Have you considered moving from Anki to SuperMemo?
- Thread starter Jon
- Start date
I am considering. What got me interested again is that I got a taste of the incremental reading workflow through the Anki plugin. I edited the plugin to try and make it feel more like Supermemo (for example to carry pictures from a parent note to an extract, along with several reference fields instead of a single one). But still, the scheduler in the plugin doesn't work and I can't dismiss notes without deleting them, so it is really a makeshift version of what I believe SuperMemo can offer.
I am also enticed by the knowledge tree and the concept maps and the promise of SM-17 being a superior algorithm and all that.
What deters me, though, is that I work on a MacBook pro and there are issues with rendering the fonts on the retina screen. Installing Windows and SuperMemo is not a problem (Parallels Dekstop), but once I launch the app the fonts in the UI are all blurry and pixelated. I am willing enough to try and adopt to SuperMemo's quirky UI but sacrificing the smooth fonts for reading is a deal breaker for me. Particularly because I have so much control in Anki, using CSS to style notes for comfortable reading -- (actually that is the only reason I got acquainted with CSS).
I am also enticed by the knowledge tree and the concept maps and the promise of SM-17 being a superior algorithm and all that.
What deters me, though, is that I work on a MacBook pro and there are issues with rendering the fonts on the retina screen. Installing Windows and SuperMemo is not a problem (Parallels Dekstop), but once I launch the app the fonts in the UI are all blurry and pixelated. I am willing enough to try and adopt to SuperMemo's quirky UI but sacrificing the smooth fonts for reading is a deal breaker for me. Particularly because I have so much control in Anki, using CSS to style notes for comfortable reading -- (actually that is the only reason I got acquainted with CSS).
I have used both Supermemo and Anki in my time. I eventually switched from Anki because while the Supermemo interface is quirky, I find its functionality to be far superior. Let's take just the ability to Dismiss without deleting. That is a big issue from my perspective. I am constantly dismissing items but I don't want them nuked from the Knowledge Tree.
Like yourself, I use Parallels Desktop. I have an iMac, but use Supermemo on my HP laptop. The font sizes were a little small on the iMac for my ageing eyes. So I looked into scaling them, and from memory there were issues with losing a bit of clarity if I did that on my iMac, while Windows had scaling functionality built in.
How bad is the blurriness?
Edit: There was some quoted (internal) research on the Supermemo site that stated the advantages of the SM17 algo over SM16 algo. There have been many iterations since the SM2 algo. The "feel" of it is that Supermemo cuts down frequent repetitions much faster than Anki. But consider that comment to be completely unscientific!
Like yourself, I use Parallels Desktop. I have an iMac, but use Supermemo on my HP laptop. The font sizes were a little small on the iMac for my ageing eyes. So I looked into scaling them, and from memory there were issues with losing a bit of clarity if I did that on my iMac, while Windows had scaling functionality built in.
How bad is the blurriness?
Edit: There was some quoted (internal) research on the Supermemo site that stated the advantages of the SM17 algo over SM16 algo. There have been many iterations since the SM2 algo. The "feel" of it is that Supermemo cuts down frequent repetitions much faster than Anki. But consider that comment to be completely unscientific!
Last edited:
Ordinarily, it is possible to suspend cards in Anki. Suspended cards don't show up for reviews so it's like dismissing them in SM. Unfortunately, the third party incremental reading plugin ignores that and still shows cards you suspended. So I am forced to delete them (otherwise the broken scheduler would keep showing me the same cards again and again)
I experimented with different scaling and resolution settings both in Parallels and Windows. I managed to improve the situation but it's still not good.
It seems that for users with retina screens it's a tradeoff between a better algo and SM functionalities VS blurry fonts. I imagine some people would rather put up with the blurry fonts. I wish I was one of them. If I were to commit to Incremental Reading I'd be spending a lot of time reading blurry text. I just can't bring myself to do it.
I experimented with different scaling and resolution settings both in Parallels and Windows. I managed to improve the situation but it's still not good.
It seems that for users with retina screens it's a tradeoff between a better algo and SM functionalities VS blurry fonts. I imagine some people would rather put up with the blurry fonts. I wish I was one of them. If I were to commit to Incremental Reading I'd be spending a lot of time reading blurry text. I just can't bring myself to do it.
I understand your frustration with blurry fonts. If you can use the better algo, the time saving on repetitions should be significant. There is an opportunity cost to not using the superior SM17 algo. And let us not forget, they keep improving the algo over time.
Now to the Incremental Reading...When you say you would spend a lot of time reading blurry text, what are you referring to exactly? Reading an entire book as blurry text? I just want to make sure we are on the same page before I answer further.
Now to the Incremental Reading...When you say you would spend a lot of time reading blurry text, what are you referring to exactly? Reading an entire book as blurry text? I just want to make sure we are on the same page before I answer further.
Yes, essentially I imagine myself doing almost all my reading in SuperMemo (apart from fiction for example). Is this common among seasoned users?
My thinking is this: if I can read in a way that improves my retention of the material (IR) -- then why would I ever make the choice of reading in a way that will lower my retention (ordinary reading)?
Well, because the fonts are nicer with ordinary reading, that would be a reason. A slightly unfortunate reason.
I think I will start a trial run of SuperMemo, using it in parallel with Anki. I'll keep my usual routine in Anki and build up a new unrelated collection in SuperMemo. I'll give it a few weeks to see what happens.
My thinking is this: if I can read in a way that improves my retention of the material (IR) -- then why would I ever make the choice of reading in a way that will lower my retention (ordinary reading)?
Well, because the fonts are nicer with ordinary reading, that would be a reason. A slightly unfortunate reason.
I think I will start a trial run of SuperMemo, using it in parallel with Anki. I'll keep my usual routine in Anki and build up a new unrelated collection in SuperMemo. I'll give it a few weeks to see what happens.
The reading of material from SuperMemo itself amongst heavy users is sometimes done directly from SuperMemo. However, there are many other heavy users who extract the material as they go. One particular user comes to mind where they do this:
While reading a pdf, extract each relevant part to CintaNotes using keyboard shortcuts > This provides a aggregation of the notes on one page> At the end of reading (the chapter, for example), copy to a mother topic in SuperMemo > Split this Mother topic into subtopics > When ready, cloze delete these subtopics.
It is a great workflow and one I may adopt myself, or at least elements of it. It lets you read from a pdf with its superior formatting, but still break the topics down for processing in SuperMemo.
While reading a pdf, extract each relevant part to CintaNotes using keyboard shortcuts > This provides a aggregation of the notes on one page> At the end of reading (the chapter, for example), copy to a mother topic in SuperMemo > Split this Mother topic into subtopics > When ready, cloze delete these subtopics.
It is a great workflow and one I may adopt myself, or at least elements of it. It lets you read from a pdf with its superior formatting, but still break the topics down for processing in SuperMemo.
That sounds like a potential solution!
But then again I like the idea of SuperMemo handling the schedule and letting me skip from article to article*. How do you replicate this with Cintanotes? Does it even remember where you left off or does this approach assume you'd read an entire pdf at once?
*as far as I understand Supermemo will even throw in a review or two here and there?
But then again I like the idea of SuperMemo handling the schedule and letting me skip from article to article*. How do you replicate this with Cintanotes? Does it even remember where you left off or does this approach assume you'd read an entire pdf at once?
*as far as I understand Supermemo will even throw in a review or two here and there?
Cintanotes is just used as a go-between to efficiently get the information from your pdf into SuperMemo. In fact, there is a problem with pdf's themselves in that they are difficult to get the text into SuperMemo. So, using the above described method, it acts as a go-between. Cintanotes is a bit like Notepad in Windows, but a bit more advanced with handy keyboard shortcuts for speed.
So if I go down this route, I would do a batch of extracting at one time. e.g. a whole chapter. Since you are only scanning through the chapter to decide what to extract, I believe it is faster than actually reading the entire chapter.
Please note, this Cintanotes method is not my idea but someone elses. They deserve all the credit.
So if I go down this route, I would do a batch of extracting at one time. e.g. a whole chapter. Since you are only scanning through the chapter to decide what to extract, I believe it is faster than actually reading the entire chapter.
Please note, this Cintanotes method is not my idea but someone elses. They deserve all the credit.
Thanks, I'll check it out. Polar bookshelf also looks interesting: https://github.com/burtonator/polar-bookshelf
Another potential solution I am considering, is to highlight the pdf and then, once I have finished either the chapter or ebook, I could then run it through something like https://www.sumnotes.net/
I haven't tried the service yet but it seems it can strip out all the pdf highlights into one document. That could be a big time saver. Then, you can paste that doc into a Mother topic and break down into separate flashcards, which are then cloze deleted. It is an amalgamation of both the CintaNotes method (but without the CintaNotes) and the sumnotes extraction. This way, you could - for example - read and markup the pdf file on an iPad (or phone), which could be a much more comfortable experience.
I haven't tried the service yet but it seems it can strip out all the pdf highlights into one document. That could be a big time saver. Then, you can paste that doc into a Mother topic and break down into separate flashcards, which are then cloze deleted. It is an amalgamation of both the CintaNotes method (but without the CintaNotes) and the sumnotes extraction. This way, you could - for example - read and markup the pdf file on an iPad (or phone), which could be a much more comfortable experience.