Forums

Feature Suggestions time management ideas

Do studies in 30 minute intervals, not much longer, everything less would use less than 100% your potential. In these 30 minutes, do the learning first, then the repetitions, so you can finish even if the reps are not done completely. Deal with however many words to repeat are left by developing a strategy. In my case, I just ignore stuff I know 100% and use every day (or get enough repetition outside of memrise) - no strategy is valid for everyone though, so find what suits you best.

What was it again you need a timer on memrise for? ;)

It would be good if you could tell memrise, that say you want to study 30 minutes a day and it would just feed in new words automatically to keep that flow going. And it would force you to have to click a button or something to keep going, kind of like the time boxing feature of Anki.

not everyone wants to do exactly 30 minutes a day, first of all. if you only have 30 minutes a day to work on language study, do you really want to spend 100% of it on rote vocab quizzes?

second of all, implementing an optimal N minute workout on the site requires calculation, or you will add too many words and then never really get to them all, as has happened to me. then you're forced to ignore a lot of words you don't know because there are too many to water, and you find yourself adding 0 words for month(s) to catch up, which makes the site seem not so effective then.

hence my suggestion.

If you only have 10 minutes, well, so be it. How is memrise supposed to know how long you'll need to finish a certain number of new words and repetitions? People's speed is different.. that's the main problem here. Any idea how to solve that problem?

Don't get me wrong, I know exactly where you're coming from. Been there, done that - 8 weeks only reps, no new words, an ignore list in the thousands. However, I came here with a decent vocabulary already under my belt, much of what I "learned" here wasn't exactly new, so ignoring stuff poses no danger.

crowd sourcing? duh. if you don't think an idea is useful just say that and be done, why do you ahve to argue everyone down?

You can ask for something and say "how you do it is your problem", or you can make a suggestion on how it could be done. What was your question again?

This is the suggestions forum. It says "Get creative and tell us what new features you would like to see." So actually I don't make these suggestions hoping you will come here and "answer my question" by trying to convince me my suggestion is useless/unneeded. Actually it comes across as a bit annoying and rude. Because I am making suggestions. For memrise. Not asking for "how do I do this without suggestions," this isn't the help forum actually. So don't ask me what my question was.

I don't get where that aggressive attitude comes from. AFAIK this is a forum, i.e. a place for discussion, and that's what we've been doing. Did I call your idea bullshit? Nope. I made a suggestion on how to manage your time without a feature like this, asking why we'd really need such a thing here. Put the flame thrower back in your pocket.

I can see how you might not realize how irritating you are behaving, so I will leave you to your abilities and not blame you for them. Cheers.

Still on the war path, huh?

@ThatHorse - your idea is intriguing but I'm not sure how easy it would be to implement with an SRS type of system. SRS implies a certain adaptability based on how well you retain what you learned. In a way - I could see it be possible - but wouldn't the SRS have to create a baseline level of a person's individual speed at retaining information before it could generate an optimized study plan? What happens if outliers such as trouble words decrease the speed at which new words should be/are added? What happens if you happened to be exposed to words that you already know? Wouldn't that artificially increase the rate at which you are exposed to new words that you might not be ready for? Should Memrise create this system or refine the SRS, provide multiple SRS models (fast, medium, slow) based on language difficulty, prior user exposure to the language, or work at other features such as a trouble word crackdown list?

Personally - I like to be in control. I add words based on my error rate the previous couple of days and how well I feel I am retaining the material. However, I too, like IkenaiAndi - had to stop adding words for a long time just to stabilize. When I felt the waterings were manageable, I re-started adding words (but at a much slower rate).

agreed!! a time-management feature would be stellar! love this site but definitely spend too much time cycling back through unnecessary webpages over the course of an hour or so.

@jennifer What I was imagining wasn't something that forced everyone, but through crowd sourcing we could certainly get some estimations of the future time cost for learning each word. Hundreds if not thousands have already gone through a learning curve with the word. There probably are a number of different stereotypes for each word, i.e. the system can probably already tell based on how the planting and the first watering or two went, whether you already knew the word before and the word's cost will be like scenerio A, or whether it will be more like scenerio B where the word is totally new to you.

A "I wish I was done now" button would let you unbiasedly record how long you want to spend on the site. Then when you go to consider adding new words, you could see a message "It looks like you want to spend 22 minutes on the site right now, based on past feedback. Learn more." You can click on that to review your past hits of the button, remove them so they stop counting, or input manually your goals.

Then somewhere you could see an "outlook", based on your own watering speed for instance, right now it looks like you need 26 minutes to complete all the watering and planting set out in your queue. Specifically when planting words it could show something like this.

Now Mon (Today)-26m Tues-30m Wed-30m Thur-30m Fri-29m etc. If you add these words Mon(Today)-30m Tues-31m ...

We already have error rates for these words and crowd sourced time costs and so on, so these estimates hopefully would be useful.

Perhaps 1 week is not enough time. Even then, you could draw some really small little calendars, two of them side by side, ea. with numbers in each box. two 7x4 boxes ea. with a # inside of it could be very clear to see actually. Then you would see what adding is going to do.

@thathorse - I never imagined that this would be forced on anyone nor did I think that you were implying anything of the sort. You raise a good point with the crowd sourcing and memrise stats on user groups/word acquisition history. I am sufficiently intrigued by your idea. I still wonder about how easy this would be to implement, especially when certain core features are still not completely fixed, i.e. patrol, dictionary. I guess we would need memrise tech to weigh in on the cost/benefit of this feature. However, your idea does have merit. I could foresee me using a feature like this just to keep myself on track.

In addition - I would be interested to learn how well I am learning as compared to a group of people who have similar stats to mine. I can predict one thing about this group - bad spellers. This might actually elicit something resembling a competitive streak from me.

@ThatHorse, you are right that we have got the data on which words are likely to take more time to learn that others, and so the time that it should take to keep on top of different amounts of learning. This is complicated by a few things, eg, the speed with which each individual learns, and the personal tolerance for making mistakes. So, for example some people feel that they are failing terribly if they get below 90% in a learning session, while others feel that if they are getting above 80%, they probably aren't pushing themselves hard enough.

We are keen to allow some level of adaptation to personal preference, whilst at the same time not giving too much choice so as to become confusing, and also encouraging people to use a learning pattern that should be the most efficient.

The "I wish I was done now" button is a very interesting idea for how to implement part of this - I will have to think the effects of that through a bit more and see how it might work in a prototype. Thank you for the suggestion.

There is a linked issue here that the algorithm at the moment isn't flexible to thinks such as if you have been seeing a word in contexts outside Memrise, and therefore know if much better than Memrise thinks you do. This is particularly true of very basic words, and building blocks in the Mandarin characters. This is a problem that we are working on, and will deal with better soon. This should help to reduce the bulge of reviews scheduled once you have learned 1000+ words in a short time; some of which are of course necessary, but some of which are superfluous. It is a fine balance, but one that we now have enough data to get a good solution for.

@Jenniferhunter, you are absolutely right, all of this will be on hold until we get the features already on the site totally working and bug free. When you say that the dictionary is not working, could you be more specific? I think that it is now bug-free, though it may not be as flexible or fast as it could be. But it should be working ok. So if it isn't a bug report would be much appreciated!

Best wishes

Ben

The more I think about it, the simpler it seems. When you come to this site, perhaps you want to spend 20 minutes (or insert number here). If you spend it wrong, it's going to hurt you in the future. Sometimes you should spend that entire 20 minutes just planting. Sometimes you should spend it just watering wilteds. Sometimes you might have less than 20 minutes of watering but you still shouldn't plant anything and would be better off pre-watering (watering in advance, the ones that will wilt soonest--an option we need to have), perhaps with added attention to detail, slowly optimizing our mem choices and creating mems as much as possible.

Right now it's very likely we are all going to make the wrong choices for the above situations, so anything to help remedy that would be welcome.

@benwhately - sometimes, I looked for words and I knew they should be present, but I don't see them when I used the search function. I haven't used the dictionary (mandarin) recently because I gave up. I'll start using it again and see if those same issues come up.

"here is a linked issue here that the algorithm at the moment isn't flexible to thinks such as if you have been seeing a word in contexts outside Memrise, and therefore know if much better than Memrise thinks you do. "

An additional issue with outside knowledge is that sometimes, I learned different meanings than the memrise versions. This creates confusion - but it does work out eventually.

@ThatHorse - I think I get your major issue and why you want some memrise input in how to manage your time. I do understand as I too learned too many characters too quickly and then got overwhelmed by the pinyin. It got to the point where I didn't want to go onto the website. It really is a case of slowing down - for example, now I definitely have a higher %correct every day. I definitely don't feel so helpless when I have pinyin questions. Perhaps something as simple as a user setting stating that if my error rate increases to 25% could be useful. Then I should get a suggestion that I should increase waterings and not plant new words when my error rate passes a certain threshold.

BTW - are you a programmer? This is not meant to be offensive, but sometimes, things that seem simple to implement are actually a good deal more complicated whereas other things may seem complicated, but it's actually very simple to implement. Your suggestion feels like something that would require some serious coding time.

I programmed all through high school and two years of college but that was a long time ago. I have other hobbies now, didn't want to do it 40 hours a week at the time.

I think parts of my suggestion would be major (but really cool), other things would be much more simple and still critically useful.

The bottom line is, I think everyone has been burned by adding too many words. It's too easy to do this. We might even prefer to add words as opposed to keep up with the watering.

How about...

(1) Words don't actually get planted until it really seems like we know them

(2) Words can get unplanted if the fail is horrible and consistent (or at least we can choose to unplant them)

(3) Just put a minute value on our watering session. 150 1 words: est. 60 minutes.

I like all three of your suggestions, especially #2. #3 looks like a good way to stop "adding words abuse" as well as a good way to force someone to add words instead of getting complacent and focused on maintaining a 100% average. (I agree with you about the programming - My personal idea of hell is programming for 40 hours.) Hopefully - memrise will consider this as an added feature.

I really like the option to unplant words. Outside of memrise it's what you do naturally - you turn back a few pages in your 'teah yourself' book and go through the basics on that word again. Here on memrise the initial planting and watering through to harvest process is a lot more enjoyable than repeated watering (e.g of a word where your memory is say 50% accurate). So it would be nice to turn the clock bak on some words.

And your suggestion 1 would help minimise the number of words that came to need replanting.