Feature request: learner’s quick summary

#1
It's hard for me to figure out if a feature has already been proposed (or meta-feature request: a public bug tracker so you can mark things as duplicates and we can search easily among feature requests :)), so apologies if this is a duplicate.

I often find myself wanting a "quick learner's summary" view of words I look up when studying them. Ideally this quick summary should take into account some set of words I already know (as defined by a flashcard deck, probably) and use those to display a couple of sample sentences (which Pleco should have in copious amounts depending on how many dictionaries you own!) I'm likely to be able to actually parse given my level. I understand that word segmentation is imperfect and all that, but I'm wishing here!

Apart from sample sentences, this is what I think would be interesting in the summary page from a learner's perspective:

  1. If it's a compound word, other common words I already know that share characters with it (so I can make some sort of mental connection there)
  2. Since Pleco has the nice character decomposition data, if a word contains characters I haven't come across yet but that share major components with words I have come across, that's also good to know. All too often I'll find a character that looks familiar, sounds familiar, but might have a different radical from one I know, and it's kinda painful to figure out what I'm thinking of. Of course, I should be able to recall that sort of thing, but I'm learning so often can't unless I'm very familiar with the other character.
  3. If the word contains characters I've already seen, it would be great to see some sort of extra emphasis if the reading is different from what I'm used to, so I don't accidentally take a mental shortcut with the pronunciation.
  4. Other words I know that share English definitions with this one: if I'm looking up 发展 and already know 开发, it'd be good to display that. Of course, this particular example would be caught by #1 above, but you get the idea. Even better if there's enough metadata in the dictionaries to call out comparative usage notes. For example my OCE dictionary has the former under "(build up, expand)" and the latter under "(convert, improve)". I don't know whether that data is stored textually or in some sort of machine-processable structured form, but it would be amazing if that were possible.
  5. (Probably wishful thinking, but) English thesaurus hops to words I already know, like #4. My mind isn't going to remember the subtle differences between two English translations. This is probably far-fetched :cool:
What all these have in common is attempting to give me, the learner, some sort of "anchoring" of a new word to words I already feel somewhat comfortable with.

Similarly, it would be great if there were some "suggestion" mechanism that used similar thinking to what I include above. That is, if my known words set includes a bunch of words and through those I already know all the characters to super common word that I haven't seen before, it would be great if some mechanism would suggest that to me, so I can pick off "low hanging fruit".

Anyway, this might all be pretty out-there, but I've longed for this sort of functionality for ages in a bunch of learning software I've used. I'm a programmer and wouldn't even mind implementing this all myself, but can only dream of getting ahold the high quality linguistic data available inside my Pleco app. Pleco platform SDK? ;)
 

mikelove

皇帝
Staff member
#2
Some of this at least might be doable, yes - we're pretty laser focused on flashcards at the moment and not adding a whole lot of non-flashcard-related stuff we haven't already done until we get a finished release out with those, but in general we're working towards making the definition screen a lot more customizable (put whatever you want in whatever tabs you want), and so anything that's useful + can be coded easily can potentially go in there without too much effort or too much disruption to our existing UI.

SDK: probably not, part of the reason that we're able to get our hands on all this excellent data is that we're very scrupulous about protecting it and it's next to impossible to do that if you're letting third parties query for whatever they like.
 
#3
Some of this at least might be doable, yes - we're pretty laser focused on flashcards at the moment and not adding a whole lot of non-flashcard-related stuff we haven't already done until we get a finished release out with those, but in general we're working towards making the definition screen a lot more customizable (put whatever you want in whatever tabs you want), and so anything that's useful + can be coded easily can potentially go in there without too much effort or too much disruption to our existing UI.
Yeah, your recent flashcard tweets are what got me thinking about this, mostly tangentially because all my suggestions rely on some notion of “known words” (which would likely be a flashcard deck or several), and because I usually want this sort of thing when I’m in a “deliberate learning” session using flashcards. Fully understand that it’s pretty tangential and probably far from your primary concern right now, but getting even a little bit of this someday would be nice!

SDK: probably not, part of the reason that we're able to get our hands on all this excellent data is that we're very scrupulous about protecting it and it's next to impossible to do that if you're letting third parties query for whatever they like.
Yeah, that was more of a tongue-in-cheek suggestion. But I do think it’s an interesting challenge to design a sufficiently limited dev environment (using pure languages, probably) to do nontrivial processing without allowing arbitrary IO that would allow leakage of important data. But that’s pretty pie-in-the-sky!
 
Top