This first bit differentiates a "backend" that refers to your font and a "frontend" that you normally work with. Why? Macros, macros, macros. Let's put it this way: You could have a myriad of styles. That's what many in the WYSIWYG world do. Then ... they have to keep track of them all. But TeX, as a Turing-complete programming language that could theoretically be used for lots of things, has the design of local redefinition, so that "sans" within a group or a macro can be redefined to something other than delicious without you needing to worry about keeping track of global style changes. In short, like the Unix "small is beautiful" philosophy (and good programming) you make parts of your document that work, you put the working parts together, and the whole thing should have a good shot at working. It will also be remarkably consistent and behave in a manner that is more regular than, for example, Word's typesetting engine. And if you are working on a source for different outputs or even different possibilities of publication, you can switch out a font without having to muck about through every font reference in your document, saving hours of labor.
\starttypescript [sans] [delicious] \setups[font:fallback:sans] \definefontsynonym [Sans] [Delicious-Regular] \definefontsynonym [SansItalic] [Delicious-Italic] \definefontsynonym [SansBold] [Delicious-Bold] \definefontsynonym [SansBoldItalic] [Delicious-BoldItalic] \definefontsynonym [SansCaps] [Delicious-Caps] \stoptypescript
As I understand, this next bit maps the "expectations" of the typesetting engine to the capabilities of the font. For example, when working with InDesign/InCopy and importing things from Word, there's a difference between character style mappings and local overrides. Word processors usually employ local overrides. That means usually the typesetting engine picks from the alternative of a font face for bold, italic, and so on, or it takes the default face and changes its rasterization to create a faux bold, italic, and so on. In good typesetting, however, you actually link character styles with actual fonts, just like you would when hand-compositing blocks of type in a typecase with lots of clamps and so on. This bit looks like what character styles do in InDesign. The more I get into InDesign and work with my designers, the more I understand the basic typographical principles that make TeX and ConTeXt elegant.
\starttypescript [sans] [delicious] \definefontsynonym [Delicious-Regular] [name:Delicious-Roman] [features=default] \definefontsynonym [Delicious-Italic] [name:Delicious-Italic] [features=default] \definefontsynonym [Delicious-Bold] [name:Delicious-Bold] [features=default] \definefontsynonym [Delicious-BoldItalic] [name:Delicious-BoldItalic] [features=default] \definefontsynonym [Delicious-Caps] [name:Delicious-SmallCaps] [features=default] \stoptypescript
The following is not in the Ubuntu-distributed manual, so I am not sure about it.
\starttypescript [delicious] \definetypeface [delicious] [ss] [sans] [delicious] [default] \stoptypescript