Let me give an example to illustrate how I envisage a useful learner dictionary, Terti-Roberte Gonzaloque, and why it should be a Latin-Latin dictionary. I had already started work on this, and reading the Jones' book (pp.70-74) reinforces the conviction that it is a good thing (still a daunting task, though).
good thing? certainly! feasible? hardly. if it be a begginer work that you want, it is done already in lingua latina. though scattered in it many pages, to do such a job would only require to collect the occurances and cite them as example.
please correct
sciunt indicates knowledge over something, but the bird merely has the ability to sing, thus potest would be a far better choice.
but it's fun to attempt (I'm not saying I've succeeded) a clear, simple natural-language verbose definition that broadens one's vocabulary with closely related terms and allows verbs to be introduced in a way that is otherwise hard to illustrate. ("Lingua per se illustrata" is Orberg's nice way of expressing this approach.) If I can't guess what 'beccum' means, the definition of 'beccus' in the dictionary rectifies this. But what if I can't guess 'beccus' from 'beccumque'? The ideal interactive dictionary tags each word with the correct link, of course.
see what i have said above...
anyway, though everything you say thereafter is truth, these things are done already, i see no point to redo a work that has been done already.
i dont want to denigrate the effort you have put already on this, however i must say that you could be applying your time in a work that would be far more enjoying for you, far more rewarding and much more usefull for the kittens, or any learner for that matter, a work that, albeit not new, it still remains to be done to certain books: which is to write a begginers edition to this and that work with the lexica in latin. see our annis' aoidoi for instance, in its documents, besides the commentaries which do not concern this topic, each word has its meaning given in a friendly layout. see osberg's edition of caesar which give the meaning in latin for those not introduced in the 1st volume. - it remains yet to be done a complete edtion with a lexica in latin, this is much like a merging of thse two approachs aoidoi gives the meaning in english, osberg do not give all of them.
this something feasible [in fact is almost done already at perseus you would have to see each entry of each word, at least those which you dont know the meaning, as you would anyway, and collect the proper manings and then translate them into latin] , and would be more fruitful for you, becaseu, as yuou say, you are a learner yourself, and there is no great learning them to see them in the proper context. (that is why i think perseus' tool is so great) (well, there is the copying of text, but this is a subject matter for another rant
)
notice that the process involved are very much those which you would have to perform to create a dictionary. should you want to, someday, carry through the project in a more complete manner - like a pro dictionarist - the foundations would be laid out already. just as a tree does not grow upon weak roots and slight trunk, you should not lay your work upon badly worked concepts, which this would be without the proper knowledge of the possible meaning of the words. futhermore such work certainly produces fruits more rapidly, since the socpe is limited to the words in the work.
im just suggesing this because, as i said, it involves the same proccess and you can do this as you read the works - and not waste time in two different activities. i dont know whether this is something that pleases you, but let the suggestion stand.
What I realise now, thanks to Jone's persuasiveness, is that I need to start reading out loud more frequently AND begin the arduous task of 
) or Wiktionary. A lot of people usually forgets that it is not a primary source of in
the latin-latin is not feasible to do 
You set me up, edonnelly, by suggesting someone email the guy but I forgive you. 