I miss incremental reading with a passion. I wholeheartedly wish more people would discover it. Maybe then more learning tools incorporating SRS and incremental reading would come to market. I’m tired of having to choose between the mobility of Anki and the incremental reading feature of SuperMemo (I’ve probably made the switch 4-5 times in the past few years).

Every few weeks, I google “incremental reading,” hoping more people have gained interest, but to no avail. Most people who have discovered its power continue to use SuperMemo contentedly. Most people try SuperMemo for a short time, are frustrated by the interface, and proceed with vigorous online criticism. Somewhere in the mix, incremental reading has been labeled an overrated selling point for SuperMemo that has no real value. This is extremely unfortunate.

The truth is in all my years of study, reading (professional and recreational), and language study, I have never discovered a more powerful tool than the SRS / incremental reading combination.

Anki’s iPhone / iPad app is probably the biggest step SRS software has taken in the last ten years, but I don’t believe Anki will implement incremental reading anytime soon. A plugin was designed a couple of years ago that toyed with the idea, but it seems that’s been abandoned. Anki’s designer said specifically that he doesn’t believe incremental reading is a feature his users are interested in and he has no immediate plans for such a feature.

In order of importance, these are the features my ideal learning tool would include:

• an SRS engine;
• incremental reading (ALL text formats: PDF, html, etc.);
• complete mobility and synchronization between devices;
• plugins for third-party reading and browser applications (e.g. Firefox, Kindle, etc.) and would automatically import highlighted text back to the program, along with source information; and
• complete media capability (audio / video)

Such a tool would revolutionalize learning. For a motivated learner, it would blow away ALL currently available  and orthodox forms of education.

A functioning and mobile incremental reading plugin for Anki? A reader with an SRS engine? A real mobile version of SuperMemo? It would be a start. I’ve been waiting for these developments for years now, but they don’t come. Where is my perfect learning tool? Do I have to design it myself?

Use any program extensively enough, and you’ll have ideas for improvement. Here is my own wish list for the next version of SuperMemo. I would love to hear more ideas…Who knows? Maybe we’ll get the attention of the SuperMemo development team.

• PDF conversion plugin. Reading incrementally is great–I’ll never go back to linear reading. Unfortunately, only plain text and html can be processed by current versions of SuperMemo. Most ebooks are in PDF-related formats. I probably spend as much time formatting poorly converted PDF text in SuperMemo as I do studying.
• Tab-separated value file (*.csv / *.tsv) import feature. SuperMemo allows Q&A format imports but that’s only good for text-only information. I (and I suspect many others) have moved beyond the text-only flash card format. Almost all of my elements contain some media, images at least and, in many cases, video or audio. I almost never import items into my collection (I much prefer to build knowledge from my own incremental reading and extractions), but a .tsv import feature would allow me to use output from SUBS2SRS, a program that creates video flashcards from movies (great for language acquisition).
• A reader. I read everything in SuperMemo: books, emails, news articles, etc. This post gave me an idea. Have you ever used Google Reader or subscribed to an RSS feed? Wouldn’t it be incredible if SuperMemo had this feature?
• My ULTIMATE wish: SuperMemo incorporates its own browser and/or email client. Think of the possibilities. Browse the internet in SuperMemo. Import a page or entire site to your collection to be read incrementally, or simply make extractions from web sites as you surf the web. All overhead of time and effort spent on importing eliminated. If you use an online email client (e.g. GMail), you could process your emails the same way (without having to use the Outlook workaround like in current SuperMemo versions). A browser that is your incremental reader and SRS software. Now that would be powerful. Combine this with a PDF-reading feature, and SuperMemo could literally revolutionize learning.
• Audio-to-text converter. This one is a little silly maybe (other programs already do this…Dragon Naturally Speaking, etc.), but it would a nice powerful feature for SuperMemo. There’s a LOT of great material out there that is ONLY available in audio, mostly lectures. I like learning from the audio format, but having a companion text in my SRS software would guarantee integration of knowledge. As it is now, I type notes as I listen, like I would if I was in a classroom.

Post your ideas!

Audio flashcards from your favorite film? No problem!

It’s been easy enough to create text/audio-only flashcards in SuperMemo using SUBS2SRS’s output. Two text and an audio component, import the audio file…done. I saved my first card as a template with objects, and introduced the .tsv file into my learning process to be incrementally processed. I’ll extract one movie line at a time, and apply the original template to new extracts. It will take 10-15 seconds to convert each extract into a text-audio-text flashcard (importing the audio file, formatting and placing the text, etc.), and I’ll likely get faster over time. Still far from ideal–a .tsv import function would eliminate the need for this whole process…but I’m satisfied with this workaround.

Audio flashcard for SuperMemo using SUBS2SRS output.

Video flashcards–not there yet.

My next challenge is to create a video flashcard. Anki users  have been doing this for a few years now. Using SUBS2SRS, and with a little preparation in Anki, users can play bits of movies or dramas (with target language subtitles) and Anki will quiz their listening/reading comprehension, using source language subtitles in the answer field. The value of video flashcards is immeasurable. They allow you to study using your contextual (video of person speaking), listening (speaker’s voice), and reading (subtitle) comprehension simultaneously.  They make study much more entertaining. 

Now that I’m over my initial frustration of SuperMemo’s inability to handle .tsv files, I’ve defined my mission in the simplest terms: to create an element that will play a short .avi file with sound. The element would also have two text fields–one that contains a subtitle for the video in my target language, and another with a subtitle in my source language. Only this last text field should appear at answer time–all the other components at question time.

First attempt at video flash card.

I’ve run into a couple of challenges.

1) My .avi file has sound. However, once it’s imported into the video component, it loses its audio. I’m assuming that whatever SuperMemo uses as its video player doesn’t handle audio. This is why I created a separate sound component that plays the audio portion of my video. The problem is I can’t seem to get both the video AND sound component to play automatically at question time. At the moment, only the video plays…I have to manually click the play button on the audio component.

2) The video component is playing my .avi file much too fast (maybe double speed?). It plays at regular speed on any of the players I have (GOM, WMP) so I know the file isn’t the problem.

I haven’t been able to find any pertinent information online, so I’ll just keep chipping away…

> I personally suffer due to the PDF pain as much as you, as I do need to
> read a lot of scientific papers that often come in PDF-only format. I
> adopted a number of strategies that depend on the type of article: from
> incremental copy&paste to visual learning (i.e. learning with picture copies
> of the text). The bad news is that the next version of SuperMemo does not
> make much progress in this area. However, I will yet once again forward your
> appeal to a colleague for a thorough technical re-evaluation.
>
> I have a sneaky suspicion that Adobe got a bit of vested interest in
> preventing an easy to-HTML conversion, or the problem is indeed technically
> insurmountable. However, a good plug-in editor from Adobe (like an IE’s
> TWebBrowser uses by SuperMemo) would provide a way around. Last time I
> looked into that issue, we did not have that option.
>
> Please rest assured that, as a user myself, I won’t give up on this
> subject, even though it has been dragging for a couple of years now.

For those of you interested, members of the SuperMemo Yahoo Group have posted their own workarounds to importing PDFs. After trying everything under the sun, I have a few tricks of my own, but they all involve a lot of work (mostly formatting issues). Depending on the type of PDF, I might use ABBYY FineReader, a Word conversion, or a simple copy-and-paste, or a combination thereof. However, what is really needed is a better .pdf-.txt (or .htm) conversion software (that more accurately retains formatting, images, etc.) or a PDF import plug-in for SuperMemo.

For those of you unfamiliar with “subs2srs,” it’s an incredible program for language acquisition. Feed it a video file, along with subtitle files (target language and first language), and it will essentially output flashcards that either have a video or audio clip with screenshot (your choice) in the question portion, and the corresponding subtitle material in the answer.

Well, it doesn’t actually create flashcards but rather multiple media files and one master file of tab-separated values (.tsv) that will point the SRS program in all the right directions (where to find each video clip, audio clip, and screen shot). Ultimately, every spoken line in the processed video becomes an individual flashcard, or item in SuperMemo terms.

I’m posting this here because of the sparsity of information online regarding this topic. I was up until 6 a.m. this morning working on a way (and trying to find information online) to make this work. If had read this post, or one like it, I wouldn’t have bothered (OK, I might have)–so I hope I can save someone that heartache. Most importantly, I hope this post reaches someone who has found a workaround since this topic was last addressed AND/OR in case someone knows whether .tsv import will be supported in the next version of SuperMemo. Get in touch with me! As far as I know, every other SRS (Anki, Mnemosyne, etc.) has .tsv import as a basic feature.

In the meantime, I’ll be taking the “subs2srs” output files that I have and manually creating SuperMemo items using the media / subtitles it has parsed. Better than nothing but, again, like the currently available solutions to the PDF challenge (finding a way to get PDF material into SuperMemo),  it’s a hugely inefficient workaround.

At times like these, I really wish SuperMemo would become popularized (like Anki) so that a community of programmers would take the initiative and work on plug-ins, etc…

P.S. Another frustrating challenge I encountered during this experience, and one with very little information online, is successfully converting Korean .smi subtitle files to a format that subs2srs recognizes. I used Subtitle Workshop, an amazing program. However, although Subtitle Workshop displayed the Hangul just fine, the converted .srt files appeared as junk in subs2srs. After much head-banging, I found the problem/solution: you need to change the encoding to Korean under “Advanced Options” in subs2srs. The same goes for GOM Player, by the way (play with the advanced options). Again, I hope to save others this time and frustration.

In this article, Krashen proposes that reading narrowly (i.e. books in one particular field, or by one particular author) is more beneficial to language acquisition than reading widely. By reading narrowly, we are repeatedly exposed to a certain vocabulary set and we gain a background knowledge for further reading in that field.

Most EFL readers follow the opposite principle. They feature articles on a variety of topics from as diverse a range of fields as possible. The idea is that this exposure to variety will endow students with a broader vocabulary set and a well-rounded grasp of the language. However, if students truly learn to read by reading, and reading should be pleasurable (Comprehension Hypothesis), then these textbooks are highly ineffective.

Every student has different interests. A textbook with ten articles on ten different subjects will likely contain 7-8 articles that a student has no interest in. This has certainly been the case for me. Those learning a language should select reading that interests them, with very little regard for level or variety.

This year, that obsession is SuperMemo–incremental reading, in particular.

I’ve been using SuperMemo for about five years now. For those of you unfamiliar with SuperMemo, it’s a Spaced Repetition System software…the origninal SRS software, in fact.

SuperMemo’s promise is that, if the software is used daily, the knowledge in your brain will reflect at least 95% of your SuperMemo knowledge collection. It fulfills this promise by forcing you to actively recall information just when that information would normally begin to “decay” in your brain (out of misuse). Luckily, the interval between the moment of “decay” becomes longer with each repetition. Knowledge becomes more firmly implanted in our brains with each “recall.” The challenge is recalling knowledge frequently enough to keep it there, but as infrequently as possible (more time to learn new stuff!). SuperMemo makes this calculation and gives you the nudges you need.

In addition to these standard SRS features (you can find several other, easier-to-use programs that work on this principle), SuperMemo has another component that is particularly addictive: incremental reading. With this feature, you may import any number of text or html files to SuperMemo, of any size, and read them incrementally and in parallel.

Knowledge evolves in SuperMemo. An imported book is broken up into important pages. Noteworthy paragraphs are highlighted and processed separately. Paragraphs become sentences, and sentences ultimately become actively-recalled items (question-answer pairs or cloze deletions). It all sounds like glorified memorization until you try it yourself.

I’m enamoured with SuperMemo’s incremental reading for several reasons.

• I’m endowed with confidence that my reading is purposeful. Not only does this confidence make reading more enjoyable, it also increases my sense of responsibility in choosing worthwhile material. I will never forget the important ideas contained in a book I’m reading. That feels good.
• Incremental reading allows me to read widely. One of the realizations you come to early in your experience with incremental reading is that linear reading, or reading one book at a time and working your way through a “reading list,” is vastly inferior to reading books and articles simultaneously, in parallel. It’s more efficient to read in parallel. It’s more entertaining to read in parallel. Most importantly, one’s power to associate seemingly disparate ideas is increased by reading in parallel. Worried about when you’ll get around to that book on your reading list? I don’t. I import it into my knowledge collection and read it in tandem with my other reading, as quickly or slowly as I like.
• Remembering the ideas and principles I encounter in reading allows those ideas to truly integrate with my own ideas and experiences. They become part of me, instead of “going through one ear and out the other.” I can weigh an idea I learned ten years ago against one I learned yesterday, despite elapsed time. Incidentally, this is also the best counter-argument against those who would question the value of memorization in an age when information is at our fingertips.

SuperMemo 2006 has become freeware. Try it, and revolutionize the way you look at reading and knowledge.

I was back in Korea, and it was happening – a wedding ten years in the making.

My wife and I divided the workload — I would handle my mother and sister’s visit and our honeymoon, and she the wedding planning. I think I speak for most guys when I say this is a choice arrangement.

A typical Korean wedding has several hundred guests. I remember what 500 people look like from my brother-in-law’s wedding a few years back. Beyond the challenges common to any large wedding, the Korean bride and groom are expected to visit every guest table during the reception. 500 guests translate to 50 table visits. Needless to say, it was with my full approval that my wife and mother-in-law decided our wedding would take place at a smaller venue (smaller weddings are known as “House Weddings”). We limited invitations to 300, of which we expected roughly half to show.

Only half? Are you sure? Yes. So sure, in fact, that the contract with the wedding hall was for approximately 150 guests. I was aghast. What if more show? And besides, why are we inviting people who we know won’t come?

Like everything else in Korea, weddings have a system.” There is a right way to do it, and everyone does it that way. It’s the best way, so we do it that way.” These enigmatic answers are the best you can expect when questioning the very specific processes that  characterize the Korean way, whether it’s getting married, buying a car (white, black, or silver?), opening a business, raising children, shopping for clothes, or an outing with friends. Everything is done in a very specific way, with very little room (or tolerance) for deviation. It can be painfully annoying, but when you’re the minority complaining YOU are being annoying. But I digress.

Why only half?

A wedding invitation in Korea is essentially an invitation to deposit money into a ‘wedding account’, opened by the bride’s and groom’s parents specifically for that purpose. It’s understood that attendance is optional (although relatives are really expected to come). The wedding account balance is the parents to keep, who are footing the bills after all. Oh, didn’t I mention that? Traditionally, the bride’s parents pays for the wedding, and the grooms’ for the newlywed house (or down payment, 전세, at least). Believe it or not, the cash gifts may actually outweigh these costs of the wedding. What kind of racket do these parents have going on?

Keep in mind that most of the people receiving wedding invitations from my wife’s parents, including relatives, long-time friends, co-workers, and peers, have children of their own. These children have gotten married, or will get married, at which time my wife’s parents would or will have made their own cash donations. So, when your child gets married it’s time to get that money back. Add that to the already considerable number of reasons Korean parents pressure their children to marry and, in our case, why we felt so pressured to be married here and not in Canada. This is another example of how a marriage here goes so far beyond the union of two people. If found the family considerations stifling at times.

As usual, my being a foreigner put everyone in an awkward position. Would my parents be buying a house? What would my family’s role in all this be? It saddened my wife’s parents that they were being deprived of the full wedding process. We had already foregone many of the pre-wedding traditions, such as multiple meetings and gift exchanges between the two sets of parents.

I come from a small family to begin with, and my father had passed just months prior. My wife’s parents dream of SY marrying into a large, supportive family had been dashed the moment they had agreed to our marriage.

What I expected to be an overwhelmingly happy occasion in my life was being overshadowed by alternate, but equally unwelcome feelings of guilt and defensiveness.

Unfortunately, a gloved, waving hand greeted me as I boarded the bus and I immediately began sending apologetic text messages. Sure enough, the trip took 10 minutes longer than most.

My experience today isn’t the norm, however. The Korean bus driver will almost definitely get you there on time. 

In the daily effort to meet what I can only guess must be a ridiculously tight stop schedule, the Korean bus driver probably runs 50 red lights, narrowly misses running over a dozen pedestrians, cuts off at least as many cars, and causes two dozen passengers to fall while boarding or disembarking. The buses here will affect you somehow, either as pedestrian, passenger, or driver. 

My own experiences with Korean buses have inspired the following list of driving practices that seem common to all Korean bus drivers. For all I know, they may even be part of Korea’s Ministry of Transport’s training program.

• Turning lanes are actually express lanes. When traffic is stopped, use these express lanes to bypass law-abiding drivers. Cut back into the regular lane (or even all the way to the opposite turning lane!) at the intersection and wait confidently for the light to change.
• Increase speed at red lights. Otherwise, crossing pedestrians and traffic may hinder your progress.
• Do not stop the bus in designated passenger pick-up / drop-off areas. Stopping in these areas will mean more passengers successfully boarding your bus. Stop far enough away from groups of passengers so that only the strongest in the pack make it to the door in time.
• Randomly fail to make stops. Particularly effective if you’re behind schedule, this practice requires that you ignore shouts from passengers attempting to board or disembark. See next.
• Subject your passengers to a constant stream of 60s / 70s pop songs. MOT studies have shown that Bee Gees b-tracks or Beatles remixes have a calming effect on passengers, particularly in life-threatening situations.
• Use the bus’ climate controls to make passengers as uncomfortable as possible. A good guideline is summer = full a/c, winter = full heating, regardless of outdoor weather conditions. If one is available, equip yourself with a bus whose windows do not open for maximum effectiveness.
• Do not stop the bus for boarding or disembarking passengers. This only results in lost time. Never allow the bus to come to a complete stop – otherwise, passengers may seize the opportunity to ask a question. Passengers must not interfere with your primary objective – purposeless speed.