Cardflasher

Introduction

Cardflasher is a simple flashcard program I made that is nonetheless quite featureful. The main concept, borrowed from Granule which in turned borrowed it from Leither's cardfile methodology, is segregating cards according to how many times you've gotten it correct.

Things I didn't like about Granule that motivated me (maybe different now):

You had to edit decks in the Granule GUI unless you like painful XML editing
Decks/Cards/Cardfiles had some brittle hard links using file paths, so you couldn't just send them to a buddy
Identical cards in different decks were not detected
Only two sides to a flashcard (learning Japanese, at least three are useful)
Crashy, heavyweight, no console UI

So, I derived these features from the above complaints:

The deck files have an extremely simple format, but also an optional advanced format
A card is exactly identified by its contents, so a deck is just a list of cards and is totally file path independent.
Many decks can be passed to cardflasher, which automatically merges duplicates.
A single file keeping track of cardbox levels, in ~/.cardflasher/ instead of per-deck stuff. Easy to back up, automatically detects if you have already practiced something in some new deck.
Arbitrarily many sides to a card, and you specify what goes on the front and what goes on the back each practice session. The cardbox data is stored according to this, so going Japanese->English has a separate level from going English->Japanese
In addition to trimmed-down GTK UI, there is a nice, simple console UI, which I use to do cards remotely.
Written in OCaml, pretty darn lightweight and crash-free.

And here are the things I got rid of that were probably useful:

Only ten boxes - in cardflasher you tend to keep pushing cards higher, which makes it hard to know where you are at
No expiration dates - you have to remember by yourself to review things.

Screenshots

GTK interface
GTK interface

Console interface
Console interface

In this shot, I've used shell globbing to pass a huge pile of decks to cardflasher, which automatically merges them, detecting duplicates. And then I passed --summary so I could see where all my cards were.
--summary demo

With the --each option, I get a summary by file, to see where I should concentrate.
--each demo

This is actually a screenshot of me editing a deck in vim, so you see how trivial the file format is (it has more complex options, but forget about those for now). In the deck shown, --given 0 shows the english, --given 1 shows the katakana, and --given 2 shows the kanji. Every side that isn't given is on the back of the card when you flip it. You can do combinations like --given "1 2" and cardflasher automatically keeps a separate level count for each pair of given/not given sets.
file format demo

Downloads

The program is in cardflasher-* while the other files contain flashcards I have typed up. The file japanese-* contains complete cardsets for the books used at UCSC. I would welcome someone assembling deckfiles of common words according to topic or frequency.

4/7/2006

11/12/2005

11/11/2005