syncing tags

zeppo's Avatar


23 Feb, 2018 04:17 PM

It appears that when i delete tags with the mac app, the action is not syncing to the iphone app, and worse, are getting reinstated on the mac. Seems to take multiple attempts and syncing to get it to work.

  1. 1 Posted by zeppo on 24 Feb, 2018 07:37 PM

    zeppo's Avatar

    I've created smart stack for notes having a particular tag. I am noticing when I select the stack, open a note in the stack and delete its tag (in order to remove it from smart stack) and then two finger swipe to the next note to remove its tag, etc, it may be that even though I am seeing the tag deleted, it is not actually getting deleted. In other words it may be that this is the issue causing cards to keep the tags I am trying to delete rather than the sync to the iphone returning the tags to the notes. Or it may b a combination of issues with both. At any rate, I'll delete tags this way only to see them return to the notes and I am still trying to figure out the cause.

  2. Support Staff 2 Posted by drewmccormack on 27 Feb, 2018 11:06 AM

    drewmccormack's Avatar

    So you delete the tag, and it reappears on the same note? You are using the info panel in the note editing area. Is that right?

  3. 3 Posted by zeppo on 28 Feb, 2018 01:39 AM

    zeppo's Avatar

    Yes, and now I am sure it is the sync that is restoring the tag. I've taken note of the cards and deleted the tags on the mac and gone back in to double check they remained deleted. But then after the sync the next day, they are back.

  4. Support Staff 4 Posted by drewmccormack on 28 Feb, 2018 08:40 AM

    drewmccormack's Avatar

    That sounds like a sync that perhaps was half complete being interrupted. It could be that if you let the sync complete entirely that the tags do disappear, but it is not guaranteed.

    I plan to test your data soon. I am just really busy and need to find some time to do it.


  5. 5 Posted by zeppo on 01 Mar, 2018 03:50 AM

    zeppo's Avatar

    I leave the phone app open until the sync bar goes away. On the mac I once the sync icon ends, I click on it again to make sure it doesn't have more it wants to sync. Then I close the app. Then I go back to the phone app and see if it wants to sync again, which usually it does, but it doesn't take as long, maybe 8 minutes instead of 30-45. Then I go back to the mac app and sync one more time. This is what I mean when I have said I "baby" the app.

    But as I have finished today's due notes and synced the apps in this fashion today, the tags that I have deleted on the mac app day after day, are once again back.

  6. Support Staff 6 Posted by drewmccormack on 01 Mar, 2018 07:39 AM

    drewmccormack's Avatar

    I’ll try it and see if I see the same. If so, it is most likely just that almost a gigabyte of data simply takes a long time to sync, to the point where it is maybe not a good fit for the way we sync.

    But I’ll see if I can learn anything about why it still takes 8 minutes to sync. That surprises me. An incremental sync like that shouldn’t be more than about 10 seconds.


  7. 7 Posted by zeppo on 01 Mar, 2018 02:26 PM

    zeppo's Avatar

    8 minutes is when things go well. It is usually 30-45 minutes (not seconds).

  8. Support Staff 8 Posted by drewmccormack on 01 Mar, 2018 02:32 PM

    drewmccormack's Avatar

    I can imagine a new sync might take that long with a lot of data, like you have, but it should certainly not take that long if you are completely up to date, and just study a few notes. That should only take seconds.

    Will investigate. If I don’t see the same problem, perhaps the issue is old data, which our new reset may help. It is a complete clean slate in the cloud.


  9. Support Staff 9 Posted by drewmccormack on 02 Mar, 2018 01:15 PM

    drewmccormack's Avatar

    I've started testing your library. I've only done preliminary tests. I want to wait a day and see if the scheduling of a new day causes some issues.

    Here is what I have so far:

    1. Your library is not that big in terms of notes, but it is huge in terms of study events. you have 10000 odd notes, but 220000 study events. Ie you average 20 study events for every note. I suspect that is why syncs take longer than I would expect for a 10K note library.
    2. First iMac sync took about 12mins (it's a good iMac)
    3. The iPhone 7 then synced that data down in around 30-40 mins
    4. Last step in initial sync: the iMac then did another 12 min sync at which time each device has the other's data fully
    5. I then quickly went through 100 due notes on iPhone. Synced to cloud in about 10-20s. iMac picked up with data and updated in about 10-20s
    6. I went through 100 notes on the iMac. Results were similar. 10-20s per sync

    So far everything is working as expected. The initial sync is very long, but the number of study events explains that. To improve that aspect would require quite major changes to the data internally. Doable in a major update, but otherwise too risky and too much time. And there are very few people with as many study events as you have built up.

    I still want to see what happens on a new day, because there is some retention updates each day, and I want to see if that is very expensive.

    The other thing that can be expensive is a clean up operation that will take place when there are a lot of files in the cloud. I want to see that too. It would presumably also take many minutes. Maybe that is what you are seeing sometimes. Could possibly disable that cleanup for large libraries.

    Will test this stuff Monday and let you know.


  10. 10 Posted by zeppo on 02 Mar, 2018 05:21 PM

    zeppo's Avatar


    FYI , in case it matters, I rarely study notes on the imac.

    What are study events? The number of times a card has been studied?

    One thing about the scheduling algorithm : If it is set to continue to apply the same formula to long term notes, I would suggest that it be changed. For instance I don't think if you grade a response to a note that was last studied a year ago as correct , i don't think it makes much sense to schedule the next date two years later, and in turn if that same note is graded correct two years later, scheduling it after that for four years later, etc etc. I doubt the science behind spaced-repetition extends into the realm of years. It probably is more likely that once you are testing long term memory of around a year, you might as well cap the gap between study events and reschedule correct notes at a 12 to 14 months later every time from then on. But you might add an option at this point to add a flag that "retires" the note and stops scheduling for it. Or by the same token an option after a correct response to move it to a "retired" stack, though I don't like this as much unless it retains info about what stack it was originally in , let's you return it to its previous stack should you "un-retire" it. But the user could occassionally peruse the retired stack at whatever interval he sees fit to see if he has gotten rusty on any of them. I'm sure this is beyond the vision of what you intended to achieve with the app. But I can say that spaced-repetition flashcards have made a huge impact on my long-term learning attempts. At some point, you reach the long term stage dealing with info retained for a year or longer that is still susceptible to being forgotten, and it would be cool if your app could continue to be an effective tool for that. Scheduling cards at gaps of 2 years > 4 years > 8 years , etc, as the currently algorithm probably does, makes little sense.

    Another thought of perhaps a simpler short term fix to just to get the app to handle the issue of large number of events is to cap the number of events so that the algorithm only considers the last 10 or what ever is relevant to taking a card from short term scheduling to long term of a year or more.

  11. Support Staff 11 Posted by drewmccormack on 02 Mar, 2018 06:02 PM

    drewmccormack's Avatar

    I think if you honestly grade the notes, ie right or wrong, it should work for years. If after two years you see a note appear, and you know it and grade correct, you probably will know it in several years more. Of course, it depends on you being honest about what you know. If you actually don’t fully recall the info, you should grade wrong, and go down the snake to the start :)

    Kind regards,

  12. 12 Posted by zeppo on 02 Mar, 2018 09:17 PM

    zeppo's Avatar

    On what science do you base that? You are just making an off the cuff assumption. You haven't any experience using spaced-repetition to study the same material for a period of years , otherwise the problem you app has handling what you term study events would not be a revelation to you, because you would have already encountered it yourself. i have described this syncing problem to you for years, and it has always been a mystery to you as to what the problem could be. I always grade my notes honestly even for things many people might ignore, like simply missing where a word is accented when spoken. And I can tell you that there are many things I will forget if I do not put them in practice, and this definitely holds true for a foreign language. I really think it would be to your advantage, from an engineering standpoint, to listen to people who are actually using the app to its fullest long term potential and not just for short term use to study for one semester and then be done with that material.

    There's a large group of foreign language studiers who don't have the advantage of constant immersion in the language they are learning. If I was immersed in the language I was studying and knew I would continue to be for years, I would probably discontinue use of the app because the return from using it would pale in comparison to simply studying and adding what you have learned to your actual use of the language. There is a different mechanism to storing language you have learned in memory when you are actually using it as compared to a flashcard app. The app is far inferior in that regard. But it is still an excellent tool in absence of anything better.

  13. Support Staff 13 Posted by drewmccormack on 02 Mar, 2018 10:15 PM

    drewmccormack's Avatar

    It’s all based on the scientific research used to develop supermemo, which is where spaced repetition was developed. It’s based on forgetting curves. If you recall something you haven’t studied for a long time, it is actually well established in your memory. In that case, it may well be perfectly reasonable to go years without looking at it again.

    That’s the standard SR algorithm, but nothing is stopping you from changing to a once a year regular schedule after a while. The app is flexible enough to allow that. You can even change back again to SR later, and it will continue on.

    In terms of the sync issues, you are underestimating how difficult it is to make an app like this sync. I am much more aware than anyone of those difficulties: I sell the technology from Studies for use by other developers (, and they are happy to pay, because it would take 6’months or more to get it right themselves. And even then, if you throw half a million rows of relational data at it, it will take a long time to sync. There is a limit. At some point the current generation of technology just can’t do it fast enough.


  14. 14 Posted by zeppo on 03 Mar, 2018 01:22 AM

    zeppo's Avatar

    I understand that the algorithm stems from supermemo, as that is pretty much where every flashcard app derives their spaced-repetition from. But I have never seen any trials of spaced repetition that continue beyond a couple of years (and maybe only a couple at that. So the algorithm (especially the open sourced one) is likely just an extrapolation of spaced repetition study over a matter of months. This, based on actually studying and using spaced repetition for a number of years, is where I think it is flawed to assume that the curve will continue in a way that expands the study interval at an ever increasing rate. Now this is assuming you rule out other influences on learning besides the spaced repetition, such as simply being able to use (read/write/speak) vocabulary on a regular basis. I think, for long term notes, to believe that after getting a note correct after an interval of two years means that the next interval you are likely to finally forget the note is for instance, 5 years, so it should be studied again in another 4 and a half years to prevent this, etc is simply an off the cuff guess based on an extension of a curve, rather than actual longer term trials behind it to really determine where the real next "forgetting point " most often occurs. In a sense I have actually had my own long term trials with notes and I certainly don't think going from a 1 1/2 year interval to a 4 year interval is a good idea. The idea of switching to a once a year schedule might be a solution I'll have to look at.
         From the standpoint of being difficult to program and not offering enough financial reward as a reason for not implementing changes, that I get. What I don't get is when the reason is saying a change won't matter to learning with the app based on a guess rather than actual use. But we as users just make the suggestions with the hope that the programming is feasible or that once it becomes feasible, the suggestions won't be overlooked because they may really improve the app. I like the setup of the Studies app with the combination of the mac app and phone app. I have tried other apps, but probably haven't put enough effort into really putting them to the task some of it due to the resistance of an old dog learning new tricks I think. Anki is very popular, but I don't like the website desktop model. Anyway I hope your app continues to develop. Seems like maybe every 18 months or so I'll do a google search to check out what others are using , particularly for language study, and I don't come across Studies very often This always surprises me because I have become so committed to its use.

  15. Support Staff 15 Posted by drewmccormack on 03 Mar, 2018 03:30 PM

    drewmccormack's Avatar

    > On what science do you base that? You are just making an off the cuff assumption.
    The science is the “theory” behind spaced repetition. You can say there is no long term data, but accepting your own experience would also not be very scientific, ie, it is one person’s experience. If that works well for you, I encourage you to try switching to the one year interval after a year or so of SR.

    The “theory” I was talking about is the same for months as for years. If you learn a new word, and 2 years later, when you test that word, you still know it, it is not strange to say you know the word very well. It is even more likely that in another 2 years you still know it, because you have just reinforced it even more, so expecting it to be there in 5-6 years is not unreasonable.

    This is just a description of SR, nothing more. It is based on the theory that you forget things with an exponentially decaying curve, and that the rate of decline decreases the more you have studied. This is the same for months as years, and I frankly don’t see why it would be any different, as the example I gave above hopefully clarifies.

    Of course, it all relies on “grading” properly. If you are not true to yourself, and you grade a word “right”, when you actually don’t know it very well, the whole system falls apart.


  16. 16 Posted by zeppo on 03 Mar, 2018 06:02 PM

    zeppo's Avatar

    > "so expecting it to be there in 5-6 years is not unreasonable."

    not unreasonable as an hypothesis to test, but very flawed as an assumption that it very likely to hold true. History is filled with assumptions of this sort that fail to prove true. There is a vast unknown in how memory works, but one thing is known for sure, and that is it is very complex. Empirical data from users is much more interesting to me that a very simplistic extrapolation that what holds true for human memory with repetitions over 1 to 2 year spans will hold true in terms of continuing to greatly widen study intervals over a 5 to 10 year span.

  17. Support Staff 17 Posted by drewmccormack on 05 Mar, 2018 08:49 AM

    drewmccormack's Avatar

    OK, I've now see the long sync you mention, after a day. I am now going to build special versions of the app to find out why this happens. Probably still take me a few days, but getting there.

  18. 18 Posted by zeppo on 05 Mar, 2018 06:15 PM

    zeppo's Avatar

    Thanks.  At least now I know for sure it isn't some bug with my cloud account.  I knew it probably wasn't, but couldn't rule it out and sure gives me some peace of mind.
    Let me know if tags show up again after deleting them on the mac.   Keep an eye on them for a couple of days.
    Man, if you can solve this it would be awesome!
    Also, if it has to do with cards that have a long life with many study events in their history, it would help to cull some flashcards if you can add some indicator to cards whose current study interval is, say,  more than 12 - 16 months (by interval I mean the amount of time from their previous study date to the next scheduled).  Then I could take a look at whether I think its reasonable to retire those cards without having to go into the "info" panel on each and every card.  Or if that is difficult to implement, and similar solution that could alert me to these type of cards.
        On Monday, March 5, 2018, 3:49:59 AM EST, drewmccormack <[email blocked]> wrote:

  19. Support Staff 19 Posted by drewmccormack on 05 Mar, 2018 07:50 PM

    drewmccormack's Avatar

    I think you may be able to filter notes with a due date long into the future by creating a smart stack with a due date rule in the future.

    Kind regards,

  20. 20 Posted by zeppo on 05 Mar, 2018 08:48 PM

    zeppo's Avatar

    Ok, If I understanding you correctly you are saying you may add that due date rule filter option in the future.  If its no more trouble, please add as well one for date last studied, or length of time since last studied.
    I can try to achieve some approximation of cards with long study intervals  by using filters for run of correct responses, year retention, and total number of responses.

        On Monday, March 5, 2018, 2:50:44 PM EST, drewmccormack <[email blocked]> wrote:

  21. Support Staff 21 Posted by drewmccormack on 06 Mar, 2018 07:28 AM

    drewmccormack's Avatar

    Ah, I thought those filters were already in there. I think I know why they aren’t. The due date is not connected with the note itself, but with an order of the facets. Would be good to have those filters, but it would be more difficult to do than the ones that already exist.

    Indeed, I was going to suggest using total responses as a good measure of a note that you have studied a lot.


  22. 22 Posted by zeppo on 08 Mar, 2018 03:36 AM

    zeppo's Avatar

    I've mentioned this but I figured I'd remind you of it since I myself was reminded of it with today's sync.   The time of the sync seems to grow considerably when I have made changes to notes on the mac (which is where I do my editing).  But I don't know that it coincides with the number of notes I have made changes to.  Seems like if I just make changes to three cards it has as much an impact on sync time as if I made changes to 20 cards, though I can't say for certain.   Adding new cards has the same effect.   

        On Tuesday, March 6, 2018, 2:28:11 AM EST, drewmccormack <[email blocked]> wrote:

  23. Support Staff 23 Posted by drewmccormack on 08 Mar, 2018 07:42 AM

    drewmccormack's Avatar

    I've been testing with the development app for several days now. I haven't seen anything odd, but it does have a few changes from the release version. Syncs are often a few seconds, and at most a minute or so at the start of a day, when there have been a bunch of updates.

    Just tested adding 4 notes, but didn't see any longer sync time, so that may be a coincidence.

    Will continue with daily checks, and see if I can trigger anything. Otherwise I will release what I have, which includes the must better reset function, and hopefully it will help.


  24. Support Staff 24 Posted by drewmccormack on 08 Mar, 2018 07:44 AM

    drewmccormack's Avatar

    Some questions:

    Do you start a new study session of All Due Notes each day? And do you finish them all? Or do you start a session, then keep using "Continue" until you have covered them all?

    Or do you perhaps even start many new study sessions each day? Do you finish all the sessions you start?

  25. 25 Posted by zeppo on 08 Mar, 2018 06:17 PM

    zeppo's Avatar

    I always just study All Due Notes, and then use Continue if I have to stop and resume later. The only time I ever start another study session is by accident, as sometimes happens when I am working on the mac app and intend to click on a stack name to open it for viewing/editing, but instead the "study" or "study due" pop up appears just before I click the mouse instead, causing a study session to open, which I then have to exit (very annoying by the way).

    I have also turned off cellular for Studies on my iphone and for the last few months, have started turning off wifi on the iphone as well while the Studies app is open. This is to prevent attempts at syncing until I have finished the day's due notes so that I can avoid interrupted sync attempts. When I am done studying the day's due notes I'll then turn on wifi and open the app to sync when I know I will have a chunk of time when I don't need to use the phone and can allow for an uninterrupted sync.

    I do the same for the macbook app, turning off wifi before I open it, though I don't have to be as careful with it because it doesn't take as long to sync.

    After the iphone app sync completes, I typically will try to initiate one more sync on the iphone, which usually will be very short or unresponsive (ie, has nothing to sync).

    Then I open the mac app with wifi on to sync it. Once that is completed, I click the sync icon one more time to make sure there's nothing more it wants to sync. Once I'm sure its synced all it wants, I close the app and if I have time, I'll sync the iphone app again.j

    Sometimes I have delayed finishing off the day's Due Notes until very late in the day and if it is too close to midnight, I may avoid starting a sync that might last past midnight when new cards will be scheduled, just to avoid scheduling during the middle of the sync process. Or I may have time to complete the iphone sync before midnight, but rather than start a mac sync that may continue past midnight scheduling, I'll wait until after midnight and allow scheduling to occur on the iphone and sync that first. That way I'll know what the accurate number of Due Notes should be. Then I'll run the mac sync.

    All these habits have been developed in order to avoid problems that cause the cloud sync to auto-shut down on me, often forcing me to go through the cloud reset protovol you told me about when that occurs. This used to occur a lot, but by babying the app the way I have mentioned, I have reduced the occurence to every couple of months.

  26. Support Staff 26 Posted by drewmccormack on 08 Mar, 2018 07:58 PM

    drewmccormack's Avatar

    I guess you are studying hundreds of notes a day. Is that right? My copy has 750 due, but is of course not kept up to date.

    I still haven’t see a sync take longer than about 2 minutes (after initial sync). You said yours was taking 40 mins? What sort of iPhone is it? If it takes 40 mins, does it eventually finish OK, or does it remain in a incomplete sync state?

    One last question, probably unnecessary: do you have the most recent copy of the app on each device? I have seen a few errors which I only fixed recently.

    Kind regards,

  27. 27 Posted by zeppo on 09 Mar, 2018 01:47 PM

    zeppo's Avatar

    More recently I have been studing about 150 to 200 per day, because I have been regularly adding notes. But usually in the past I might allow the due per day to get down to about a consistent 80 cards a day and then add about 30 new cards and when the average was back to about 80, repeat and add more cards. The sync has always been long. For me long syncs go back to the inception of cloud sync. If I don't edit notes, the best sync times I get might be 12 minutes. But its been a long time since I have seen that because I have been creating and editing cards for the last few months (and it only takes a few changes to impact the sync time). This morning I have 155 due notes for the day.

    In recent months the sync times have been 30 to 45 minutes. Once the sync is complete it typically is apparently complete because I always open the app again to see if another sync will begin and it usually doesn't. Once I have confirmed its doesn't want to sync anymore, then I will sync the mac app, which doesn't take near as long. Usually the mac might take a minute, although occasionally it has a long sync of 8-10 minutes, but not often.

    I have the most recent Studies version on each device. The mac is version 1.2.4 installed Feb 9. I don't know how to tell what version I have on the iphone but it is the one released on Feb 17 judging by the app store history. Usually I take a screen shot whenever I update this app to document it, but I didn't this time, so I'm not positive if I updated it prior to sending you the library. (I have one of a prior update I installed on Feb 9th). I set turn restrictions on for app installs and deletion on my phone for added security, so I don't get notifications when a release comes out, I just check them periodically, or if an update appears on the mac, then I'll check the phone.

    I have an iphone 6, up to date iOS. Now that you say you haven't had long sync times since the initial sync, it makes me worry about problems with my icloud account again. I guess I can try syncing it on another internet connection and see if that makes a difference. Or I can try syncing with cellular on and wifi off.

  28. 28 Posted by zeppo on 09 Mar, 2018 01:50 PM

    zeppo's Avatar

    Oh, and if you have 750 notes due and haven't flipped through the notes each day, maybe the long syncs aren't showing up because it is not having to sync graded notes and schedule according to graded notes. I have mine set to grade automatically as correct if i flip to the next card without grading.

  29. Support Staff 29 Posted by drewmccormack on 09 Mar, 2018 02:04 PM

    drewmccormack's Avatar

    I have been flipping through about 100 a day, but already had a backlog of about a week or more.

    It is odd that you iPhone takes so long, but the Mac not so long. My Mac is faster, but not more than about 2x faster.


  30. 30 Posted by zeppo on 09 Mar, 2018 03:14 PM

    zeppo's Avatar

    Keep in mind that I do not study on the mac (except occasionally, maybe once a month). So the mac is typically not uploading an updated study history to the cloud, it is only downloading one from the cloud (after first uploading the days study from the iphone). If I do study on the mac instead of the iphone, the mac will take longer. Though it doesn't reduce the iphone sync time to anything like what you are getting. it will still take the iphone a long time. I can try studying on the mac and then syncing it before I do any iphone sync and see how long the iphone takes to sync today.

Comments are currently closed for this discussion. You can start a new one.

Keyboard shortcuts


? Show this help
ESC Blurs the current field

Comment Form

r Focus the comment reply box
^ + ↩ Submit the comment

You can use Command ⌘ instead of Control ^ on Mac