hi markos, i saw above you were curious about my view on this. sorry i can't visit here as often as i'd like...
my view is at the end, but let me give the lead-up first - like most people i assume, i use lexical resources in two ways:
1. to get an overall understanding of a word (outside the context of reading a specific text)
2. to understand the meaning of a word in a specific sentence
for use 1 - the general one - i personally think the classic references LSJ and OLD (and cunliffe say for homer) are necessary but not sufficient. by necessary i really mean that i don't see monolingual or other resources as replacing them now or later. however my personal experience is that they are not sufficient in various ways, and monolingual lexica are among the other resources that can help top-up what you don't get from LSJ, OLD etc alone.
probably the best e.g. for me of the insufficiency of the classic lexica are for very common words like prepositions. i think if you read the specialist books on prepositions, you get a much better sense of the different meanings and how they interrelate, compared with only relying on say the LSJ entries.
building up your mental semantic networks, in particular, usually requires more than LSJ, OLD etc alone. take a concrete e.g - crito begins:
"τί τηνικάδε ἀφῖξαι, ὦ κρίτων; ἢ οὐ πρῲ ἔτι ἐστίν;
πάνυ μὲν οὖν.
probably the words in here that wouldn't figure on a typical beginners greek vocab list would be τηνικάδε, πηνίκα, ὄρθρος and maybe also πρώι. if you looked up say τηνικάδε and πηνίκα in LSJ only, i think what you wouldn't get is how the τηνικάδε / πηνίκα (and ἡνίκα) set are closer in their semantic network to the word ὥρα than other similar time-related sets (e.g. τότε / πότε / ὅτε), whereas the monolingual entry for πηνίκα linked above (which i've seen before, it must come from an ancient lexicographer) brings this out better. once you've made that connection through ὥρα, this means that when you conversely see ὥρα used, e.g. phaedrus 229a "ῥᾷστον οὖν ἡμῖν κατὰ τὸ ὑδάτιον βρέχουσι τοὺς πόδας ἰέναι, καὶ οὐκ ἀηδές, ἄλλως τε καὶ τήνδε τὴν ὥραν τοῦ ἔτους τε καὶ τῆς ἡμέρας"., it reinforces the semantic network you've mentally built up and calls to mind the τηνικάδε / πηνίκα / ἡνίκα set again as correlating well over to τήνδε τὴν ὥραν τῆς ἡμέρας - you have a constantly reinforcing feedback loop the more you read, here through the node ὥρα that you get from the monolingual but not necessarily from the LSJ.
on the other hand, the monolingual dictionary is clearly not sufficient on its own. if you looked up ὄρθρος there alone, you wouldn't get the right sense and how it relates to other time periods - i forgot that i actually discussed this on an earlier post (http://www.textkit.com/greek-latin-foru ... =2&t=10990
nor even are monolingual lexica necessary to understand semantic networks of words – currently they are useful resources but you could imagine a more systematically mapped out semantic network of πηνίκα say. you can always try to envisage what would be the optimal learning resources for future generations - but for now, since that work hasn't been done, i think the approach "LSJ+" is better than "LSJ alone" (i doubt anyone would disagree with that) and that monolingual resources can have a place in the "+" if used with caution and if you spend a bit of time with them (e.g. in the above linked work, you actually get a better sense of τηνικάδε from the πηνίκα entry than from its own entry).
for use 2 however- i.e. reading a word in a specific context - i think LSJ, OLD etc are usually sufficient, but not actually necessary. the clearest example for me is when reading works with ancient scholia – my favourite work for reading is joshua barnes' iliad, with scholia picked out and printed on the same page to make it easier to read without consulting cunliffe or LSJ. if i have come into reading a work prepared, having learnt in advance the general sense of all words used in the work (something i've recommended before), then often scholia can be sufficient to point you in the right direction when you have a mental blank looking at a word, and you feel your semantic network for the word being developed through another feedback loop. it's not so much monolingual lexica that i am recommending for this use (i.e. collections of definitions taken out of context and arranged alphabetically) but monolingual scholia on the same page of a text.