (good and bad) use. Does anyone know of such a work?

The trouble with footnotes in this case are that you are trying to make better use of a page, so rather than take up space at the bottom of the page that could be used for text and leaving large unused areas, sidenotes will use otherwise unused space; when the reader has located a sidenote once they will surely not have any trouble finding the next one, they will not be difficult to spot after all. Footnotes present the problem that the optimum measure is often a lot shorter, given the point size is smaller, and you have to consider whether you should stick to the text measure or use the optimum measure, equally sidenotes containing long urls will also present problems in linebreaks or even having to break the addresses over more than one line. The best way to make this decision is to look at your provided copy and see if the footnotes might present either problem, if they are really heavy you might consider doing them as endnotes (either end of chapter or end of document) but they are less user-friendly. As a personal preference and with all other things considered I prefer a sidenote to a footnote.
Tim

One other way to deal with a wide page/narrow measure is to have a significant left indent. In other words, the headings might start as much as six picas to the left of where the text starts. This should work well as long as you have one or two headings per page, and for academic work that shouldn’t be a problem.