The PRs just got merged though 
1 Like
While it would make the wording easier, I don’t think it makes sense to include them if the feature is intending to limit cards. If you’re suggesting it be changed so it’s instead a reps limit, then that could result in users having study terminated when intraday learning cards are still in waiting. That seems like it would lead to wasted effort/more forgetting.
I didn’t see @snowtimeglass’s comments before merging the PR, and did not mean to prematurely halt discussion.
2 Likes
We don’t need to count reps - simply, if intraday learning cards from yesterday are appearing today let them count towards the limit. As I write this I realise we are talking about different things, whereas you’re thinking about counting the same cards multiple times if it becomes a intraday learning card from a new card, I’m thinking about intraday learning cards from multiple days before that are due today. @snowtimeglass are you talking about the former or the later?
Edit: If what I want is added, blue+red+green will always be equal or less than the maximum limit. It’s good to see that all of this was just misunderstanding each other, I hope we will reach a consensus now.
If anything, about the latter. I suppose I have concerned about the case with the latter type* of the cards, especially in the example:
*: intraday learning cards which are already overdue that day.
I’m still not sure that Dae is thinking that way as mentioned above, but in any case I’m also somewhat concerned about the point that Dae mentioned, i.e.,
could result in users having study terminated when intraday learning cards are still in waiting.
I mean, I have yet to dispel the impression that this would occur even in the latter case with “just one count for one card”.
To confirm, it means that the count of intraday learning cards is capped at “Maximum cards/day” when the number of the waiting intraday learning cards has been greater than the number of “Maximum cards/day”?
(For example, if an user has 20 waiting intraday learning cards and their “Maximum cards/day” is 5, then the count of their intraday learning cards to be studied that day becomes 5?)
Dae said, “could result in users having study terminated when intraday learning cards are still in waiting.”
I ask everyone to differentiate whether a situation is actual or probable. If this is a concern for a user, they would increase their maximum limit, would they not?
Now, I don’t give a very good argument if a user has to make changes in limits just to finish the intraday cards. But, here is the important part, an intraday card that is overdue is much of the same nature as that of an interday card. Read further.
I notice that Dae is talking about study being terminated midway, which will only happen if you count repetitions. If you count cards, whether a learning card from 2 days before has 24h interval or 8h interval shouldn’t matter. That is, every learning card is a de-facto interday card when it is counted.
The count of all learning cards are capped by the maximum limit, but we don’t count a card twice. If a card has been counted as review card or a new card it is not counted again when the state becomes (re)learning. But the card will be count towards the maximum limit after the day rollover.
Are there still concerns about not treating intraday cards differently?
The code counts reps. It just works out the same as cards at the moment, because we don’t include intraday learning cards, and they’re the only ones that may be shown more than once. If we start including intraday learning cards in that count, it will end up being a rep limit instead.
Ensuring intraday learning cards don’t end up getting counted more than once would require keeping track of which ones have contributed to the count that day, and it adds a bunch of extra complexity.
I thought you only need to count the cards when gathering them.
The current situation would seem like a baffling contradiction to many and it is worth a lot to make things easier to understand. However, I can’t evaluate the merits-demerits of any necessary code changes, so I’ll leave it upto you to decide.
Not to nitpick, but to confirm my understanding or misunderstanding, you mean “every learning card that is overdue is a de-facto interday card when it is counted.” here?
I did mean that. From an user’s perspective, there is no difference.
To be clearer, what I’m trying to say and what I’d like to confirm is: you mean “every learning card that is already overdue that day (i.e. each card searched by is:learn is:due -prop:due=0) is a de-facto interday card when it is counted.” here?
The more you repeat a particular phrase, the more we lose the context of something. I am sure you’re not misunderstanding me. The sub-day intervals of intraday cards mean that study getting terminated midway isn’t particularly helpful for the learner. But that is not the case if you’re not counting reps and you’re looking at overdue intraday cards. I think the search term is right.
snowtimeglass, any reason why are you being so specific?
It is not because I care about its so specific definition but because I thought I may be misunderstanding what you referred to there, and I supposed that by being specific, the potential misunderstanding of mine may be revealed and resolved.
“every learning card that is overdue is a de-facto interday card when it is counted.”
My understanding still doesn’t catch up with this thought well.
It is because I suppose as follows at the moment:
Intraday learning card is a learning card that has not yet successfully passed an intraday step followed by an interday delay.
Therefore, when an intraday learning card is overdue, that means the card is encountering an interday delaying without successfully passing even the intraday step.
So, when counted, an overdue intraday learning card seems to be more of a disproportionately delayed immature learning card than a de-facto interday learning card.
I think you are correct, although as I said, we have already forgotten about the context in which I said what I said. I meant overdue intraday cards getting buried for the next day is not much different from a interday card getting buried for the next day. That was the point of similarity my analogy was about. But perhaps to you I seemed to mean overdue intraday cards are perfectly similar to interday cards in all aspects. Be assured that is not what I meant. I was simply arguing that there is no more reason to include interday cards to the limit than there is reason to include intraday cards to the limit.
I agree that that perfection is not needed. I don’t and didn’t think you seemed to mean overdue intraday cards are perfectly similar to interday cards in all aspects. That perfection is impossible and unnecessary.
To confirm, my question is not whether they are perfectly similar but whether the difference between them is really not important when considering the suggested change.
Dae probably isn’t adding them to the limit, so at this point much argument is futile. But I do think the difference isn’t important. After day rollover you have already broke the chain of reviews. It doesn’t give you much advantage to study the overdue cards the next day over any other day. I can think of exceptions though (if you were studying late at night and you continue to study after day rollover, for example).
When focusing on the current situation, which doesn’t add intraday learning cards to the limit, “Maximum cards/day” doesn’t seem to represent its actual role in practical sense. The limit number actually works as “Maximum reviews/day”.
The meaning of “Maximum cards/day” is so intuitive and inclusive, so it would be inflexible proportionately.
" ‘Maximum cards/day’ is always ignored by intraday learning cards, and can be ignored by New cards."
That “Maximum cards/day” seems to be inflexible contradiction as an expression for that limit number.
I think we are too used to “Maximum reviews/day” that “Maximum cards/day” seems unituitive. I suggest waiting and if people complain, the changes can be reverted then.
For me, personally it’s not that hard to imagine why leftover intraday cards are slightly different in nature from other cards. Though for all practical purposes they should be included to the limit but it doesn’t seem dae is keen on working this out. If you or somebody else agrees to solve this issue, he might accept a PR. In any case, I would wait until the next version is released before I conclude “Maximum cards/day” is a bad change.
I stayed out of the original discussion about whether this change should be made or not (only offering some suggested language). But I now regret not speaking up about my concerns. I realize the train has left the station (multiple PRs are merged) – but here is an example of a concrete user who is going to be made more confused by changing “maximum reviews/day” to “maximum cards/day.”
Reddit - Deck ignoring card limits [Including a screenshot since folks on reddit often delete their posts.]
Responding to that user clarified the issue for me –
Those 27 cards got there in one of these ways –
- Cards that you introduced as New today and are still in Learn
- Cards that you studied as Review today and lapsed into Relearn
- Cards that were in Learn/Relearn yesterday and left unfinished (i.e. not graduated to Review)
- Cards that are on a 1d+ learning/relearning step
The ones from #1 or #2 have already been counted as part of today’s daily limits on New or Review cards. The ones from #3 or #4 were counted as part of the daily limits on a previous day.
Granted, the user was confused even without the change! But if the change is released, it will be much harder to explain to someone why they still have Learn/Relearn cards to study beyond their max “cards” for the day.
I am not advocating for Learn/Relearn cards to be subject to any of the daily limits – I agree with @dae, that is impractical and bad for learning. I am advocating that these changes be rolled back.
[From a number of different mentions in different discussions, I understand that @sorata is concerned with the uses of the word “reviews.” But I don’t think it’s all that confusing, and I think there are better ways to clarify it in the documentation/manual/tooltips, without changing the names of features. Rather than complicate this thread further with discussion on that broader issue, perhaps that should break off into its own thread about the use of the word “reviews” to mean multiple things, and whether it is problematic.]
2 Likes
From your experience, where do most people learn how to use Anki? If they read the tooltips or the manual, how did this user come to believe the review limit limits every card in his deck. (Perhaps the source of such confusion should be identified?)
But I am finding it hard to believe that the average Anki user will decrease his review limit and check for changes in his learn count. If serving such users is the goal, might we not be looking at the wrong places? I am saying, how can you come up with anything that wouldn’t confuse this group?
The real help one can provide to them is probably through support. Am I wrong? So the question is, is it any harder to explain one exception to maximum limit than to explain two exceptions to review limit? I think the tooltips are better now at least.
However, most people who use Anki, probably wouldn’t be confused by this change. But that’s a hypothesis. @Danika_Dakika I suggest we keep the changes for now. If you come to notice the current wording is creating more confusion among users (new users in particular) we can revert the changes then.
But in the future, I would like Anki to include all cards (including intraday learning) in the limit.