掲示板 Forums - A couple things about the word dictionary

A couple things could be improved about the word dictionary. Here's a screenshot of the beginning of a kanji review session where I get 5 new kanji, and 2 problematic words fit on 1 screen.
For the first highlighted word, you actually got the whole definition as it was in the original database, which means whoever entered this definition didn't really know what 埋蔵 means (hence the"??"), and I need to change the definition. In the second word, somehow you only picked the end of the definition, the figurative meaning. Maybe an issue with how line breaks were handled during the copy of the database. I have to edit this one too. But I can't do it from here, and not from the dictionary page of those word either. My only opportunity will be when the word comes up during a test, but then I might be in a hurry... It would be nice to have the possibility to edit the definition from this screen and from the word dictionary.

Also, for each kanji that I learn, I try to learn at least one word for each of its readings, even if it means adding many words that are way beyond my current level. I just don't know how I could ever remember that the on'yomi of 埋 is マイ if I don't know any word with this reading. So I have to pick either 埋蔵 or 埋葬, whichever is more useful. Currently, the way I do this is I go to jisho.org, and compare the frequency indication. 埋蔵 is JLPT1, so it's between 6,000 and 10,000, while 埋葬 is simply a "common word", meaning it's between 10,000 and 20,000. So I choose 埋蔵 (the definition of which I will need to edit) as "my" word with the on'yomi of 埋. I also regret I don't have an indication of frequency when I see example sentences with words highlighted in orange (meaning they're unknown) and I wonder if I should add some of them to my schedule or not. 
It would be more simple if the JLPT level and frequency indication of words was shown by renshuu.
Perfection (not of this world?) would be with one colour for unknown words that are within the frequency target defined by the user (say, still orange) and another colour (grey?) for words beyond the frequency target... With the possibility to define the frequency target as 1500, 3000, 6000, 10000, 15000, 20000, 25000, 30000, 40000, 50000.

I read on jisho that this "common word" indication is based on a ranking of the most common 20,000 words in groups of 500. If you have this list, it would be nice to indicate it too: if I have the choice between learning a word that is around the 12,000th most common and the 18,000th, I'd rather know and choose the first, as I might actually see it in a text at some point.
And if you have these 40 groups of 500 words, they would also be a very interesting schedule. For example I'd immediately add the first 6 groups to my N3 training, I'm sure that would add at least a couple hundred of very useful words not covered by the N3 list.
If you don't have this list, a few frequency lists are available:
there's the frequency list of the Japanese Wikipedia, on wiktionary, up to the 20,000th word. But the article says that 166,000 words (de-inflected) were found. Whoever built this could probably provide you the whole list.
There's also a list of around 35000 words from newspapers, by Alexandre Girardi, which would be great for people aiming at finally becoming able to read the newspaper, in some distant future.
There's also a 60,000 word list based on blogs. Maybe the most useful list to work on conversational Japanese, since blog posts are often written in a language resembling conversational speech?
.

It would be really great if you could somehow display an indication of the frequency, switching to jisho several times per session is a bit silly.
For the definition, I don't mean suggesting a change of the common non English definition, I meant changing my custom definition from the dictionary, the same way I do during tests. The most frequent issue with French definition (and very likely with other non English definitions) is that the beginning of the definition (before a linebreak) is missing. So, if you see why that happened, you can reload the definitions in the affected languages and fix at least hundreds of definitions in each affected language (something like 5% of the words, in French).
If it helps I could spend a few days doing the tests in Italian, then Spanish and then Russian instead of French, and I'll see if the same thing occured in those languages. My German is rather poor but I can at least test my N5&N4 vocabulary list in German. The problem is rather easy to spot, because typically the literal meaning is omitted, there's only the figurative meaning,.

A possibility to add the frequency info without adding visual complexity could be to have an option to sort words (whether in a dictionary search or the words containing a kanji) by frequency groups: N5 N4 N3 N2 N1 "common", and then the rest. So the user wouldn't see N5, N4... in front of each word, but they'll be in order of usefulness, which is what matters. jisho uses this order... It would get  even better than jisho if you split the "common" group in 2 (before or after #15000...), or if you add another group beyond the 20k common words (like the 35k common newspapers words...), but any sorting by frequency would help. They also use the wanikami lists, whatever that is.
Take the kanji 補. The words that should appear first are 候補, then 補う, then a few N1 words like 補給.

Instead, just after the most common word 候補, many rare words follow.
I couldn't find oginau in the first 3 pages. Maybe I missed it, but it's not on top anyway, which hides how useful it is.
It would probably be an improvement to show words in frequency order (and akasatana within each group?), with only one line added to separate what the user considers as useful (for him) and completely superfluous, whether it's the N2 level, N1 level, common words, the 20k common words, or the 35k common words, or 50k.
So, in this example, this page would display the 8 JLPT words, then 6 more "common words", then a line saying "rare words", then the rest (if the user selected "common words" as the useful vocab limit).
Hope this helps.
.

Should be MUCH better now, but it needs additional testing. Also, jlpt markers are in the vocab dictionary, and the other frequency markers will be up soon!

A one-letter typo caused that one! Should be fixed.

Aaah, much better, yes, all the words on top are "common words". And with the frequency markers in the vocab dictionary, I guess I can remove jisho from my bookmarks now, your indications are much more accurate. Here comes some testing:

So, all the words on top are common words, but the 2 most words we're supposed to learn first according to JLPT are placed 9 and 10, as you see. So maybe there's something to do to bump them up?
If I want to check the frequency of, say, 補欠, I go to the word dictionary tab and ask for hoketu, much easier than before, using a separate tab for jisho. But would it be possible to get directly there by tapping / clicking on the kana ほけつ? That would be ideal... Currently there's no direct link from the kanji dictionary tab to the word dictionary tab. Except when there are example sentences, in which case I can tap the word in the sentence. But among the 10 "common words" above, 3 are missing a sentence...
I just noticed, using a language other than English is useful here: if the definition is in English, I know it's almost certainly not JLPT. If I had paid attention to this, I wouldn't have mentioned this problem at all. I'm glad I did, it's quite an improvement.

...but please don't completely get back to the separate entries! I think the merging is a brilliant change. You've achieved an amazing number of improvements, in the 3 little months I've known the site, and that was already starting from the best Japanese learning tool there is.

How about keeping terms with kanji-kana variations and alternate kanji variations together and terms with different kanji or othography separate? 

For example: 見逃す　and 見のがす, and 基本形　and 基本型 would be combined; but 　見落とす and  見落す,  and   硬い　固い　堅い would be separate.

Would make it easier to find what you're looking for.

For now, split them out into separate lines with all the icons, but I'm thinking of wrapping a few more of the icons into the settings dropdown box since you are less likely to use them for the alternate terms. Thoughts?

Edit: for alternate readings, wrapped all icons into a dropdown box, left sentences out.