Decker's font system is a bit more limited than most bitmap font renderers, because it adds a fixed number of "padding" pixels after the strike of each glyph. If a font says it needs 2px of padding between glyphs, then you'll get 2px between "ML" and 2px between "><" — even though those glyphs have much less visual weight, therefore need to be squished closer together to look right.
Traditional bitmap font renderers give each glyph an "advance" width that can be less than the number of columns of pixels that it draws, so each glyph can have a spacing appropriate to its shape. It makes for nicer text, but I suppose it also makes the rendering code more complex, so I can understand why Decker made the tradeoff it did.
