Dear IBEX users,

I wrote a modified version of the Question controller for picture-sentence 
matching task for people in our lab. I guess this may be useful to other 
users so I am sharing it here. To use it, just put the .css file in the 
css_includes directory and the .js file in the js_includes directory.

The controller is called PictureAccept. It takes as arguments a sentence 
("s") and a list of url's which should link to pictures ("as"). "as" can 
also be a list of lists, which each sublist containing a key and a url, as 
in Question. In this case, a response key is associated with each image. If 
"as" is just a list of url's, participants will need to click on the 
pictures.

The result file contains a number indicating which picture was chosen (from 
1 to n) and the time taken to answer (this is explained in the comments of 
the results file).

Example syntax (A, S, D being the response keys for pictures 1, 2, 3 
respectively):

["Example","PictureAccept", {s: "The sentence you want to show",
                            as: [["A","http://...picture1.png"],
                                   ["S","http://...picture2.png"],
                                   ["D","http://...picture3.png"]]

The width of the pictures is forced to be 16em in the .css file ("img" 
entry) and they always come with a black border. You are free to modify 
this as you need (especially if you would like to include more than 3 
pictures). You still need to host your pictures somewhere outside ibex farm.

I did not test it a lot, so it may not work perfectly on all browsers, or 
if you try something too specific (many or big pictures...). You'd rather 
not look inside the .js file if you enjoy beautiful code as it is very 
unclean (I did minimal modifications of the Question controller, and left a 
lot of things which are useless here).

You can see an example here:

http://spellout.net/ibexexps/Emmanuel/Picture-Matching/experiment.html

I hope you will find it useful!
Alexandre