Graphic is a simple font that is best used for big text.

Original size: 15pt

*

A font which has a spurless, sans-serif, pixelated polygonal look which is somewhat reminescent of fonts used in VHS technology.

A lot of applied science went into this design. It's designed to remain legible on all media in all use conditions, provided that one uses the original size or a multiple thereof. Numerous technologies and mediums were employed to realize this objective.

"Diaspora" was tested and refined for use with/on/against:
     • CRT, LCD & e-Ink screens
     • image formats & compressed imagery (GIF, JPG)
     • printers (inkjet, bubble jet, laserjet, & thermal)
     • analog video & multi-generational copies (VHS, Super 8)
     • digital video (AVI, MP4, MPEG, WEBM, WMV)
     • 3D and voxel models (Blender, MagicaVoxel, POV-Ray)
     • dynamic scaling hardware (game consoles and capture devices)
     • imagery plugins & filters, including image degraders
     • image scaling/interpolation hardware & software
     • image recognition hardware & software

These all have traits which degrade, distort, compress, glitch, or otherwise alter imagery in various ways. This design aims to minimize the loss of legibility from these effects and to attain the best scores possible in various forms of imagery analysis. So far, this has proved extremely useful, as it can remain fully legible even when extreme JPG or video compression are applied to it thousands of times.

A piece of software I helped write, called the Marinan Imagery Deconstruction AI System (MIDAS), is being used on captured images of this font. The end objective is to realize the design which has the best all-around Marinan Interpretability Value (MIV) for all the tested platforms - the design which is considered by MIDAS to be the most legible in the most media under the broadest range of use conditions and quality levels.

MIDAS uses a set of considerations made with both humans and computers in mind, so a high MIV does not necessarily equal a better font - it just means one that the system thinks is easier to visually interpret. Note the use of the phrase "visually interpret" as opposed to "read". MIDAS tries to determine how well people and computers can tell what shapes are, not how much enjoyment they'll get from reading or how much strain they might undergo while doing it.

*

VERSION HISTORY:

1.0.0 - initial release.

1.0.1 - More Latin support added.

1.0.2 - First batch of tests run.

1.0.3 - gjy5&ßẞ were improved, some glyphs added.

1.0.4 - Second batch of tests run. Space width reduced.

1.0.5 - Experimentally converted to a rounded spurless design, then converted back to a plain spurless after testing. A few new ligatures were added.

1.0.6 - Cyrillic and Greek enter development. Many of these letters must be altered to be distinct from their Latin counterparts.

1.0.7 - Some spacing values changed to increase internal consistency. More difficult tests are being devised. However, since only I seem interested in this type of work, this project is going on hiatus for some time.

*

See also: AMFA, a font built with similar considerations in mind