Feature suggestion for flashcards (multiple choice options)

#1
It seems the choices in "multiple choice" tests are more or less random, from categories specified. As it stands now, I often find I can guess by recalling just one character.

It would be useful (from my POW at least!) and make guessing a bit harder if options presented were somewhat related - e.g. the system, when adding "wrong" choices, prioritised definitions for words that share one+ character with the correct entry, if they are available.

Just a possible improvement, FWIW.
 

mikelove

皇帝
Staff member
#2
We've played around with some ways to make multi-choice harder but haven't found one that works well in practice yet; what you're describing works great if you have a large enough vocabulary and if the word you're looking at happens to have at least one fairly common character in it, but it doesn't work so well otherwise, and I'm reluctant to roll out a feature like this if it's going to be inherently broken for most of the people using it.

However in 4.0 we are planning to let each card link to a bunch of specific other cards to draw its choices from (cards can be linked to each other in various ways) - in theory you could use that to do what you're describing, and at some point we'll probably even add the ability to import those lists of linked cards (albeit maybe not right away in 4.0) to make that easier.
 
#3
what you're describing works great if you have a large enough vocabulary and if the word you're looking at happens to have at least one fairly common character in it, but it doesn't work so well otherwise, and I'm reluctant to roll out a feature like this if it's going to be inherently broken for most of the people using it.
What about implementing exactly what I suggested - IF the pool of cards available for selection includes one or more words with one character same as in the card being tested, THEN add all/some of those as choices; fill up the remaining slots as usual? I do not see how it would break anything even on very small databases because it is only a conditional tweak to an existing algorithm.
 

mikelove

皇帝
Staff member
#4
That's going to feel messy, though; people will wonder why it works sometimes and not other times, and for reasons that have nothing to do with the difficulty of the card itself. Not saying it can never happen but it's hard to make it a priority. (I suppose we could look at doing something to add, say, just one card with the same character to the list of choices if we can find one, but that's the sort of minor quality-of-life tweak that it'll take us several 4.x updates to get down to the level of considering)
 

mikelove

皇帝
Staff member
#6
In general we do want to make multi-choice more challenging, I'd just rather focus on doing it procedurally, e.g. by generating similar words that are total nonsense, making subtle changes to the pinyin, etc; I'm less inclined to build a feature that relies on a lumpy / inconsistent aspect of Chinese words.
 
#7
To be honest, I do not understand how nonsensical words would make choosing the only “sensical” word more challenging and how modern Chinese being predominantly bisyllabic is inconsistent. Then again, mine was simply a suggestion, I thought it was a no-brainer, apologies.

After all, multiple-choice is easy by definition, compared to strict self-grading or maybe literal comparison of user input to the card - which works only to Chinese, not from.

Conceptually, if anyone cares, I always thought flash cards weren’t an organic and well-used/loved part of Pleco (Pleco has always been and remains the greatest Chinese dictionary / dictionary shell).

I’m currently in the middle of extricating myself from flashcarding and rethinking it. This is likely the reason for all my questions and problems. Anki remains the best flash card environment tho, partly because of its level of openness and configurability - plugins ;)

I’m not asking to make Pleco open in a similar way, just suspecting that Pleco flashcards will never catch up to Anki’s. They definitely will not if they continue trying to be a general flashcards/SRS system, and ignore traits specific to the Chinese language. Exploiting the limitedness of a subject domain is a great way to produce an exceptional product serving that domain (again, that’s what makes Pleco so good - it’s constructed around servicing Chinese dictionaries, not French and not even Japanese ones!)
 

mikelove

皇帝
Staff member
#8
I’d actually disagree with you there - we do quite a lot for Chinese and are doing more, I’m just not sure about your specific suggestion. (Sometimes I wonder if I’d be better off just accepting suggestions without providing commentary / feedback - no need to get into arguments about this stuff, and anyway plenty of ideas I like end up not happening / ideas I dislike end up happening)

Flashcards spent a long time getting ignored simply because we had too much else to do (iOS and then Android ports and then a few years of breakneck OS updates to deal with) - they no longer are, and the new system will let you replicate Anki’s behavior almost perfectly if you want it to along with doing a tremendous amount of stuff Anki doesn’t do.

If people still prefer Anki after that we’ll cheerfully go back to focusing most of our energy on dictionaries, but flashcards have been a major part of Pleco for long enough (and are still popular with a sizable enough % of our users) that we have to at least try to bring them up to date.
 
#9
mikelove said:
I’d actually disagree with you there - we do quite a lot for Chinese and are doing more, I’m just not sure about your specific suggestion. (Sometimes I wonder if I’d be better off just accepting suggestions without providing commentary/feedback - no need to get into arguments about this stuff, and anyway plenty of ideas I like end up not happening / ideas I dislike end up happening)
I meant providing some specific treatment for flashcards, given they are only (meant to be) used for studying Chinese. Of course you do a lot for Chinese in other parts of your app, I just did not notice much specific to flashcards - a handful of options available from a hundred+, maybe :)

Re arguments, there's always the old XKCD:

And it's very nice that you still have time and desire to provide essentially individual feedback to users - I'd expect Pleco is not a two guys and a dog operation by now!

mikelove said:
If people still prefer Anki after that we’ll cheerfully go back to focusing most of our energy on dictionaries, but flashcards have been a major part of Pleco for long enough (and are still popular with a sizable enough % of our users) that we have to at least try to bring them up to date.
If you ever have a resource contention and face a choice - more dictionaries / dictionary features vs more flashcard tweaking - I vote dictionaries. Pleco is literally THE tool any serious learner of Chinese uses. You probably do not know - there is Chinese guy here in AU with a sizeable podcast https://www.chinesetalkeze.com/journey-to-chinese-fluency/ - in every episode he asks his guests to recommend a tool they cannot live without. Nearly everyone who learned Chinese in the past 5-ish years mentions Pleco as THE tool. And it's not flashcards they mention ;)

Flashcarding is easy if one is computer literate - exporting from Pleco is easy (+ "chinese support" plugin in anki or something).

PS I guess I have too much time to procrastinate on forums instead of learning Chinese :cool:
 
Last edited:
#10
(Sometimes I wonder if I’d be better off just accepting suggestions without providing commentary / feedback - no need to get into arguments about this stuff, and anyway plenty of ideas I like end up not happening / ideas I dislike end up happening)
That's one of the dangers of putting your thoughts out on an open forum - the discouragement you can feel when someone who... probably knows a little less than you do on a matter... comes to tell you how you should do things. (That's not singling out you werewitt, but this is a general problem I've seen many times in many places)

I appreciate hearing your thoughts, even if it's not what I want to hear.

Thank you for that.
 
#11
Hi werewitt and Mike,

I would also greatly miss Flashcards on Pleco if it were to be removed. That was more of a threat to show where the discussion was heading. I very much doubt that Anki ever reaches Pleco’s level of usefulness without any Chinese dictionary to look things up while studying. Also, comparing Pleco 3.2‘s aging flashcard system to the current Anki system isn‘t quite fair. Version 4.0 is supposed to be configurable to behave almost exactly like Anki‘s algorithm. Sometimes focusing singularly on the technology around it distracts you too much from the actual aim, which is to learn words. Pleco 3.2’s flashcard system is still very useful for that.

Discussions should only serve the purpose of making one’s point of view clear to the other side. Not to resolve all differences, but to fully explain oneself so both users and Pleco can complete their picture of what the the other side wants. Once one’s ideas have been clearly communicated, then it’s the other side’s job to think about them and work with them. So maybe it’s advisable for Mike to always leave a door open for a suggestion in his wording because he’ll never know for sure he won’t ever do anything in that direction, and the user will more easily understand that they can’t get a straight answer simply because many decisions take time.
 
Last edited:
#12
Not sure why anyone would think @mikelove got emotional over this little discussion - I definitely haven't. From experience, user opinions need to be repeated many times over until they receive any sort of decisive feedback, either "yes we will work on this feature in XXX" or "won't do" :)

It is nice Mike is very forthcoming with his feedback, including the "won't do" type. This is not too common in software vendor world. Then again, interweb is conducive to all sorts of misunderstandings.
 
#13
I wouldn’t press too hard; in my experience, it just destroys sympathies and doesn’t give you anything. ;)
 
Last edited:
#14
I wouldn’t press too hard; in my experience, it just destroys sympathies and doesn’t give you anything. ;)
I'm not pressing too hard, I honestly hope I'm not.

But - this is the internet. In the very worst case - a new account takes about 2 minutes to set up, big deal ;)
 

mikelove

皇帝
Staff member
#18
No, not emotional at all, I just don't want people going away unhappy because I explicitly said no to something when I didn't have to + might actually still end up doing it.

In general, yes, focusing on underlying goal makes sense, but the aim in 4.0 is to build the technical foundation for a flashcard system that we can use for another decade (so that we can in fact get back to focusing on dictionaries) - part of the reason the current flashcard system is so limited is that it wasn't built to do the sort of stuff people are now asking it to do, a lot of the things people want us to add (which we may not add in 4.0 but will add eventually) would be pretty much impossible without a bottom-up rewrite.
 
#19
Indeed; I was just expressing a general tendency, of course it doesn’t apply to everyone in equal measure. I am certain that 4.0 will be even simpler to use, with a powerful engine under the hood and configurability in the right places. I’m also impressed by the fact that the core is pure C code. :)
 
#20
Yeah, it always helps to look back on the purpose of it all. :) Techies run the risk of digging into the tech and forgetting what it's for.
I wouldn't start down this rabbit hole. Flashcards on a phone with modifiable settings (and apparently some scientific support for SRS in general) are a techie thing. The moment flashcards stop being purely stacks of paper, their purpose becomes "to learn words with better retention / quicker / more words etc" and fiddling with settings is supporting exactly that goal. Otherwise they would have remained literally an electronic version of stacks of paper to be drilled in sequence / entirely randomly.

:cool: Pardon a philosophical aside.
 
Top