+- kanji koohii FORUM (http://forum.koohii.com)
+-- Forum: Learning Japanese (http://forum.koohii.com/forum-4.html)
+--- Forum: Learning resources (http://forum.koohii.com/forum-9.html)
+--- Thread: Mighty Morphin Morphology (/thread-7486.html)
Mighty Morphin Morphology - overture2112 - 2011-08-26
nest0r Wrote:And when I study, I just choose ip1known or ip0mature to toggle between vocabulary word cards and video comprehension cards.Minor issue that occurred to me: if a i+1 known vocab card becomes i+0 known, the selective study of ip1known doesn't review it. Of course, after doing ip1known (vocab) and ip0mature (comprehension), you can just study the remaining due cards without selective study on (which should just be the i+0 known ones).
Mighty Morphin Morphology - nest0r - 2011-08-26
overture2112 Wrote:So it will remove the ip1known? I thought that tag was kind of its own separate thing.nest0r Wrote:And when I study, I just choose ip1known or ip0mature to toggle between vocabulary word cards and video comprehension cards.Minor issue that occurred to me: if a i+1 known vocab card becomes i+0 known, the selective study of ip1known doesn't review it. Of course, after doing ip1known (vocab) and ip0mature (comprehension), you can just study the remaining due cards without selective study on (which should just be the i+0 known ones).
Actually, did I set this up wrong. I only have MorphMan monitoring my subs2srs deck's Expression field, but I think I was assuming that somehow MorphMan would know when that vocabExpression became i+0 mature because it was monitoring it for that, since that's the field it's told to auto-fill. Like it was monitoring Expression field in general, and vocabExpression for maturity, and sentenceExpression for known?
So I should be having it monitor both Expression and vocabExpression? Does that mess with the sentenceExpression, then, as it fills the sentenceExpression with the sum, e.g. hypothetically it would fill sentenceExpression with both Expression and vocabExpression?
So if the vocabExpression isn't being monitored for known-ness, but only for maturity, then I guess the Expression would never have its single unknown updated by the vocabExpression, but it would get updated as i+0 mature eventually. Perhaps when that happens the 1 unknown can be added to the known.db? Or periodically the mature.db can update the known.db?
Otherwise the only time an Expression would become i+0 known from i+1 known is when another card you're reviewing differently has the same morpheme and becomes known, in which case it would actually be okay to drop the original ip1known card from the schedule via SelectiveStudy not noticing its (removed?) tag, since you'd be learning it to i+0 maturity from elsewhere (assuming you're not doing the same thing in the other deck, then I guess they'd cancel each other out? ;p).
Also, so you're saying that if vocabExpression or Expression then become i+0 known rather than i+0 mature, it removes the ip1known tag? Because maybe it should just leave that tag till it's i+0 mature, right, since only the initial focus is on it being ip1known, after that it's whether it's ip0mature, unlike other fields/cards where the iPlusN known-ness is its own spectrum.
Mighty Morphin Morphology - overture2112 - 2011-08-26
nest0r Wrote:Also, so you're saying that if vocabExpression or Expression then become i+0 known rather than i+0 mature, it removes the ip1known tag? Because maybe it should just leave that tag till it's i+0 mature, right, since only the initial focus is on it being ip1known, after that it's whether it's ip0mature, unlike other fields/cards where the iPlusN known-ness is its own spectrum.It was to avoid new cards which are already i+0 known, but I was dumb and forgot selective study lets you change behavior for reviews and new cards separately. So what you really want are tags:
i+0 mature => 'ip0mature'
i+1 or i+0 known => 'ip0or1known'
[maybe have this?] otherwise => 'ip2+known'
Then use selective study to show only ip0mature or ip0or1known tagged cards for reviews and hide ip0mature cards for new.
Also, I'd appreciate if you have any ideas as to better names for those tags. The ones above are accurate but feel inelegant. 'SentenceCard', 'VocabCard', and 'TooHard' or something are a tad cleaner but I want to keep things somewhat general. Maybe I'll add it to the configuration but I want a sane default.
Mighty Morphin Morphology - nest0r - 2011-08-26
overture2112 Wrote:What about All Mature or Completely Mature or Fully Mature for ip0mature, 0-1 Unknown, and 2+ Unknowns? Then people can change defaults to different stuff like Mature, Comprehensible, and Incomprehensible, or whathaveyou.nest0r Wrote:Also, so you're saying that if vocabExpression or Expression then become i+0 known rather than i+0 mature, it removes the ip1known tag? Because maybe it should just leave that tag till it's i+0 mature, right, since only the initial focus is on it being ip1known, after that it's whether it's ip0mature, unlike other fields/cards where the iPlusN known-ness is its own spectrum.It was to avoid new cards which are already i+0 known, but I was dumb and forgot selective study lets you change behavior for reviews and new cards separately. So what you really want are tags:
i+0 mature => 'ip0mature'
i+1 or i+0 known => 'ip0or1known'
[maybe have this?] otherwise => 'ip2+known'
Then use selective study to show only ip0mature or ip0or1known tagged cards for reviews and hide ip0mature cards for new.
Also, I'd appreciate if you have any ideas as to better names for those tags. The ones above are accurate but feel inelegant. 'SentenceCard', 'VocabCard', and 'TooHard' or something are a tad cleaner but I want to keep things somewhat general. Maybe I'll add it to the configuration but I want a sane default.
Mighty Morphin Morphology - nest0r - 2011-08-26
So I can assume I don't need to worry about monitoring the vocabExpression/sentenceExpression in the Fields to check?
Edit: On a side note, is it possible to have a whitelist option for Morph Man in addition to blacklisting? So you could show only nouns, for example?
Mighty Morphin Morphology - overture2112 - 2011-08-26
nest0r Wrote:So I can assume I don't need to worry about monitoring the vocabExpression/sentenceExpression in the Fields to check?Sure thing.
Edit: On a side note, is it possible to have a whitelist option for Morph Man in addition to blacklisting? So you could show only nouns, for example?
Mighty Morphin Morphology - nest0r - 2011-08-26
So just until I end up writing more extensively on my strategies for the wiki, I should mention the way I'm tinkering with word search (シークワーズ) at the moment (via: http://www.armoredpenguin.com/wordsearch/) is to display the desired morphemes (such as knowns) in the GUI, then sort them by length here: http://textmechanic.com/Sort-Text-Lines.html , then select a batch of the desired minimum lengths, save to .txt (I think it must be UTF-8), and import into the generator. After that you can save as .html, .pdf, or what I prefer is snipping it with the snipping tool and pasting into Windows Journal so I can draw lines through the terms.
Haven't really played with the crossword generator (http://www.armoredpenguin.com/crossword/) yet, despite all this time recommending it. Since it requires clues, I think that'll be trickier. But likely I can just use clozed sentences, going from Rikaisan's save text or Anki to a .txt with a vocabulary word and the meaning or clozed sentence on each line.
Mighty Morphin Morphology - nest0r - 2011-08-27
Oh and check this out, I eliminated everything but the following and the import still worked:
<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
<vhdatabase.sqlite xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance">
<vocabulary><word>morpheme1</word>
</vocabulary>
<vocabulary><word>morpheme2</word>
</vocabulary>
<vocabulary><word>morpheme3</word>
</vocabulary>
</vhdatabase.sqlite>
(I saved that as an .xml and imported with no problems.) Technically the original format has the <vocabulary> and <word> on separate lines, but that doesn't seem to matter. /n00b
Edit: Know what else is weird? None of those numbers/punctuation show up in my known.db single column Morph Man list. I don't get how they got in there in the first place. I must have messed something up before. This time I didn't have to do anything to import the list. Literally all I did was show them as single column, copy/paste into UltraEdit and replace ^ with <vocabulary><word> and $ with </vocabulary</word> and copy/paste that into the above format, and voila!
Edit 2: I guess this is where having the option to show inflected forms would be nice, since that'd fit reading material better.
Mighty Morphin Morphology - overture2112 - 2011-08-27
nest0r Wrote:Edit: Know what else is weird? None of those numbers/punctuation show up in my known.db single column Morph Man list. I don't get how they got in there in the first place.The gui database viewer has a blacklist field at the bottom right, which defaults to punctuation and something. The i+N caluclation and unknowns fields are filled with everything though (no black/whitelist yet).
Edit 2: I guess this is where having the option to show inflected forms would be nice, since that'd fit reading material better.
I might need to do some refactoring to get inflected forms to be kept track of properly- I'll try to avoid changing the database format if possible though (then again deleting all your dbs and having it recalculate everything is pretty effortless now though).
Mighty Morphin Morphology - nest0r - 2011-08-27
Yeah, I thought it was strange that the 。 was listed as a 記号 on that Mecab page but it was still in my list. I must've done something weird before to have numbers and punctuation in the mix.
Edit: On another side note, how would a person add a non-Anki list of words they know and merge it with the known.db, or is that even possible without the card information. Like, could they extract/save the list as a .db from the GUI, then open it and the known.db as A and B and show their Union, then save as a new known.db to replace the old one? Or I imagine the known.db has its own special formatting so you can't easily create a new one. Maybe you'd have to import the list as a deck and set its known threshold to 0 in the .cfg, so it gets appended?
Mighty Morphin Morphology - overture2112 - 2011-08-27
nest0r Wrote:On another side note, how would a person add a non-Anki list of words they know and merge it with the known.db, or is that even possible without the card information. Like, could they extract/save the list as a .db from the GUI, then open it and the known.db as A and B and show their Union, then save as a new known.db to replace the old one? Or I imagine the known.db has its own special formatting so you can't easily create a new one.The known.db is just an ordinary db. You could put the words in text file, import it into a database using the gui, the merge it with known.db (and save as known.db).
Mighty Morphin Morphology - nest0r - 2011-08-27
Great! Okay, last feature request to note then I'll shut up for a while. ;p
Since sometimes words are written in hiragana instead of kanji, it could be useful to display readings as hiragana? Optionally instead of katakana, perhaps? And if there's an export function related to the vocabulary.xml, hopefully that can let you choose whether to export said column (as a word entry in itself).
Edit: Maybe that'd be a bad idea though. I don't know. I'm not good at future-proofing.
Mighty Morphin Morphology - nest0r - 2011-08-27
Okay, I'm beginning to think that (inflected form) known morphemes is good for general browsing stuff, without hiragana readings or even katakana words (to prevent false positives as they overlap), and there's utility in also being text-specific to find unknowns and highlight those when reading a particular novel and the like.
Yep.
Mighty Morphin Morphology - nest0r - 2011-08-27
... Now that I've actually started reviewing these new super cards. Something has occurred to me. The intervals are still the same. So if the monitored morphemes become i+0 mature and the Expression field is used to fill the sentenceExpression and I review that ip0mature card... the intervals will be very large, no? Even though it's effectively a new card.
I only noticed this because I have my extemp vocab deck set up to fill a productionExpression with the vocabulary word when it's mature, but even though it's now fresh to me, the interval is months long. I guess when doing selective study, I can just fail all the cards tagged ip0mature?
Edit: Haha, no that doesn't work because it empties the filled field and the tag. Well... uh, damn.
Mighty Morphin Morphology - overture2112 - 2011-08-27
nest0r Wrote:... Now that I've actually started reviewing these new super cards. Something has occurred to me. The intervals are still the same. So if the monitored morphemes become i+0 mature and the Expression field is used to fill the sentenceExpression and I review that ip0mature card... the intervals will be very large, no? Even though it's effectively a new card.Yes, the initial interval would be high. But then again, I figure the comprehension cards are like long term safety checks since you're not really learning anything new, so most likely a large initial interval is fine? If not, we could always manipulate the interval a bit after promoting.
Mighty Morphin Morphology - nest0r - 2011-08-27
overture2112 Wrote:Well for me, the whole point of making a card a production card or comprehension card is to now go on developing those skills without having to worry about knowing the lexical information, which was developed in a focused, transfer-appropriate fashion (recognition cards for recognition, production for production) and/or as a pre-learning stage to reduce overhead for these new cards. The change in format based on whether I know the vocabulary within the card makes it a completely different card type where I'm not grading the card based on vocabulary knowledge, but instead the spaced retrieval is now developing this new skill in a focused way that has its own learning curve beyond just one retrieval every x months.nest0r Wrote:... Now that I've actually started reviewing these new super cards. Something has occurred to me. The intervals are still the same. So if the monitored morphemes become i+0 mature and the Expression field is used to fill the sentenceExpression and I review that ip0mature card... the intervals will be very large, no? Even though it's effectively a new card.Yes, the initial interval would be high. But then again, I figure the comprehension cards are like long term safety checks since you're not really learning anything new, so most likely a large initial interval is fine? If not, we could always manipulate the interval a bit after promoting.
So having the schedule adjusted so it's like a new card (similar to if you were to unsuspend a card of the same fact with a different template) would be great, but wouldn't that affect the maturity? (Actually I think not, I tried rescheduling cards past the mature threshold to fake maturity and it didn't work.)
Also, if failure means the card gets demoted from i+0 mature (depending on whether there are other cards with that morpheme that are also mature), that means you can never fail and retest a card as that new type of card, right? Instead if you fail a video clip comprehension card that you're grading based on your listening + parsing, you have to retest it as the old format (e.g. recognition vocabulary card) that you reached maturity with, except with unnecessarily small intervals since you technically didn't fail it in that format/criteria and probably still know it well in the old format/criteria.
Edit: I just thought of a temporary solution in a completely different direction using the unmatures field and having a conditional message when it's empty, but it goes back to having 2 templates and doing manual unsuspension.
Edit 2: By the way, if you're curious about what I meant in the past about the region of proximal learning, it's that studies show that when students are studying in any given session, they tend to do better and be more motivated/persevering when they study the easiest items first (easiest of materials they haven't learned, so in other words, easy but not too easy i.e. difficult but desirably difficult). This becomes more efficient as well when you're talking about time constraints (similar to answering problems on a test that you have a chance of answering rather than mulling over questions where you draw a blank).
Mighty Morphin Morphology - overture2112 - 2011-08-27
nest0r Wrote:Also, if failure means the card gets demoted from i+0 mature (depending on whether there are other cards with that morpheme that are also mature), that means you can never fail and retest a card as that new type of card, right? Instead if you fail a video clip comprehension card that you're grading based on your listening + parsing, you have to retest it as the old format (e.g. recognition vocabulary card) that you reached maturity with, except with unnecessarily small intervals since you technically didn't fail it in that format/criteria and probably still know it well in the old format/criteria.Good point. Perhaps it's best to use separate cards then, having morph man just add i+0 mature tag and i+0 known tag so you can effectively control the enabling of cards via selective study.
That way if you fail a comprehension/sentence card it has no effect on the vocab card. That said, what if you fail a vocab card? Should it fail the sentence too? With the above, if you fail all vocab cards with some morpheme it would take away the i+0 mature tag.
1) If you have the selective study review filter only show you "i+0 mature" tagged sentence cards, it would effectively disable the card until the component morphemes are all mature again but that could be bad since it could become very past due by the time that happens.
2) You could use selective study to only show "i+0 mature" tagged new cards and any sentence cards for reviews, with the downside that it may be rough as your reviews aren't necessarily i+0 mature anymore.
I'm leaning towards 2. Also, it seems Anki doesn't let you specify via selective study that a card must have tag A and tag B, but I think something like "show only i+0 mature and hide vocab card" or "show only i+0 known and hide sentence card" works?
Mighty Morphin Morphology - nest0r - 2011-08-27
I think that #2 would work, yeah. I guess having the template is kind of necessary to ensure they don't compromise the tags/SelectiveStudy once thresholds are passed.
It's not too big a deal, I think, if you reach maturity with one card and take that opportunity to start reviewing it in a new context, because then you're still getting reinforcement, and simply failing a vocabulary card I think won't indicate a complete failure of your knowledge, so much as a bad day or whathaveyou. By the time you fail a mature vocab card upon the next review after it passes the mature threshold, you'll have reinforced it in a new context multiple times with the new sentence card and its smaller intervals (decreasing your likelihood of even failing it, actually).
Mighty Morphin Morphology - nest0r - 2011-09-05
So to sum up as I've experimented, at the moment what I've done is, for example, I have my vocabulary deck using two templates: Recognition and Production, with the Production cards suspended. Morph Man 2 monitors the field holding the kanji form of the word. I have a conditional field set up on the Recognition cards so that when the unmatures field is empty, I get a message telling me it's mature and I can now unsuspend that fact's corresponding Production card. That's for per card notifications mostly for new cards, as I also went into the browser and unsuspended all Production cards that were tagged ip0mature. Then I just switch between templates using SelectiveStudy. I do similarly with the subs2srs deck and its Vocabulary/Comprehension cards.
So what the eventual change would do regarding how MorphMan auto-tags cards, is it would allow me to leave everything unsuspended and not worry about the conditional unmature field and just funnel it through the SelectiveStudy.
I've given up on getting something nice set up for RTK users who use Anki and MorphMan, where they could get a notice when the kanji are mature, since MorphMan doesn't specifically handle/monitor the RTK kanji.
Mighty Morphin Morphology - gwedig - 2011-09-14
I love this tool. It has made finding those near-comprehensible low hanging fruit sentences *so* easy, it's great. And it's really addictive to go in and sort my deck based upon i+N and see how many new cards I can unsuspend. I probably do it more frequently than I should...
One feature that would be a huge help to me would be to have more than one iPlusN field. I'd like one for known and one for mature. That way I can sort out the 'easy' and the 'very easy' cards more systematically. Is there any chance of getting something like this?
Thanks for putting this tool together.
Mighty Morphin Morphology - overture2112 - 2011-09-14
gwedig Wrote:One feature that would be a huge help to me would be to have more than one iPlusN field. I'd like one for known and one for mature. That way I can sort out the 'easy' and the 'very easy' cards more systematically. Is there any chance of getting something like this?Yea, I'll expose the mature i+N and known i+N data separately. I should have an update sometime this weekend with a handful of suggestions that have been building up.
Mighty Morphin Morphology - andrewkun - 2011-09-15
Amazing stuff! Thanks!
One thing I am having trouble figuring out: after the job runs my cards are tagged with the various fields but the study order does not seem to have been modified. Is there a manual step I need to take to make that happen?
Mighty Morphin Morphology - overture2112 - 2011-09-15
andrewkun Wrote:Amazing stuff! Thanks!It should modify the creation times, so if your study options are set to 'show new cards in order added', it should work automatically. Is this not the case?
One thing I am having trouble figuring out: after the job runs my cards are tagged with the various fields but the study order does not seem to have been modified. Is there a manual step I need to take to make that happen?
Mighty Morphin Morphology - andrewkun - 2011-09-16
I'm pretty sure it's not. I verified by adding the tagged fields to my card's display and went through a few new cards.
I see this card as my first new card:
new morphemes: 3
morphman index: 41048
vocabrank: 48
and then this card:
new morphemes: 0
morphman index: 8000
vocabrank: 0
The deck is set to: "Show new cards in order added"
Would love any suggestions! (I tried checking and optimizing my DB which didn't seem to help)
Mighty Morphin Morphology - overture2112 - 2011-09-16
andrewkun Wrote:I'm pretty sure it's not. I verified by adding the tagged fields to my card's display and went through a few new cards.Strange. Can you verify that Morph Man completed updating the deck (maybe it's big and didn't finish)? You can check the last update time in the configuration window.
I see this card as my first new card:
new morphemes: 3
morphman index: 41048
vocabrank: 48
and then this card:
new morphemes: 0
morphman index: 8000
vocabrank: 0
The deck is set to: "Show new cards in order added"
Would love any suggestions! (I tried checking and optimizing my DB which didn't seem to help)