Frustrated with fonts and formatting

Topics/tags: Rants, typography, marketing, long, rambly

Late in the spring, I spent some time updating my workflow for digital memos and letters to accommodate the College’s revised typeface choices and wordmark. I thought I had things the way I wanted them.

Then I decided to submit a sample to Communications [1]. In doing so, I realized that I should probably use Renner* rather than Josefin Sans as a substitute for Futura, even though I liked the latter font more. Renner* is closer to Futura.

Of course, Renner* had an a that I don’t like and didn’t have monospaced numbers. But it’s an open font, so I could feel comfortable changing it. And I did, creating a font that I called RennerAlt. Making a variant was less complicated than I thought it would be. FontForge was remarkably easy to use. I felt good.

Amazingly, Communications approved my design except that (a) they wanted me to make my left margin align to the left edge of the G in the wordmark, and (b) they wanted me to use ragged-right (aka left-justified) text. I appreciated the general approval, particularly because the letterhead had my personal Web site rather than the grinnell.edu that I refuse to use [2].

Then things started to go wrong.

Most importantly, html2ps would randomly crash on body text (in Besley*) with a G, but only on Linux and not on my Mac. (My Mac has a slightly older version of Ghostscript. Does that make a difference?). And it didn’t crash on every G, only some. Is that an issue with the design of Besley*? I don’t know. It may mean that I have to change the choice of typeface. It’s frustrating. And it’s one of those bugs that comes and goes.

What about the suggestions from Communications. I like fully justified text, so I’m ignoring the request to use ragged-right. But it seemed reasonable to try the wider left margin. Unfortunately, html2ps does not behave sensibly. First, I tried adding a margin-left entry to the body tag. That made no difference. Next, I added a margin-left entry to the paragraph tag. That made no difference. Just to make sure that I was getting the term right, I tried margin-left with blockquote. That worked. I can’t tell why it doesn’t work with the others. But I don’t want to use an extra blockquote around the body; that’s inelegant and incorrect. So I kept trying.

Eventually, I discovered that I could add margin-left to a magic @page style entry. That worked. But it did some strange things to the encapsulated PostScript file I included as it attempted to fit it into the new width. So I had to change that file, which I did mostly by experiment. I didn’t like 1.5-inch left margins [3] because they felt too large. I decided that I’d reconsider whether or not to follow that recommendation.

You think those would be enough problems. But they weren’t. While things looked great on screen, when I printed out the letterhead and showed it to a colleague, the first thing they asked was Why are there these white spots where bars cross; is there an xor of sorts? It appears that the printed Renner* has a problem in the locations where portions of the letter cross. It happens only in the printed copy, not onscreen [4]. It happens with both the original version and my alternate version.

I asked myself, Is it time to go back to the drawing board, or at least back to Josefin Sans? But summer hit, so my decision was simply to leave all of the problems until later.

Later finally arrived this fall. I had a new memo to write, and I decided that it was worth a bit of time to play with fonts and perhaps even with document spacing. Industrial Type has upgraded and renamed Renner*. It is now Jost*. There are many more variants available, including nine weights in the standard version and even more if one is using new OpenType technologies [5].

I was hopeful at first. Jost* supposedly includes an option for the simpler a that Futura uses. It also has the monospace numbers that I want, which it calls tabular numbers. However, for some reason that I cannot understand, my workflow did not work with Jost*. It may have been a naming issue. It may have been something else. But I kept having problems with one of the faces I used. Eventually, I decided to create versions with alternate names; for some unknown reason, that approach worked.

I like Jost* more than Renner*. I may be able to stick with it. If not, at some point, I’ll probably design my own alternative to Futura. I plan to call it Pastura.

What about that wider left margin? Let’s see what I can do about that. Currently, the left and right margins are 2.5cm [6]. That means that the text width is 8.5in - 5cm, or about 6.53in. If we switch the left margin to 1.5 inches and the right margin to 1 inch, we end up with a text width of 6 inches. I initially guessed that I would need to inversely scale the width of the bounding box for the Encapsulated Postscript. That bounding box is 536 units wide. My calculations suggest that I need a bounding box that is 583 units wide, or 47 additional units.

Nope. That scaled the wrong direction. Okay, let’s try a bounding box that is 492 wide (536 x 6/6.53), or 44 fewer units. Yup, that seems to get the sizes right. Now, what about placement? I know! I’ll temporarily add a vertical line that expands beyond the header and into the document. That will help me do the alignment. Wow! That was much more efficient than the last time I tried it.

Does it look right? Let’s compare it to the sample document that I received from Communications [7]. My simulation is close, but not perfect. But I now have enough tools to play with the margins and such that I think I could achieve something closer to their example. However, it appears that they are using a 9pt font and both the left and right margins are 1.5 inches [8]. As much as I’d like to be a team player, I’m not willing to use a 9pt font, nor am I willing to use a 1.5-inch right margin.

Hmmm … If I’m willing to make those compromises, can I make others? What if I shift the laurel leaf left a bit? As long as I’ve aligned the text with the left edge of the G in Grinnell, I should be fine. Maybe a left margin closer to 1.25 in, with the laurel leaf shifted appropriately, will work. Here’s the current state of my sample, including the vertical line I am using for alignment [9].

Here’s an example I got from Communications.

Are they close enough? The typefaces are different but similar enough to meet our goals. I am using a slightly heavier weight, medium rather than book. My first inclination was to say that’s okay. However, on second thought, I’m going to fix that. My left margin is a bit smaller, but not annoyingly so. My right margin is certainly smaller. However, that’s a decision I’ve made intentionally, as are the fully justified margins. I can tell the difference between the 9pt and 10pt typefaces, and I do prefer the 10pt. It looks like the red of the laurel leaf is slightly different. I’ve used { 0.854902 0.160784 0.109804 setrgbcolor }. Is that correct? What does our visual language page say? In standard RGB, Grinnell Red is (218,41,28). Yes, that translates to those percentages. Oh, that’s right; I grabbed the percentages from the SVG file: Of course they are correct.

It appears that I now have something I’m relatively happy with and that matches enough of the new stationary standards that it sufficiently meets the goal of a common look and feel [10]. I’ll take that as a success.

[1] I’d rather be straightforward about what I’m doing.

[2] When they let me put personal content, such as these musings, on grinnell.edu, I’ll reconsider that choice.

[3] The left edge of the G in the wordmark is approximately 1.5 inches from the left edge of the page.

[4] Now that I’m examining the problem a few months later, it appears to have happened in the onscreen version, too.

[5] I’m not. I also don’t need very many weights at present.

[6] I didn’t choose the 2.5cm margins; they seem to be the standard defaults in html2ps.

[7] I appreciate their willingness to share a sample letterhead with me. I do, however, wish that our official Style Book was ready.

[8] The size of the right margins are hard to tell in a ragged-right example.

[9] Yes, it will go away in the future.

[10] It meets those goals better than most of the official documents I’ve seen from the College in the past three months.

Version 1.0 of 2018-09-23.