Pipe Minimal

typeface-pipe-minimal-demo

I’m calling the typeface Pipe Minimal, at least for now; easier than calling it “the typeface”.  It’s aggressively minimalistic in terms of the core design elements, and those elements could be commodity bits of piping, just a bucket full of straight bits and an elbow joints in PVC or copper.

The above demo plate is what I put together as soon as I’d crossed the threshold of having an actual usable font file; I managed to sort out a few problems that only came up once I tried to get FontForge to export, and got a set of lowercase characters into an .otf file and Photoshop was happy to use it.

But when I went back to FontForge to start working out upper case glyphs, I found myself frustrated by a couple things:

1. Really bogged-down performance when I tried to add a couple more guideline arcs and lines to the guide layer, the backdrop that I’m using to to create consistent shape/size in my glyphs.  I wanted a few more guidelines but more was (honestly pretty surprisingly) killing the program’s responsiveness to the point of being unusable.

2. I wasn’t even feeling confident about where, exactly, to lay those new guidelines out.  There seemed to be a few different logical options that contradicted each other a bit, and it became clear to me that I needed to think through the problem.

So I took a break and scribbled on graph paper and thought a bit about my guidelines situation.

typography-layout-diagram
circles within circles within circles

The core of it is that I designed the typeface as a a collection of razor-thin lines originally, a skeleton vs. meat-on-the-bones thick-walled characters.  And those thin drafting lines fell on a grid nicely; thickening up the walls changes the situation in slightly complicated ways and makes a character that was e.g. twice as tall as it was wide no longer hew to quite the same proportions once it fattens up some.

And so I started over from scratch with a razor-thin line instead of laying out exterior and interior walls, and that has worked fairly well. It feels like something FontForge doesn’t quite like — the thin lines can basically disappear visually if they fall along a guideline (which mine do, religiously) and only closed shapes like the circular loop in many characters will show up in the miniature glyph preview/navigation window.  So it feels a bit like flying blind.  But it works.

I redid all the lowercase letters that had made it into the sample up top, and verified that I could take those skeletons and expand them into thick-walled characters, and with that sorted out proceeded with the rest of the characters I was aiming for: upper case letters, numerals, and basically all of the punctuation that shows up on a standard US keyboard.  Enough that there won’t be obvious conspicuous gaps when writing casually using the typeface.

pipe-minimal-weight-samples
Three different weights.

And once I’d gotten the full set done and one fatten-and-export step done, it was easy enough to do a couple additional passes of fattening and exporting.  So now I have demo versions of Pipe Minimal in three different weights: thin, regular, and bold.  They’re not finished — I’ve got a bunch of little things I need to futz with, and need to figure out kerning lookup tables and keep reading up on FontForge to figure out which other things I don’t know yet that I’m missing.  But it was a good day for getting past the intimidating impression that FontForge left on me at first glance yesterday.

As a kicker, here’s the initial graph paper sketchbook takes on a bunch of the new glyphs; a couple of those uppercase characters changed from the initial sketch, but I also fiddled with a couple different takes on numerals (ultimately I like having them be three units high, splitting the difference between the four-unit uppercase and primarily two-unit lowercase letters) and scrawled out four potential ampersands all of which I ended up throwing away in favor of a different variation on the uppercase E.

typography-glyph-sketches

Author: Josh Millard

I manage and help moderate the community website MetaFilter, where I go by "cortex"; in my spare time I get up to all sorts of creative nerdery on the internet and in Portland, Oregon.

4 thoughts on “Pipe Minimal”

  1. Nice work! Font glyphs are a bit special: they must be outlined shapes, not stroked lines. There is a way for FontForge to take strokes and convert them into paths with a user-defined pen, but I forget where it lives in FF’s deep menus.

    Other font weirdnesses include: must have nodes at extrema of paths, paths must not overlap, and node coordinates should really be integers. So many of the reasons are lost to ancient messy corporate spats between Adobe and Apple/Microsoft, but we still have to live with ’em.

  2. Font glyphs are a bit special: they must be outlined shapes, not stroked lines.

    Yeah, this was an interesting revelation when I first started in — I’d been designing the glyphs as lines in Inkscape and then adjusting the rendered stroke width and style to get the actual look of the letters right, and when I came over to FontForge I realized that wasn’t really gonna translate. So my first go was setting up a donuts-and-boxes template on the grid layer with a medium-ish width and outlining the glyphs from there. Worked okay to a point but seemed like width variation was gonna be a pain.

    There is a way for FontForge to take strokes and convert them into paths with a user-defined pen, but I forget where it lives in FF’s deep menus.

    Aye, turns out it’s “Element -> Expand Stroke”. That’s the approach I’ve been taking in this second go, once I established that it was workable. Which, the quicky weight variations I’ve put together all involved just apply the same Expand Stroke operation to the entire typeface, which gets the basics done pretty well. Some cleanup details after that still, some of which I think I can resolve ahead of time by modifying the underlying strokes. But in practice to get things juuuuust right I’m gonna have to probably take a little more care with the per-glyph details to get the Line Cap and Line Join variation how I want it.

    Other font weirdnesses include: must have nodes at extrema of paths, paths must not overlap, and node coordinates should really be integers. So many of the reasons are lost to ancient messy corporate spats between Adobe and Apple/Microsoft, but we still have to live with ’em.

    Yeah. I’m still figuring out how and to what extent I can mass fix those issues with a command prior to font generation, vs. getting a long list of error messages at generation time.

  3. I really like the square of negative space created between the c and the o in “.com”, as it appears in your header graphic. The use of the stencil face originally confused me, but now that I understand it I really like it.

    But you know what would be even more on brand? If the center of that negative space had its fractal heart cut out with (half?) a small black square.

Leave a Reply

Your email address will not be published. Required fields are marked *