Forget FontForge... how can I rethink this new media class assignment?

January 19, 2009 9:55 PM

NewMediaArtProjectFilter: Help me rework a digital art assignment involving font-making software and database-esque art.

I teach a new media class in a university visual arts program. I am planning several class assignments around forms which are often found in new media art, specifically the loop and the database. For the loop project, students are to create a looped animation using Photoshop or Flash. For the database assignment, I had planned to have students create a font, which can be thought of as having a database-like form (a non-hierarchical collection of discrete items which can be reconfigured in various ways). My definition of the database form in new media art is drawn largely from Lev Manovich's The Language of New Media.

Note that this is not a typography class and I am not teaching students to design proper typefaces; the focus is on creating visual art which has a database-like structure, and I think the font format would be a useful way to introduce this idea. The resulting "fonts" would be more like dingbat fonts than traditional typefaces, and could be thought of as systems for creating art. For example, I will be showing the class Paul Chan's Alternumerics and Kelly Mark's Letraset Drawings.

I had planned to use FontForge (open-source font creation software) for this project, but it is proving impossible to work with - I haven't been able to install it successfully on Mac 10.4, and the lack of adequate documentation online makes me hesitant to use it in class. There does not seem to be any alternative free or open source font-making software available. FontStruct is very neat but its modular approach is a little too restrictive.

Since creating a working font does not seem possible, what other database-like forms might be useful for this project? I can certainly have students create an "alphabet" of visual forms without the end result being a computer-readable font file, but this falls a little short of what I wanted this project to accomplish. One thing I really like about the idea of creating a working font is the ability to then "translate" existing texts into a visual system one has created (the Alternumerics link above is really the best example of what I am imagining here).

So what I am asking is, how can I reconfigure this project? Is there a way to easily create fonts without FontForge or expensive software, or can you think of an alternative way to have students create art which is similarly database-like? I am thinking some kind of collaborative project (where each student creates a small component of a larger work) might be the way to go.

This is a small class of visual arts students who have access to Adobe Creative Suite 3, are fluent in Photoshop and Illustrator, and are not programmers.

I know this is rather convoluted and specific, but any comments or suggestions are more than welcome. Thanks!

posted by oulipian to Media & Arts (2 answers total)

This might be too far for you or your students, but you could design a "pixel" "font" of sorts. In other words, instead of thinking of the font as made up of discrete pieces (like ascenders, descenders, and serifs), think of them as made up of "areas" such as top-left and bottom-right. In the simplest case, these could be binary, either on or off, and, given a big enough grid, you could make something perhaps resembling something like a glyph. But, given my druthers, I'd not go with the binary thing and instead do a 3 x 3 grid and let students design their own tiles (a set of n tiles), and those tiles would be assembled by hitting the database. So the artistry is in the tile sets and how carefully the students think about that. One option for the "system" part is to either allow or disallow tile rotation. That's a simple thing that makes a big difference as to how I'd approach the assignment...

This would be nothing like Kelly's Letraset drawings. More like the bags Michael Beruit did for Saks Fifth Avenue (img). Doing things like Kelly did / does via algorithms is very tough. The closest thing I've seen is the new Jay Z video for "Brooklyn (go hard)". Source code is available. It's not all that visually good IMO, but then again, I haven't done anything better, so I feel bad criticizing. I have done some computational generation of related shape stuff and it's a damn hard problem to systematically describe "good" relationships between objects. When they're irregular shapes, like glyphs or characters, it's even tougher.

Sounds cool. Hope it works out for you. Is the syllabus for the class online somewhere?
posted by zpousman at 7:28 AM on January 20, 2009

Oh, and I forgot to mention this above, but it could be collaborative, too.

Have students borrow one another's tiles and configure them into their glyph designs. I could even imagine the project getting a bit more like an exquisite corpse thing. If you had students design tiles for the 3 x 3 grid but they also need to design "connect points or lines" on one or more sides of each glyph (a point or line where that tile could connect to an adjacent tile), when you composite the tiles as a glyph, things would then line up a bit better.

So the database would have a field for the img tile, then a field for rotation (if allowed), and 4 fields that described any (they could be "null") of the connect-able areas on the edges. Then you can have the software generate glyphs from the student's little databases.
posted by zpousman at 7:34 AM on January 20, 2009