Sharing SuperMemo Databases?

[Jonny Wang]

I haven't been able to find reference on this forum to supermemo (Palm version) databases that people are willing to share with other Chinese learners. Are there databases out there that people are willing to share?

As I'm using/have used the Chinese Made Easier curiculum (sp?) I'd be especially interested in a database with vocab from that. Speaking Chinese About China (话说中国） would be nice too.

Is it even possible to share databases? I'm still learning some of the details of the software. I hope to be able to share databases in the future as the ones I'm making get large/complete enough to be useful.

[Nibble]

Supermemo Palm does not have a built-in function to export databases that you have made on your PDA; but you can use a program like FileZ to copy the database files onto and off of any Palm PDA. The files are saved on the internal memory as "sm_[database name]", and you just have to reset the learning process when you put a new one on.

[Jonny Wang]

Gotcha. That makes sense. So, if people have made their own databases, they can share them with others. Nice. Now if I could just find somebody who has the databases I need....

[klortho]

[This is really in response to this thread, but it fits this topic better, so I posted it here.]

Supermemo - yeah. As a friend of mine said, it's definitely a lifestyle choice. These days, I do about 600 repetitions / day, which, if I work straight through, takes about 3 hours. Sometimes it just seems unbelievably burdensome, but I can definitely see the benefits, and I can't imagine how I could ever learn these characters without it. Before Supermemo, my study of the characters was all haphazard, and I just knew that I was forgetting them as fast as I was learning new ones. Now, it's arduous, but consistently uphill.

I definitely have been putting a little bit more emphasis on memorizing characters than on memorizing words, but I do put both into Supermemo. My thinking is that I want a solid foundation in the characters, period. Of course, you have to learn multi-character words, but the characters are key, and, for me, the hardest part about learning Chinese. I've suspected from the begining, and I think it's being born out, that putting more emphasis on the characters first will make the whole enterprise easier in the long run.

In general, for every character or multi-character word I put in, I have two Supermemo items. For the record, right now, I have just under 4800 character items (2400 characters) and 3400 word items (1700 words).

My strategy is to go through the textbooks I use in my class, and first add all the new characters from the "new words" of each chapter. Then, time and spirit willing, I add those multi-character words that I think are most useful and/or interesting. I also sometimes add words based on whether or not I think they'll make some of the characters easier to remember.

I also add characters and words from other sources: people's names, street signs, etc. Sometimes, if I get really bored, I add characters from the character-frequency list in Wenlin. As of now, the most frequently used character that I don't have in Supermemo is at position 1108 (欲 yù desire, wish). Oh, and one other source: my latest hobby is to study a little bit of chemistry in Chinese, so I have the characters for the elements memorized up to calcium (钙 gài, atomic number 20). Not as useless as you might think! I've already been able to pick apart some obscure ingredients on the labels of snack foods.

For each character, I add two items. In the first, the question is just the character itself, and I have to remember the pinyin and the definition. I cut-and-paste the answer from Wenlin, and then try to massage it a little. I try to keep the definition very simple. And, if it's a duoyin zi (multiple pronunciations) I have to make a choice. If both pronunciations are very common, then I make myself remember both. Otherwise, I just remember one, and put the other pronunciation into the answer with a little "for reference only" tag.

Then, I use the Supermemo duplicate (Ctrl-Alt-D) and switch (Ctrl-Shift-S) to make another item with the question and answer reversed. So for this item, I'm given the pronuncation(s) and definition(s), and have to remember how to write it.

A lot of characters are only used inside multi-character words, and some people have told me that I shouldn't try to remember them by themselves. For example, "奢 shē" is pretty much only used in "奢侈 shēchǐ luxurious", so, they say, I shouldn't have flashcards with just the character "奢" by itself. I ignore this advice, even though I'm sure it's right, just because I am pigheaded, I guess. I want to be able to look at any of the characters in my repertoire, by itself, and be able to recognize it.

For multi-character words, I pretty much do the same thing, with one exception. In the second item, I sometimes move the pinyin pronunciation into the answer. That is, for the second item, the question consists of just the definition, and I have to remember both the pronunciation and how to write it. Whether or not I do this depends on how specific the definition is. For example, one of my most recent flashcards has the question "May Fourth Movement", and the answer is "Wǔ-Sì Yùndòng 五四运动". An example of an item where just the definition would be ambiguous has the question "xiāngdāng v. match; balance; correspond to; be equal to" and answer "相当".

So, that's how I use Supermemo. I'd be interested to read how others use it. If anybody wants a copy of my database, I could try to figure out how to export it and give it to you. Let me know.

[realmayo]

Interesting. thanks very much for the reply.

I had been toying with the idea of two separate databases: one from Chinese to English and the other English to Chinese -- but you have them all together in one ... so you might be asked a character first, then an english definition next? if that works for you I'll probably do the same.

three hours every day does seem a long time -- do you normally spend that long? and judging by your post in the other thread, using Supermemo you've learned almost 1500 characters in a year ... as well as lots of vocab. are the characters starting to sink in more deeply now? or if you didn't keep Supermemoing them would they fall out of your head pretty fast?

I don't suppose anyone else has any more supermemo stories, good or bad?

[leosmith]

I've used SM extensively for Japanese. My single character study method is very similar to klortho's. The main difference is that I also have multi-character sample words in the answers of the Character to English collection. But this is mostly because there are at least 2 common pronunciations of each character in Japanese - you may not find it necessary for a simpler language like Chinese.

I also use SM for grammar. English to Target language, and Target language to English. Again, Japanese grammar is pretty complicated, so this may be a waste of time for Chinese.

My learning sessions vary depending on how much new stuff I've added recently. My sessions are about 1 hr now. I have about 2200 characters, 4000 words, and lots of grammar entries. Everything is duplicated (E-J, J-E), so I have about 15,000 total flashcards.

klortho, 3hrs would drive me crazy. I've been there before, but only when I was dumping tons of material in. Two questions for you.

First, are you entering stuff into SM without any prep work? I found that for me, this is terribly inefficient. Starting out learning with flashcards is too hard. Now I start with lists, study new stuff for several days in a row, then enter it into SM. SM is just a review tool for me. An incredible review tool, but only a review tool.

Second, how are you learning to write your characters? Are you using the Heisig method (or "story" method)? For me, advanced mnemonic methods such as this are far superior to other methods, and are the only thing that actually keeps me from forgetting.

[klortho]

realmayo, I really don't see any reason to have separate databases. I remember reading somewhere on the SM website that they don't recommend that, either. It seems like it would be a maintenance headache, and there's nothing wrong with having the questions come at you in a more random order. I'm usually incredibly anal about such things, but I actually spend very little time organizing my SM collection.

3 hours is brutal. This is about the maximum that I can tolerate. For the last year, since I've started using SM, I've been in China and just going to school -- not working -- my reps have hovered between 400 - 600 (2 - 3 hours). I think part of the reason is that my memory really sucks, and I find characters incredibly hard to remember.

When comparing characters and/or Chinese words to, for example, names of western actors, which I also throw into my collection more or less randomly, there is no comparison. I can remember the actors names, even relatively obscure ones, very reliably, after only a couple of wrong answers. Characters, for some reason, just do not stick as easily. Anyway, like I said, my memory is shit, and that's really why I can't imagine being able to make the progress that I have without SM.

leosmith, as for your first question -- I agree that learning straight from the flashcards doesn't work, but my method differs from yours. I first enter the characters or whatever into SM, and then, while doing my reps, when I come across a brand new item, or even an old item that I've completely forgotten, I then jot it down on a two-column quiz sheet. Then, I try to review the quiz sheet two or three times before the next day's reps. This method seems to work pretty well for me.

As for mnemonics -- I definitely use them, but I don't know anything about the "Heisig method". Actually, the thing that I find hardest to remember are the tones of Chinese, and I have a mnemonic system for those that works pretty well. As for other mnemonics, I try to do like the SM site says -- while doing my reps, I keep an eye out for leeches. If I come across a bad one, or a set that are interfering (which seems to happen more and more as my character repertoire grows) then I take time out to invent some mnemonics for them. I'm intrigued by this Heisig method, though, and am looking it up. Thanks for the tip.

[realmayo]

sorry to return to this again.

my puzzle is this: I want to be able to test myself on how to memorise characters, but I also want to test myself on vocab, and I know that I'd learn the pinyin of plenty of vocabulary faster than I'd learn how to write the relevant characters.

in fact, in terms of what is quickest to learn, I guess the order would be:

1 sound (ie pinyin)

2 100% sure on tones

3 recognise the character

4 able to write the character.

and obviously the characters cause a problem that is absent in, say, an English speaker learning Spanish: in that case you could have a vocab test of English -> Spanish and Spanish -> English.

what I'm looking at now is three entries per character or word I want to learn (ie Q: -> A:):

a) character -> pinyin + english (testing reading, recognition)

b) english -> pinyin (testing vocab) though throw the character in here too

c) pinyin + english -> character (testing writing)

and, klortho unlike your method, not everything will be treated in all three ways -- for example I may not be too fussed about, at this stage, learning how to write certain rarer characters.

any flaws here? or suggestions?

it may sound petty to be spending a fair bit of time planning this rather than actually knuckling down and doing some work ... but I find supermemo pretty unwieldy to muck around with and don't want to have to start all over again when I realise I should have begun differently....

[leosmith]

a) character -> pinyin + english (testing reading, recognition)

B) english -> pinyin (testing vocab) though throw the character in here too

c) pinyin + english -> character (testing writing)

Well thought out. A couple comments.

1. can you change c) to: english -> character + pinyin, or is that too tough right now?

2. you mentioned vocab, so I assume you're going to do the same with combinations?

3. how about adding sentences that cover all the grammar you learn? and maybe sentences for vocab that usage isn't obvious? (hint - making CD's in addition to supermemo will really get you fast with your grammar/vocab)

Good luck!

Leo

[klortho]

it may sound petty to be spending a fair bit of time planning this rather than actually knuckling down and doing some work ... but I find supermemo pretty unwieldy to muck around with and don't want to have to start all over again when I realise I should have begun differently.

Whatever time you spend now preparing will be worth it, I think -- it's an incredibly long road to hoe. But, keep in mind that with Supermemo, if you find something's not working, you can always just start adding new elements according to a new scheme, and gradually move old elements into the new scheme as needed, or as they pop up.

a) character -> pinyin + english (testing reading, recognition)

B) english -> pinyin (testing vocab) though throw the character in here too

c) pinyin + english -> character (testing writing)

The problem with this is that (B) isn't very realistic except for a subset of characters with very specific English meanings. Most characters have more general meanings, and so the English question of (B) won't be enough information to let you guess just one character.

So, if my guess is right, you'll probably end up with just (a) and © for most characters, which is what I have. Then, you can throw in (B) - like questions for multiple character words, the more of those that you have for each character, the better. Anyway, just 我的意见

[realmayo]

cool thanks guys, useful suggestions.

leosmith, this will indeed be for what I'd term vocab (ie usually two-character combinations) as much as single characters. & I'll have to think about bunging grammar in there too some time, sounds like a good idea. what do you mean, making CDs?

I can't help thinking that what the world needs now is someone to design a better interface for supermemo, I find it a nightmare, always terrified I'll mess something up that will come back to haunt me, be they "childless objects" or whatevers lurking around.

I was also following that other thread recently about characters, people discussing various aids to memorisation. though apparently a bit childish, it does seem that constructing a story around each character, involving its component parts, should help.

[leosmith]

Whatever time you spend now preparing will be worth it

Totally agree! Don't be discouraged by members of this forum who give you a hard time for being too organized. They're just jealous:mrgreen:

The problem with this is that (B) isn't very realistic except for a subset of characters with very specific English meanings. Most characters have more general meanings, and so the English question of (B) won't be enough information to let you guess just one character.

I interepret (B) as the classic vocabulary drill "how do you say 'xxxxx' in Chinese". I think he's just throwing the character in the answer as a reminder. I have the same drill in Japanese. You are right that it's sometimes difficult to have a specific enough english description to guess the word, or the word doesn't translate well into english. In this case I drop little hints in the question. Another issue, sometimes there are many correct answers. In this case I give myself full credit if I get just one of them. Sometimes the word is highly dependent on grammar. In this case, I use a sentence. All of these exceptions add up to maybe 10% of the vocab in Japanese; I assume it will be similar in chinese. But even if it's a little more, the (B) drill (english to target language) really pays off regarding speaking, so it's worth the trouble to me. However, everyone learns differently, so maybe the difficulties aren't worth the trouble to you.

what do you mean, making CDs?

I used a program called audacity

http://audacity.sourceforge.net/

to record all my japanese grammar sentences. There are about 1000 sentences. The standard format for this is English-long pause-target-pause-target-pause for each sentence. Pauses need to be long enough for you to respond. Never stop the recording, just like pimsleur. The effect this has is to greatly increase your response speed, well beyond what supermemo alone will do for you. If you could get a native to record the chinese side, you could cut essentially a professional quality product that's perfectly taylored for you.

though apparently a bit childish, it does seem that constructing a story around each character, involving its component parts, should help.

That's interesting you thought it seems childish - exactly the same words used by Heisig. He was a little embarassed to show people the method, because of the childishness. Here's a link:

http://www.nanzan-u.ac.jp/SHUBUNKEN/publications/miscPublications/pdf/RK4/RK%201_sample.pdf

[realmayo]

yep leosmith, your interpretation is correct.

as for Heisig, if he had his Hanzi book out now I'd be interested but, going on what you've written elsewhere, Autumn this year is the earliest publication date. I'll make up my own stories instead: I've already got the gist of 20 or so radicals as well has recognising a fair few phonetic components, so I'll use these as the basis of quasi-etymological stories.

And this with the help of quite a cool book - "Cracking the Chinese puzzle" by TK Ann. Although I believe some of his etymology may be suspect his descriptions are still memorable. Often just little things, eg he says 世 "generation" comes from 卅 "thirty" ... this is the kind of stuff I like and remember.

[BC_01]

I did a search for a post on supermemo databases and found this one. In the original post there is a request to share supermemo databases but the following posts speak mainly to the use of supermemo itself.

As we know, databases could be shared as a txt file, a PDB file, or perhaps in another format like word/excel. So returning to the original question...

Who would be willing to share a database they have made?