As mentioned earlier, last week I devoured Part I of John Lee's History of New Testament Lexicography. Lee has some interesting and provocative observations, particularly concerning the development of the practice of glossing and the need of lexicographers to focus on providing solid, tested definitions. It is important that the information in these lexicons be sifted and made as accurate as possible, and Lee pinpoints the areas where the most work is needed.
However, after reading Part I of Lee's book (which I need to read again to better understand all that Lee discusses) I can't help but think that the primary problem between lexicography and the popular/common user of the lexicon is a bit of a paradox.
The user of the lexicon typically desires to know what a particular word means at a particular point (e.g., "What does αὐθεντέω mean in 1Ti 2.12?"); but the writer of the lexicon typically wants to transmit the meaning of the word in general, divorced from specific context. The purposes of the user (word meaning in a specific instance) and the purposes of the lexicographer (general meaning in a corpus or collection of corpora) don't quite line up.
The lexicographer pores over all sorts of citations, instances, studies and such to gain a general idea of the word and write a functional definition, as well as provide some well-intentioned glosses. But the probable user of the lexicon is most likely at a particular verse in the New Testament wanting to know what a particular word "means" — quite possibly because this user wants to know what "the Greek" (that strange, magical language) really says here.
The problem that lexicographers need to solve is how to provide responsible information to this sort of user while still actually doing lexicography (instead of translation or an 'amplified' translation). Definitions are one way, but we also need to do better in training folks how to use these resources and what to expect to gain from their usage. My guess is that much NT lexicon usage is that of a magical answer key. Poor glosses and inadequate definition complicate the issue, so they are problems that need to be addressed as John Lee rightly points out. But the problem of proper usage needs to be addressed as well.