"Additionally, I don’t want a visual element of the project to be something I didn’t create myself."
Like many other approaches to font design it falls for what I like to call the "Metafont Fallacy".
When we approach the shapes of a typeface it is very natural to think of them as created by a writing implement like a brush or a pen. That is how characters were reified for the better part of 5000 years after all. Even the Romans used pen and ink despite being better known for carving into stone.
It is no surprise that Knuth designed Metafont as simulation of pen and brush.
It turned out, however, to be completely impractical to produce a typeface that looks similar to what we are used to. Similarity is crucial for legibility.
The reason for this is, that the stroke width of any contemporary typeface varies in a much more complex way than can be produced by an oblique pen. There are at least three causes for that:
- Printed letterforms have never been produced by a pen or brush
- Letterforms include sophisticated optical corrections
- Letterforms account for the flow of ink in the printing process, also known as ink traps
This is why most real Metafont fonts do not use all the polished pen and brush machinery at all, but use a very thin pen to draw the complex outline and then fill it.
Modern letter shapes are just like that and nothing that pen and brush strokes could directly reproduce.
As a sometimes semi-professional calligrapher, I can also add that the mark of good calligraphy is also to move beyond mere motions of an oblique pen held at a constant angle. Really beautiful and pleasing shapes demand precise control of angle, tilt, and ink flow all at once.
Great comment about font design in general, although in this particular article it seems like it's not fallacious at all. They seem like they're explicitly going for a handwritten effect, and are willing to sacrifice legibility to get it. The images at the bottom of the page achieve a hand-made look, and having the letters look like they were drawn with a brush or pen -- including all the flaws that come with that -- is part of the (quite impressive, IMO) effect.
An innovative and refreshing take on font design! Almost all other articles that I've come across on font design, are written by graphic designers who are at least moderately 1337 typographers, and who are pretty anal about things like kerning, and who come across as being fairly intolerant of non-designers.
Nice to see an example of how us mere mortal engineers can "design" a font, using a few rough dots, and a whole lot of arguably hacky (but also arguably quite clever) scripts. No doubt there are those who would cringe at the results; personally, I applaud them.
Ha yeah I am actually a designer, although I'm very UX focused rather than UI - and I'm definitely not a typographer. I am cringing a bit at the kerning at the moment, it's definitely still a work in progress! I guess though that my goal was to recreate handwriting, rather than to create a font. My handwriting doesn't have great kerning either.
I made something similar for OpenScad, back before it added support for text. It was like a really simple Hershey font, but using only line segments and circular arcs (no ellipses), because that was the path of least resistance.
I think the most important thing you can do, if creating a handwriting font, is simply to create multiple versions and come up with a way to pick glyphs from them at random when composing text with it. If you don't do that, your font will never, ever make it out of the uncanny valley, no matter how much work you put into getting the details right.
I get so much junk mail that tries to fool me into thinking it's addressed to me personally with script-style fonts. Some of the fonts are very nicely done, but all I have to do to dispel the illusion is to compare two instances of the same glyph and observe that they are identical.
I think if you look at the example uses of this font, it's exactly what the creator intended. They weren't going for a 'true' handwriting font, merely a generated one that looks attractive. I think the end result is a lot nicer than a classic cursive font.
I never fully automated, but python paths->svg for strokes followed by import into FontForge for width and export to ttf worked nicely for my conscription efforts
I like that idea and the way it was implemented. It reduces a font to its design and then executes the design, an excellent compression technique. I think there are plenty more applications to that idea. For example, in stead of transferring bitmaps showing pictures pixel by pixel, describe the picture using vector graphics (coordinates) or even a more basic description of those vectors.
Somewhat related but this can go pretty low in amount of code (and data) with a bit of design, here is some experiments and details on the generation of monospace blocky font with very low amount of data (data is encoded into bitfield) and how to style it through simple operators or algorithms such as diffusion curves, the aim was slightly different as it focus on using smallest amount of CPU instructions : https://www.onirom.fr/wiki/blog/25-09-2022_tiny_bitfield_bas...
The first render, where the stroke width is thin at the top, and fat at the bottom... it's still bad objectively, but it almost looks like it is heading in the direction of something that could be called interesting/creative. If the author continues, it might have something really cool in a few iterations (mixed in with more than a few that make the eyes bleed, I'm sure). Wouldn't be shocked if he eventually shakes a few insights down from the clever tree.
Perhaps you are being downvoted because you're leaving negative yet unconstructive feedback. Apparently what author created is not creative and objectively bad? Please reflect on what you feel when you receive unsubstantiated negative feedback of your own (the downvotes).
Nothing negative about it. They're objectively bad. If I was insulting artistic talents, if he had done them himself, then I could see hurt feelings. These were done by algorithm. The author didn't create them. He wrote code that did so. As I already said, if he continues, some insight might shake out.
You're all poorer for seeing negativity where none exists. You don't even realize you're doing it, do you?
"Additionally, I don’t want a visual element of the project to be something I didn’t create myself."
Like many other approaches to font design it falls for what I like to call the "Metafont Fallacy".
When we approach the shapes of a typeface it is very natural to think of them as created by a writing implement like a brush or a pen. That is how characters were reified for the better part of 5000 years after all. Even the Romans used pen and ink despite being better known for carving into stone.
It is no surprise that Knuth designed Metafont as simulation of pen and brush.
It turned out, however, to be completely impractical to produce a typeface that looks similar to what we are used to. Similarity is crucial for legibility.
The reason for this is, that the stroke width of any contemporary typeface varies in a much more complex way than can be produced by an oblique pen. There are at least three causes for that:
- Printed letterforms have never been produced by a pen or brush
- Letterforms include sophisticated optical corrections
- Letterforms account for the flow of ink in the printing process, also known as ink traps
This is why most real Metafont fonts do not use all the polished pen and brush machinery at all, but use a very thin pen to draw the complex outline and then fill it.
Modern letter shapes are just like that and nothing that pen and brush strokes could directly reproduce.