Thanks for outlining the work flow in the new app, that will help me transition.
Yes, marking things wrong immediately is gaming the system for me but allows me to just swoop right into new vocabulary, trusting I will get enough repeats in the day’s practice to actually learn them, while keeping daily reviews of older vocabulary from getting boring. The learning mode in the new app is a worthwhile alternative for sure.
I would love to use just “study” mode but unfortunately over the years, having taken some weeks off here and there, I now have well over 10k reviews due in that mode. So I would rarely see the words I am working on now if I used it. It would be great if I could reset the counter but I don’t think that’s possible? Anyway, now I just work my way through multiple individual current lists and every once in a while tackle the main study and older lists. If there’s an alternative, I would be glad to know!
When you get the mnemonics in place for individual characters within words, I’m going to give the new app another serious try. It seems that’s the only real obstacle for me now. It sounds from what you describe that continuous study mode will suit my needs once access to these individual mnemonics are back.
Appreciate your responses. They are very reassuring. After all, I can’t imagine not using Skritter pretty much the rest of my life!
Thanks for the clarification @Procellis . I am particularly interested in the possibility of being able to add new words at will in Continuous mode, as was easy in the old app.
It keeps the review of older vocabulary from getting boring. When I feel myself getting bored, I just double-click the add button and request 2-3 new words to work on, then happily continue. They “spice up” the review.
For sure, having to switch into Learn mode is a sequence of steps I’m not really sure long-term users need, but others may disagree depending on learning styles. I’m willing to try it to see if I can adapt.
Okay, so it sounds whether you defer failed cards until your next session or add them in-line to the current session is a a significant difference between Continuous and Due Cards – enough to warrant a different mode.
That said, I don’t find that which I want is tied to whether I’m in a “continuous” or “review until done” mood. In fact, I find the behavior of withholding bad cards and giving them to me in a lump at the beginning of the next session to be counterproductive, and even when reviewing-until-done (my usual approach) I would prefer to see them again “anonymously” in the context of normal review.
The reason is I know they are the cards I just got wrong. Something I have experienced both with New Skritter and with other SRS systems that have a similar behavior is that when the failures are separated in that way my next review of them is basically a waste of time. Why? Because compared to the high-entropy nature of Chinese/Japanese vocabulary, the fact that I just got these 1-5 cards wrong is a gigantically obvious thing for my brain to latch on to. It’s not uncommon for me to literally know the answer before seeing the question. This design ends up testing me on my short term memory of failures (e.g. “the character I didn’t know is drawn this way”, “the one I thought was fourth tone was actually second tone”, etc.), not the exercise of my vocabulary.
In comparison, when a previously missed word appears in the midst of “regular” reviews, I am not told I just got it wrong. I am actually testing myself on whatever is on the card. It may be the case that I recognize it as something I just got wrong – but that’s okay because I recognized it. That’s success.
Based on this, even though my goal is to review until done, it sounds like I should be using Continuous. It would be great to get a remaining card count in that mode.
Quick update. We’re looking into the scheduling on Continuous Review to make sure things are behaving as expected and that daily scheduling rules aren’t being applied. We’ve also ironed out a lot of the details for supporting popular and user mnemonics on the study cards directly along with some updates to the info screen that’ll allow for viewing contained characters. The plan is to have these things in a 3.3.7 update and we’ll be rolling out some beta updates along the way as we get things production ready.
Couple more random observations after living with the beta for some more time.
First, I agree that continuous mode as implemented largely solves my chief complaint about beta, though a bulk adding feature would really be nice.
Some other unrelated complaints:
In tone reviews, I have not been able to figure out how to input a neutral tone. No matter how lightly I touch the screen, it still registers as some tone, and requires manual rescoring. Am I missing something?
The handwriting recognition system seems to stop checking once you’ve finished the elements of the stroke it was looking for, and doesn’t check to see if you’ve continued on making other superfluous markings. Put differently, Beta should be including the point you lifted your finger as part of its evaluation of whether to accept the stroke, but it doesn’t seem like it is. In most cases, this manifests merely as an annoyance. Skritter accepts some limited very limited cursive stroke joining (e.g., writing 口 without lifting the finger such that the hengzhe ends up looking like a loop originating in the lower left hand corner of the box). In Beta, I often find myself doing similar semicursive joins, having the stroke be accepted (thinking what I had written counted for both strokes), and then being thoroughly confused when the next stroke I write is rejected. Not sure if that’s explained well, but by way of example, in 元 I might join the 3rd stroke (pie) to the second heng. Beta accepts the combined form, but only counts it for heng, and won’t accept the final shuwangou because from its point of view there never was a pie.
That’s just an annoyance. However, the failure to check the stopping point does in some cases matter: if the correct answer is 干, Beta will accept 于 because it considered the 3rd stroke finished at the bottom of the shu and didn’t notice the superfluous gou. Given that there’s actually a chengyu for just this type of situation (画蛇添足) it should probably be fixed.
It may just be my weird psychology, but the “ban” button (being the symbol for “no” or “negative”) gets miswired in my brain when I’m looking at the “more info” side panel. Since it’s in the upper right hand corner of that panel, decades of using windows has conditioned me to reflexively hit it to close the panel and I end up banning instead of closing. Obviously I discover my mistake immediately since the panel doesn’t go away, and it’s a simple fix to unban it right there. But I do find myself doing it again and again despite knowing better. Just something to think about from an UI design standpoint: perhaps put the buttons in a different order, without “ban” at the far right. Or maybe even put a real “close window” on the far right after the ban.
Another reason to move the “ban” button: The multi-second latency problem on Android means this can happen:
You tap the info button
The app freezes and does nothing
You think you must not have tapped it firmly enough and tap it again
The app unfreezes and processes two taps at that location: the first opening the drawer, the second banning the word
This has happened to me and I almost didn’t notice it. It would be better if whatever popped up in the ⓘ location had no silent, lasting side effect. Even in the absence of application bugs, double taps do happen (cold hands, bumpy rides on public transit, etc.).
For neutral tone, draw a circle. Make sure to close the loop, or it will be interpreted as a 3rd tone. As for wanting to “add legs to the snake,” when I wanted to keep going after completing a character correctly I just manually mark it wrong.
I do the same, although it can be super easy to be lazy about my grading sometimes and just let it slide.
This discussion (and comments in some other posts) actually inspired us to add a setting to our recent internal build for a “rawest squigs” mode (name subject to change). The canvas records your input (literally anything you draw), and when you think you’re done with the character you can check your answer against the actual character. Still working out a few kinks, but it is kinda the ultimate test of your writing ability, and I quite enjoyed having it enabled for yesterday’s queue clearing session.
Yes… Skritter is coming full circle. Soon we’ll be shipping everyone 田字格 and a few pens to practice on
I actually do this on a whiteboard on a regular basis, partly to practise handwriting, but also as a sanity check to make sure I’m not cheating too much. Having that feature in the app would remove the latter reason for writing on a whiteboard, although the penmanship probably needs separate practice as I feel that writing with my finger on a small screen is almost entirely separate from writing with a marker on a big, vertical board.
Strong it is, the temptation to cheat, says Master Yoda. I just now wrote 周末 but for the second character wrote 夫 (fu) instead of 末 (mo). A small devil popped up on my shoulder and said “aw, c’mon man, that’s close enough.” I flicked him off like so much dandruff and marked the character wrong despite Skritter’s pleasing affirmation that I had gotten in right, and yet, felt a brief shudder that my Skritter soul had been in peril.
