dljve / wordwar Goto Github PK

• Game sends this in render() when isGameEnd() is true
• Endgame handeler calls slimstampen.endSession() to save all parameters for the user

Record timestamp and action of user

Currently the tiles decay too fast.

• Are the current model parameters (n, c, f, F, t) adequate?
• Is the standard alpha of 0.3 adequate?
• In the decay calculation in updateModel: (1) how to handle the rare case of division by zero (2) why is the negative log sometimes taken in the nominator? Is this mathematically plausible (see Nijboer 2011, page 13) or is there an error in the implementation?

During trial mode the non-active tiles should stay very transparent, but at some point they change back to normal. Probably due to a board update when the gameobject has changed.

I have created the AI, but it still needs to be tested whether it works

The AI is run by running the server.
Two AIs are created, one creates a game, the other joins it.
Then, the server waits for a third player to join the game and gives a starting sign to the two AIs to start playing.
It should have a similar behaviour to that of a human in terms of response times and forgetting words according to slim stampen.

If the user types the translation misspelled and hits enter the label from the top should turn red (just as it does in the rehearsal phase)

Right now the keyboard stays visible. When you get into the lobby screen you have to press the back button in order to see the 'create new game' button.

Issue #5 should be fixed first for this.

Possible way:

• Use a csv file conditions.csv
• Each row has a two columns: condition and nickname
• Condition on each row is randomized (either semantic or random, both appear equally in the csv)
• Session start: the first row without a nickname will decide the board condition for the next participant
• Session end: when all data is saved, the nickname is written to the row in the csv. This links the condition with the participant.

To prevent

• an additional learning effect from visible tiles.
• comitting translations to iconic memory at the last minute.

It would be cool if the color of the tiles actually represent the activations as approximated by SlimStampen, before forgetting the tile entirely.
This would make the map of the game a more real representation of the knowledge in the brain.

Possibly different configurations

Ideally still shuffle the board, while tiles maintain high semantic similarity with their neighbors.

Start of session:

• Record start of session time to link it to participants
  this is recorded in folder name and log

At the end of the session, for each word:

• Save Response Time parameters from SlimStampen model to csv
  endSession function in SS

• Save alpha parameters from SlimStampen model to csv
  endSession function in SS

• Save time points to csv
  endSession function in SS

• Save time points of forgetting tiles (to see how long after forgetting the tile the user rehearses them)
  this is recorded in the log

• Record end of session time
  the log should record the time of the endGame message