The title is a little confusing, so I will explain my suggestion.

I like to use skritter as a way to test myself to see if I know a character or not. Now let's say I kind of forgot part of character, just the end of it. Skritter won't pass onto the next character or show the red/green lighting until I am completely done writing the character: hence I know I am not done with it. If this were to be an exam, I would have written the character uncompleted.

So my suggestion is to add a very simple answer submission mode. One would write the character however he think it's written, and then submit it where it would show if he got it right or not.

The way it currently is makes it too "guided" (imo) and doesn't force you enough to recall the strokes. For example I would try out a stroke and skritter thinks I wrote another, and puts down the correct -- but unwanted -- stroke.

If I am not clear let me know I will explain again!

This has been mentioned before, and I agree.

For example, recently I was learning 僭 and 暨. Now when I write them outside of skritter i always forget which one has a horizontal line in the bottom, because when I write it in skritter i can tell when i need to add the line from the parts i have already written.

Admittedly, I don't know how the stroke recognition system works, so I can't really speak to that.

But I do think this is a problem that can be solved with the way you use the software, specifically the grading system. Generically, I visualize the character that I'm prompted for, and if I was correct, only then do I mark it as a 3, and then 2 and 1 if I was a little off, or just plain wrong respectively.

I do agree about it being too guided at times, but in the same way, it can be overly finicky when you do know the stroke. I think there are simply limitations to what the software can do. Ultimately, you need to judge your own progress.

yeah it has been mentioned before, somewhat recently i believe :)

I think, while always trying to improve the recognition, the creators favor speed over strictness because in theory if I write it pretty much correctly, i want skritter to know it and let me move on, whereas if i write it incorrectly, even if skritter 'fixes' a stroke for me, or waits for me to add a line (like shinyspoons), then I can mark myself as a 2 or a 1 (for me it depends on how many times i make the same mistake/omission).

so basically i just repeated what zach said, but i agree that i like it this way because I can't imagine the frustration of knowing a character and having it marked wrong all the time, or having to hit "submit" every character.... oh the pain...

Right, snapping into a 'pretty' form allows for visual memory of what it should be and the associated stroke-by-stroke feedback is quite useful but the downside of both is that they do give hints (which, if large enough, require one to manually down-grade). A test mode would be nice but for ordinary practice I want that stroke-by-stroke feedback.

A stroke feedback that I'd like to see is when a stroke is supposed to cross another. For example in 力 you do not need to cross the horizontal stroke, and it won't notify you. That's kind of what I mean by strictness.

I do agree with the replies, it shouldn't be so strict that you cannot move on.

@helixness - that would be useful feedback, something like making sure that you know the difference between 已 and 己 （in 自己 and 已经 respectively） would also be quite useful.

It's not going to be technically possible for us to make the distinction between 力 and 刀 or 已 and 己, sorry!

I'm not sure about a submit mode, but perhaps it's something we can try out with the component-based recognition that's planned for later. You'd write each component, then it would submit--probably automatically, but perhaps after a slight pause or after a submit press. That way it's not snapping as you go, but it does tighten it up after each meaningful component.

Say, here's one thing I was thinking about, for when you don't know a character: what if there was a mode where it would just show you the next stroke every X seconds (configurable with the stroke order animation slider)? You would use this as a mode on the show button, so instead of seeing it all at once, you're given a minimal amount of hint at a time while still guided through a new character.

That way, you're still trying to recall as much of the rest of it as possible at each step, but the speed of showing is such that it's not slowing you down that much. If you can't remember the next stroke fast enough, it'll show up, but if you can, you can get ahead of it and use the active recall.

Sound interesting? What do you guys think?

Nick I kind of already use that by doing a few random strokes and waiting for the next hint I need to be able to write it. I would hope that the current functionality of the show button is retained if you implement something as you described.