On Tue, Sep 8, 2009 at 6:05 AM, OldGrantonian<[email protected]> wrote: > > I'm a big Mnemosyne fan, but I've never been happy about the treatment > of new difficult words that have a steep forgetting curve. Some words > initially refuse to stick in my memory. > > As an example of difficult words, here are some capitals of the world > that I would never be able to memorize using Mnemosyne alone. I need > to use my own method outside Mnemosye (known as "Der Sledgehammer") > before eventually importing them to Mnemosyne: > > Yamoussoukro > Ouagadougou > Tegucigalpa > > BTW: I apologize in advance if I offend any person who is associated > with any of these towns. > > The current "Not memorised" group does not solve this problem. (I've > always thought that the "Not memorised" group seems to be a well- > meaning "hack" rather than an implementation of a sophisticated > algorithm. No offence intended :) ) > > Here are two reasons why the current non-memorized group does not work > for difficult words: > > 1) Non-memorized words are pushed to the back of the queue. In other > words, scheduled words are presented before non-memorized words. I > assume that the steep initial forgetting curve of difficult words can > reach zero before the word is re-presented from the "Not memorised" > group. > > 2) Currently, the smallest unit of scheduling time is 1 day. So any > "Yes" answer causes the card to be pushed to at least the next day. I > know from bitter experience that if I get "Tegucigalpa" correct today, > there is no guarantee that I will remember it tomorrow. Again, I > assume that this is because the initial "forgetting curve" for > "Tegucigalpa" is so steep that it reaches zero in less than 1 day. (I > searched the SuperMemo articles for any reason why the minimum > scheduled interval is 1 day. No reason was given.) > > I need the scheduled intervals to be much shorter for such words - > measured in minutes or hours, rather than days. Eventually, of > course, the interval would become multiples of 1 day. > > I'm wondering if it would be difficult for an idiot such as myself to > use a plugin to specify the smallest unit of time that is to be used > by the SM2 algorithm.
I agree that the algorithm could probably be better for early periods of had questions; but I think I've expounded before my view that you can work around these issues by asking multiple cards for each hard question. eg. it's had to remember that Parmenides was born in 0515 BC, but it's much easier if you generate 10 or so Cloze deletions like 'Parmenides was born in __15 BC' (see my previously emailed scripts), or have multiple choice questions. (For cities, where one reads letters in groups and not number by number, I suppose one would want to delete groups of vowels or consonants, or maybe just delete by halves.) As the SuperMemo site says, if the question is too hard, you need to break it down... Have you tried these techniques? They might save you time compared to whatever your sledgehammer approach is. -- gwern --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "mnemosyne-proj-users" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [email protected] For more options, visit this group at http://groups.google.com/group/mnemosyne-proj-users?hl=en -~----------~----~----~----~------~----~------~--~---
