On 12 June 2010 07:52, Wim Woittiez <wim.woitt...@gmail.com> wrote: > That was a very interesting discussion on context dependency and > recalling cards in front of the computer, but not in the real world > etc. For the moment, I don't consider that a problem for myself. I'm > learning Spanish vocabulary among other things, and I do find the same > thing happening to me a little bit, but I consider it as a step in the > learning process.
Yes, I think the fact that this is seen as a problem may be down to possibly a misconception caused by the high utility and efficiency of SRS. After reviewing a card a few times (until its next review is maybe a month away), it's tempting to assume that you "know" the card. However, that's only one part of the learning process - more things have to happen for it to become useful in the way you ideally want it to work. Mnemosyne won't solve /all/ parts of the learning process, but it certainly gives a massive and systematic help. So if you're learning Chinese characters (like myself), you'll still need to actually read/speak/listen to the language, to take advantage of and "prime" the knowledge which is feeding into your brain through Mnemosyne. It won't just upload to your mind and automatically expand into a language-understanding/producing mechanism - you have to assemble such a mechanism from the thousands of building-blocks you're gathering via SRS, by reading, speaking, listening etc. > Perhaps somebody likes this: I also use a separate .mem for practicing > music, in this case a book with 99 keyboard exercises. I just made 99 > flashcards with only a number on them, and I practise whichever one > Mnemosyne tells me to. If I quickly get the hang of it again, I grade > it 4. If I can immediately play it smoothly, I grade it 5. If it > really takes some effort to get into, it's a 3. I make sure I have to > repeat 5-10 exercises each day, which takes me about 10-15 minutes. I > let Mnemosyne worry about the rest. Really happy with the results! That's a nice way to break it up. This is one of the areas where I think the minimum information principle is less productive. I generated about 450 cards representing two-bar chunks of the pieces I was playing for an exam. The cards looked like "Prokofiev: VF XI bars 14-15" with nothing on the back (or, rarely, notes about something I was consistently doing wrong). There was a card for each bar, so there was an overlap: if I only practiced bars 1-2, 3-4 then I wouldn't be fluent at the joins between bars 2-3, 4-5 etc. So when a card popped up, I'd find the right point on my sheet music and play those two bars a few times and mark it on how fluent my playing was - 1/2 for significant mistakes, 3- for mostly-acceptable and better. The problem was that the overhead of finding the right bars, getting into the right frame of mind and playing the excerpt a couple of times - a context switch, as it were - was quite expensive. On average, 40-60 seconds per card, and for difficult material sometimes twice that (especially as I got frustrated with how I was playing). So soon enough, I ended up with a rather large backlog which would have taken hours to clear, and gave up, going back to a simpler routine to prepare for the exam. However, the benefits of SRS are plentiful - the right amount of time is spent on each piece (or better, part of a piece). This is important, since in normal practice I often gravitate to the pieces I like most or find easiest. Next time, I'll break things up on a line-by-line basis, rather than bar-by-bar. I think. That might make the context switch relatively less expensive. Also, having cards for left/right hand only as well as together might be useful. > And, coming back to a previous discussion, perhaps a new idea. When we > learn just-failed cards, they sometimes appear multiple times in a > row, when there are almost no unlearned cards left. I don't care when this happens. I just answer it, even if it's "cheating" by just having seen it some seconds ago. If I really don't know it at all, then it'll fail tomorrow and taken care of anyway. It doesn't matter if a few snuck in "dishonestly" today - the right thing will happen automatically :) Oisín -- You received this message because you are subscribed to the Google Groups "mnemosyne-proj-users" group. To post to this group, send email to mnemosyne-proj-us...@googlegroups.com. To unsubscribe from this group, send email to mnemosyne-proj-users+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/mnemosyne-proj-users?hl=en.
