Typesetting a technical thesis

athale's picture

Hi All,

This is my first post. I am writing my Mathematics Ph.D. thesis. I am planning to use Classical Garamond to typeset the thesis. Usually people use Computer Modern (that comes with LaTeX typesetting program) or Times New Roman. Garamond is artistic and that is my choice as I want a relaxed look for my thesis. Also, I own it so would not have to spend extra money. If it is the right choice, what font you would choose for headers, a sans serif font or bold weight of Garamond? Thank you very much in advance for your time and help.

Rahul.

Chris Rugen's picture

I agree with Tiffany. I'm not a big fan of Garamond's bold weight. However, I have more experience with Adobe Garamond than Classical.

Tiffany, since I don't know, do you think that Classical Garamond and Adobe Garamond SC can mesh?

Rahul, I applaud your desire to seek out something better. Back in college I used to use Centaur with non-default margins and line-spacing because TNR in Word settings was so painful. Professors would ask me how I made my papers so nice to read. The typography was sketchy, but even those adjustments made a big difference.

My experience with Garamond is that it can get tricky at smaller sizes because of its proportions and relatively fragile forms (as compared to TNR, which was designed to take abuse). This may not matter for what you're using it for, but be conscious of that, particularly if you're outputting on an inkjet or lower-res printer (which I was often forced to do).

Because I'm not a mathematician or a LaTeX user, I'd like to know what you're using for setting the equations. Are you going with Garamond, or a more specialized mathematics face?

hrant's picture

Nice that you're looking beyond your default choices. Garamond (which cut?) will work well as long as your point size isn't too small (unless you're using the obese ITC Garamond, in which case I can't help you :-). As for the headers, I'd recommend smallcaps too (you need a separate font though, don't use any automatic smallcaps method - they create discord) - not a bold weight of Garamond - those tend to look caricaturish. If you go for a sans, I've had good results pairing Garamond with Bank Gothic (a pretty ubiquitous font too), which might be especially good in your case since it's math stuff.

hhp

athale's picture

Thank you very much for the help and encouragement. I am using Adobe Garamond, and as I have to submit thesis on A4 paper I am thinking of using it at 12pt. I have not yet seen thesis with Small Caps headers, I will give it a try :-) If I go for sans serif font, will URW Grotesk mix with Garamond? I am asking without trying myself because to use a font with LaTeX I have to spend at least a day to set it up properly.

Chris I will be using Euler math font designed by Herman Zapf. Almost all LaTeX distributions have Computer Modern and Euler math fonts installed by default. As it takes a lot of effort to use any other font to typeset Mathematics using LaTeX, I will use Euler font.

Thank you once again.

Rahul.

athale's picture

Thank you very much for the help and encouragement. I am using Adobe Garamond, and as I have to submit thesis on A4 paper I am thinking of using it at 12pt. I have not yet seen thesis with Small Caps headers, I will give it a try :-) If I go for sans serif font, will URW Grotesk mix with Garamond? I am asking without trying myself because to use a font with LaTeX I have to spend at least a day to set it up properly.

Chris I will be using Euler math font designed by Herman Zapf. Almost all LaTeX distributions have Computer Modern and Euler math fonts installed by default. As it takes a lot of effort to use any other font to typeset Mathematics using LaTeX, I will use Euler font.

Thank you once again.

Rahul.

eomine's picture

Suggestion: make it 10pt and give the page wider margins (make the text column narrower).

hrant's picture

Usually 12 is too big, but I think 10 might be too small. So, uh, 11? :-)

If you're going to use Euler then try finding a sans to match it.
BTW, here's an interesting read: http://www.daidala.com/19aug2003.html

hhp

William Berkson's picture

I agree with the comments about the width of the line. I would recommend looking at Bringhurst's Elements of Typographic Style, the section "Choose a comfortable measure" and the accompanying table (in my edition, p. 26-7, 29). After reading Bringhurst, I found in experimenting that the measure is indeed very important to comfort level in reading continuous text. If you are close in width to a 66 character line (Bringhurst explains how to measure this using his table), it really is a big benefit.

On A4 paper, a usual full width line is way too wide for comfort. With the narrower measure, you can use the margins for titles and notes, if your type setting software permits.

The other thing that is critical is the leading. Adjust it till you find it comfortable to the eye. Because Adobe Garamond has a relatively small x height, you will need less leading.

If you get the measure and leading right, it will right away look way better, and the rest you can do is gravy.

Handling equations is a whole different can of worms, but I assume you have some materials on this.

Good luck!

athale's picture

Hi,

Thank you very much for your suggestions and sorry for late reply. I was going to use Adobe Garamond but then I found out that my Softmaker MegaFont CD had Classical Garamond with Old Style Figures and Small Caps, so I used that one. William, I used following texts
http://www.ctan.org/tex-archive/macros/latex/contrib/memoir/memman.pdf
http://www.ctan.org/tex-archive/macros/latex/contrib/memoir/memmanadd.pdf
to compute the desirable text width. The first link is the manual for memoir class for LaTeX and second one is addendum to the manual. The first part of the manual talks about typesetting basics.

I have uploaded the test file with few pages of text in pdf format to my web site.
http://www.geocities.com/rahul_athale/thesistest.pdf
The text is typeset in Classical Garamond and Mathematics is typeset in Euler and Computer Modern font, although the numbers are typeset using old style figures from text font. I have set the text at 11/13.6pt. The text width is 312pt and height is approx. 514pt. Spine and edge margins are 98pt and 158pt, respectively. Upper and lower margins are 130pt and 200pt, respectively. I use the ratio 1:1.618 to divide the margins after fixing the text width and text height, they also share the same ratio. Please have a look at the file and give your opinion. I have a question, do old style figures look good in chapter title and heading?

Thank you very much for all your encouragement, without which I would not have invested so much time in typesetting.

William Berkson's picture

I see that the text reproduces part of Bringhurst's table, with credits. This is convenient, but I don't know how I would feel if I were Bringhurst! Your third link, with your test file doesn't open for me; I get an error message from Geocities.

hrant's picture

The Bringhurst thing is probably "fair use".

But yeah, I couldn't open the file either.

hhp

athale's picture

Hi,

I am extremely sorry for the inconvenience. Now, it is not working for me either, but if I cut and paste it in the other window and open from there then it is working. I don't know why clicking on the link is not working, will try to find out tomorrow. Thank you very much.

Rahul.

eomine's picture

You can upload your PDF here, Raoul (use the 'upload attachment' button). I guess the problem is, Geocities doesn't allow 'hot linking' (I was able to download it though).

It looks good. And I don't think there is a problem in using oldstyle figures for your headings. But I'd make them all-smallcaps, instead of caps+smallcaps.

William Berkson's picture

I used your copy-paste method and saw it also. I think it will look better if you find a use for the wide outside margin. For example, you could set the head & sub head numbers in the margins.

Chris Rugen's picture

I agree about the margins. Also, track out the small caps a bit. It'll help their legibility.

athale's picture

Hi All,

Thank you again for the comments. Eduardo thanks for pointing out that I was using Caps and Small Caps in the page headers. The page headers were made automatically and so I had overlooked that. I only felt that it was odd but could not recognise the oddity. About the margins, now I have started page headers from the margin. I have uploaded the pdf file as attachment.


application/pdfThesis test pdf file
memtest.pdf (119.8 k)

But I did not understand what do you mean by setting head and subhead numbers in the margin. Does wide margin look bad? Should I switch to 12pt font? Thanks in advance.

Rahul.

hrant's picture

Very classy stuff.

My only complaint is that I think you're being a little bit too generous with the whitespace.

hhp

William Berkson's picture

Do have a look at Bringhurst's 'Elements' - there must be a copy in your University's library.

The outside margin is too big in comparison to the other margins. Bringhurst has a whole chapter on all the mathematical ratios that have been historically tried for page and text blocks, and their positioning on the page. Not rules, but fascinating, especially for a mathematian, I would think. For example, he shows you how to put a golden rectangle text block on an A4 page, with a lot of white space around it, which will work great for you.

Your running head with the rule defines the text block as wider. The outside margin is naked without something in it.

You can put something in it or make the rule shorter, and reposition the whole thing. I don't know the current size of your type, but you can certainly make it bigger (not much bigger though) if you like, as you will use the space better.

Bringhurst's book is an example of how to use margins. His way is traditional. It has been revived in many computer instruction books, which have wide pages with stuff in the wide margins. Basically you want an additional column outside the text block, with enough space between it and the text block to distinguish it, but not so much that it's not related. Then do the numbers or notes right aligned on the left page and left aligned on the right page, if you have a wide outside margin. You can just align the numbers with the head, sub head or first line of the equation.

puffinry's picture

I'm also writing a mathematical PhD thesis so this is a very interesting topic for me, although I'm in my first year now so I'm not actually writing up yet! Apart from the thorny issue of mathematical typesetting, one has to contend with restrictive and largely arbitrary layout regulations.

For example, my university mandates 12 pt text, with "double or 1.5 spacing" (which I suppoose means 12/18 at least). It must be printed recto only on A4 paper with all margins at least 15mm, and the left-hand (binding) margin 40mm or more.

For my master's thesis, I used 12 pt Charter and Euler with a measure of 29 picas (to make a "double square" text block, as Bringhurst would have it). I'm sure there must be a better solution than this. I've been planning to try and put something together using one of the Adobe OpenType fonts, which have most of the glyphs you need for equations (e.g. swash caps and Greek letters). The only missing glyphs (apart from "blackboard bold" and blackletter) are the geometric symbols representing mathematical operators, which can easily be taken (perhaps slightly modified) from elsewhere, since they need only match the base font in weight and size. (The proportions and style are pretty fixed!) Adam Lindsay has done some work in this direction.

For blackboard bold, one could just eviscerate the ordinary bold uppercase. You're unlikely to need more than two or three such letters anyway.

I'd be very interested in hearing anyone else's thoughts or experiences.

hrant's picture

> my university mandates 12 pt text

If that's too big (and it usually is) you can "cheat" by using a font that's small on the body - although that will exacerbate the extreme leading. BTW, do they [explicitly] forbid multiple columns?

Anyway, your solution (big lateral margins*) works great. The only two things that bother me are that Charter (a superb choice) doesn't seem to jive with the Palatino headings; and your footnotes are very cramped - the fact that you're needing a rule is telling - I think you can make the point size slightly smaller, but do track it looser as well, and if there's a semi weight use that.

* I'd make the left bigger than the right though.

BTW, that OpenType doc looks interesting.

hhp

athale's picture

Hi,

Robin, thanks for the link. Your master's thesis looks different and nice. William, I do not have Bringhurst's book and neither does the university library, even the nearest Amazon does not have it.


application/pdfThesis test pdf file
thesistest.pdf (120.9 k)


I am now using 12pt font, text width is 348pt and text height is 563pt. The left and right margin is 124pt, and upper and lower margin is 140pt approximately. Also I forgot to mention earlier, did anybody notice, table number is same as the page number? Donald Knuth et. al. have done the same in Concrete Mathematics. Thank you all for your help.

Rahul.

puffinry's picture

Hrant: Thanks for the suggestions. Would you think it best to keep the chapter headings in Charter too? I tried that, but I found Charter’s raster-friendly design rather distracting at large sizes. (The vertical trailing edge of the ball terminal on the lc ‘a’ bothered me particularly, for some reason.)

The regulations don’t explicitly forbid multiple columns. I briefly toyed with a two-column layout, but didn’t find one I liked within the constraints.

William Berkson's picture

Rahul, in Europe if you don't have another ready source, you can buy Bringhurst at Typotheque.

'Elements' is almost universally praised as the best single book on typography, so I think you will be happy to have it.

Currently, I don't think the relation of your text block, margins and page look so good. It is too close to the page shape, which makes it dull. I think your old version, putting the heading numbers in the margins would look better, or following Bringhurst's design for a golden rectangle on A4 paper (which is the ratio 1 to sq. root 2).

The formula is this: the top and spine margin is 1/9th of the width of the page, and the bottom and outside margins are twice this. This will give you a golden rectangle in the text block, nicely positioned on the page. You can then adjust the size of your text to make this work.

In general, Bringhurst's principle is "Shape the textblock so that it balances and contrasts with the overall shape of the page". Usually, as with the above formula, the inner and top margins are less than what you have in relation to the bottom and outside margins. This assymetry 'locks' together the facing, bound pages.

Some other comments: I think your rule at the top of the page is too thick. And your title page is too shy compared to your chapter titles. The classic thing would be to use letter spaced caps or caps & small caps, but that is only one idea.

Overall, Bringhurst will give you guidance on many, many issues.

hrant's picture

Yeah, you need "Elements".

I think Charter would certainly work better than Palatino. Don't necessarily let individual glyph details get to you - and remember that you're not really the "user" of this publication anyway.

hhp

athale's picture

Hi,

William, thank you very much for writing down the Bringhurst's principle and other comments. Actually it was written in the memoir manual, the link I posted in the second post. But I had made a terrible mistake in typesetting the last sample. I have a different problem though. Now if I follow the advice of spine margin being equal to width/9 then the type block becomes around 33pica, that means around 75 characters per line at 12pt Garamond. So I certainly need wider margins. Now the question is should I divide the edge to spine margin in the ratio two to one, or make spine margin equal to the page width/9 and keep the remaining space as the edge margin? Thank you very much for your time.

Thanks for the typotheque link, but I am looking for a paperback edition of Bringhurst's book.

Rahul.

William Berkson's picture

Your eye needs to make the decision. 75 characters is the outer limit for a comfortable read, according to Bringhurst, and as your text is broken up by equations, it may be fine at that length. As your line length increases, so should the leading, so you should also test out that variable with samples before making a decision. Another option is to increase the size of the text. I don't know what your version of Garamond looks like, but if it has a small x-height like Adobe Garamond, setting it at 13 to 14 point may be fine.

Typotheque is selling a paperback edition, according to the site.

athale's picture

I am extremely sorry for overlooking that it is indeed a paperback edition. I will experiment with the margins and will post the test file in a day or two. Thank you very much.

Rahul.

athale's picture

I have changed the margins and have moved the section numbers and headers in the outer margin. Regarding comment about the tile page, the university has a fixed title page format so anyway my title page will not be added in the final thesis. Please have a look. I am looking forward to your comments. Thank you very much for your time.


application/pdfThesis test PDF file
thesistest.pdf (121.0 k)


Rahul.

athale's picture

Can anyone please comment on the typesetting in the file I uploaded on Sunday? I would like to fix the style and finish the work. Thank you very much for all the help.

Rahul.

William Berkson's picture

Having been through the PhD process myself, permit me to state the obvious: the design doesn't matter that much to you, because you are going to get your PhD based on content judged by people who are very knowledgable about that content. The visual 'sell' isn't that important to you. It is a good way to learn about typography, but not that important to this particular product.

I don't think the current version works well. You have a number of elements that by themselves could work, but don't work together to make a coherent, balanced set of pages. The running head with the rule frames the page as having equal margins left and right. Then you have the text with outside margins over twice the inside margins, and with little use of them. The effect is unbalanced.

It might work with somewhat less of an outside margin and more of an inside margin, and the running head the same width as the text. But you have to try to see.

If you are going to put the numbers of the headings in the margins (as I had recommended as an option), I think will look better if they are right aligned on the left page and left aligned on the right page. Also I would try puting the equation numbers in the margins as well, if you are going to go that route.

The large bold numbers in the margins are distracting. They need to me more discrete (non bold or smaller size or both), or don't hang them in the margins. They need to be clearly subordinate if they are in the margins. Also I would check whether in labelling the heads lining (full height) numbers are better, especially if they hang in the margin.

The titles don't need to be both bold and big. The current large size and the way they are vertically spaced breaks the rhythm of the grid. Bringhurst recommends keeping everything in multiples of unit lines. That way lines on opposite pages line up with each other. Felici says you can also add half line gaps between paragraphs, and that way after two such they come back to the grid.

You have the luxury of so much space that I would stick to the grid. You can put heads in bold caps and sub heads lc bold in your normal size. Or, more classically, you can avoid bold altogether and use caps, small caps and italic as your options.

Have fun playing with this, and don't worry about it!

piticu's picture

Athale, i think will be nice to see the final version of your thesis.

Miss Tiffany's picture

I think a more elegant choice would be to use the Adobe Garamond Small Caps.

Miss Tiffany's picture

Oops. I must have glossed over when I read the word "classical". Well, I guess you might not have Adobe Garamond smallcaps then. :^/ But, I stand by my suggestion. Margins ... I agree with the wide margins. I am fond of 2.5" on the left and 1" on the remaining three. But, you might not have a choice. Some schools give students set guidelines, I know I always gave my students set guidlines. Yeah, I am mean, but when students realized they could use certain typefaces to fill the pages quicker, well I had to stop that! ;^D

I wouldn't go with a sans, because this will be deeply read anyway. I would think the reason to use a sans is to let someone skim through it. Well, I guess you can't keep people from skimming. Anyway, if you will be using Eular (if I'm understanding it correctly) that will serve as contrast enough where you probably won't want a sans at all.

I would agree with Hrant on the point size. 10 pt Garamond is probably going to cramp the eyes of the mathematician who has to grade this paper.

Syndicate content Syndicate content