(If it wasn't obvious, this is a draft. Please don't cite or quote.)
Some points to mention:
In the jargon of lexicographers, there is a distinction between “active†and “passive†dictionaries. An active dictionary is geared towards foreign-language users who need to express themselves in the target language. It starts out from a list of foreign-language words and expressions, which it then seeks to give the most accurate equivalents of in the target language. Conversely, a passive dictionary is geared towards readers, and describes what target-language words and expressions mean, using the foreign language.
By necessity, this is merely a passive dictionary. There are several reasons for this. One is that this is, at least in part, a normative dictionary. The list of the five-letter root words, or gismu, is defined by their English definitions, which have been basically unchanged since 1994. The words that signify a grammatical function, the cmavo, have since 2003 been expanded from terse one-liner mnemonics to full normative definitions in a painstaking effort by the Language Planning Commission (BPFK). Thus, this book represents the culmination of the work of completely defining the Lojban language that began in 1987, when The Logical Language Group split from The Loglan Institute.
The other reason is a political one, and has to do with cultural neutrality. A way to make a Lojban dictionary maximally useful would be to open up a few monolingual dictionaries, and find Lojban translations for the entries that are most common. If someone were to put it in this effort, the result would be a dictionary that is heavily skewed towards the kind of things English speakers talk about, and contain none of the words in Lojban that are known to be difficult to translate into English with one word, such as dikcti or se'i.