We all use Anki for different purposes, but we all follow a similar process in our study. We have source material, we take notes, and we convert the results of our reading or study into Question-Answer pairs to drop into Anki comfortable in the knowledge that this incredible SRS program will help us retain knowledge in the most effective and efficient method possible.

An Anki implementation of incremental reading would bring your entire learning process (for digital text) into the program. My vision is this:

1. Drop your source material into the program. The more text formats supported, the better.

2. Highlight important portions of the source material -- sentences, or complete paragraphs -- using an extraction tool.

3. These highlighted portions become cards of their own which come up later during your "passive-recall" or "reading" repetitions. (Unlike actively recalled items (i.e. question-answer pairs), the timing of passively-recalled repetitions might be controlled by the user giving different source materials different priority levels).

4. You make further extractions from the newly created cards which in turn become cards of their own until....the original source material has been reduced to cards that have one important piece of information on each.

5. Convert said cards into QA pairs via cloze deletions. These cards now enter your "active recall" repetitions and are subject to the spaced repetition algorithm.

This is Getting Things Done for learning...nothing escapes your reading net. If it's worth remembering, it's granted access to your active-recall queue. Otherwise, it's whittled away during the reading process. Make this all work on a tablet and I think it just might revolutionize the way people read and learn.

I would like to see the addition of javascript hooks, so that javascript is more useful. The primary things I would like to see are:

1) onQuestionLoad() - called when question is loaded

2) onAnswerLoad() - called when answer is loaded

3) onQuestionSoundPlayed() - called when question sound is played / when play button is pressed

4) onAnswerSoundPlayed() - called when answer sound is played / when play button is pressed

5) getAnkiInfo() - returns an object with client name and version, so mobile and desktop clients can be distinguished

Also, I would like a way via javascript to tell anki to enable the sound button because there may be a javascript hook that needs it.

Users would be able to download the deck, add some cards or update others, and commit their changes and other users would be immediately noticed about these modifications, so they can decide whether to download or not the new or updated cards.

Maybe it'll be also great to allow users to decide whether the card they created or updated should be committed on the server or not (but I think this will be made easy with the new management system of the decks)

Finally, decks could be available either publicly or privately, with a key or something, according to the user's needs.

In problem solving, it is key to recall larger pieces of knowledge when only small cues are available to suggest using them (example below). For this, I essentially want to type "[A] [B] [C] [D]", and click a "Cloze complements" button that generates four cards of the form

Font: [A] [?] [?] [?] Back: [A] [B] [C] [D]
Font: [?] [B] [?] [?] Back: [A] [B] [C] [D]
Font: [?] [?] [C] [?] Back: [A] [B] [C] [D]
Font: [?] [?] [?] [D] Back: [A] [B] [C] [D]

An example:

[Bernsteins's theorem] gives [the number of toric zeroes of a sparse generic square system] as [the mixed volume of the Newton polytopes of the system]. It improves the [Bezout bound on the number of solutions]

The reason for having four cards is that when working on a given problem, any one of those four regions might come to mind first, and we want the other three to leap to mind immediately as ways of continuing the problem.

For example, when your friend says "Bernstein's theorem", you want to remember what it says and what it's for. On another day, when you suddenly need a bound on a number of zeroes, you want to remember that mixed volume can help, that "Bernstein's" is the name of the theorem, and that it's strictly better than the Bezout bound.

A "Cloze complements" feature like this would be VERY powerful for learning laws, theorems and problem solving methods.

It would be useful to have a small "media manager". It shows alls the media files (pictures, sounds etc) within this deck. Its surface is close to the fact manager:

It has 4 colums:
1. file name (deckname.media\map_africa.png)
2. type (image/sound/...)
3. facts (number of facts, using the file, e.g. '1' or '0')
4. size (kB/mB)

If the number of facts, using the media file, is equal zero, the is marked red, so you can see easily all the "orphaned" media files.

With a double click on one media file in the "media manager", Anki shows the "fact browser" and filters all the facts, including this certain media file. For this, it is necessary, to search all facts for this filename.

Another grand feature would be in this context, to have a new button inside of the "fact editor" to the media manager. It could be a dropdown menu, showing all linked media files in this fact. Clicking on one of it, it goes directly to the "media manager", filtering just this single media file.

Reason:
The use of media files is a bit intransparent. I never can see inside Anki, which media files I do have. Also, there are some "orphan" files. E.g.: I tested in the beginning to record sound for a fact. After having recorded, I deleted the fact and its link to the sound file. Months later, I realized, that there are still 3 sound files in the media directory, which aren't used by any fact. At the same time, it is a bit risky to just delete a media file, because you never know for sure, whether it is used by a fact or whether it is "orphaned".

Greetings, spectas!

To realize its true potential Anki should get more socially integrated.

The social community on the web is organized via a) browsers and b) facebook,

therefore i propose to introduce options to make

1- your online account
2- your online accounts' shared decks
3- your shared decks' learning progress

publicly visible.

User should be able to decide which of the features to use or not to use (additional features).

Together with a facebook-app tracking the progress of your learning this could turn learning into a highly motivating process comparable to computer games.

For further inspiration check out: http://www.ted.com/talks/tom_chatfield_7_ways_games_reward_the_brain.html

Also I recommend to watch what the leading competitor http://www.flashcardexchange.com/ is doing - namely: precisely stressing the social features.

That if you have a deck open, anytime a review is due it would automatically show.

My specific use of this is that I could create Pimsleur-like intervals when first learning a set of cards by setting the initial interval to .0001 or something like that which yields an initial interval of about 15 seconds. Then I use the 3 or 4 buttons to increase the interval as I am comfortable with the card.

This is a feature request for all versions of ANKI since they seem to all work the same way. In a previous version of ANKI, if one had a deck open, as soon as a review was due, it would display. Now on all three platforms (desktop, mobile and web), if I have a deck open, I have to close the deck and reopen it in order for the reviews to display (on desktop to manually hit the "review" button again)