Lexing is strictly simpler than parsing - lexing is done with state machines* , while many parsers can be viewed as state machines with stacks, AKA "pushdown automata" (http://en.wikipedia.org/wiki/Pushdown_automata). Lexers don't have stacks, which is why basic regular expressions can't do things like match arbitrarily nested parenthesis.

* Basically: make a byte-by-byte flowchart ("an int is a digit, followed by more digits, stop when it reaches a non-digit character and return the accumulated string as an int"), then convert that to code. Pretty simple. Graphviz (http://graphviz.org) makes pretty state machine diagrams.

Just knowing a little bit about automata / state machines will be enough to write a decent lexer. The Wirth PDF above explains it just a few pages (chapter 3). People often use tools like lex or ragel because writing them by hand can be rather tedious (being able to say e.g. "[+-]?[1-9][0-9]* { return LEXEME_INT(atoi(str)); }" rather than writing out the tests byte-by-byte), but they're not hard.

And yeah, the Lispers are able to sidestep parsing entirely.