|
Size: 1177
Comment:
|
Size: 2540
Comment:
|
| Deletions are marked like this. | Additions are marked like this. |
| Line 7: | Line 7: |
| A number of changes need to be made to the Resu1 experiment before it is taken back in the field. | A number of changes need to be made to the Resu1 experiment before it is taken back in the field. Many of them need to be described in enough detail to be implemented. All of them need to be ranked by how important they are to be completed before the next trip. |
| Line 9: | Line 9: |
| * Create a chooser scene so multiple experiments using the same scene-types can be selected from. (e.g. Resu1 or Katrina's new experiment). * Check all files in the list to be sure they actually exist * Make a bunch of things configurable by a delim list * Create a part to do list autobalancing, so the experimenter doesn't have to choose. This is 75% complete in C#. As soon as I check that it does the right things, I'll release the C# file and the DLL. * Create a new trial type to do a quadruple of scenes instead of a triple. Pattern of TFTF, FTFT, or TFFT (where T = target soundfile and F = filler soundfile). |
== Chooser == Create a chooser scene so multiple experiments using the same scene-types can be selected from. (e.g. Resu1 or Katrina's new experiment). '''This needs to be explicated. I have little clue what Florian means.''' It might make more sense to try to componentize scenes, but if they have any outside reference (say, to a trial list) that's difficult. == File Checker == Check all files in the list to be sure they actually exist. '''This is non-trivial and I'm pretty sure no experiment software does this because it's more trouble than it's worth.''' We run through all the lists before subjects do the experiments for a reason. ;) We could write a separate script that runs through lists of a known format that could be run on each computer that will be running the experiment. == Config file == Make a bunch of things configurable by a delim list. '''Easy to do. We just need to decide which things should be configurable this way.''' == List Autobalancer == Create a part to do list autobalancing, so the experimenter doesn't have to choose. This is 75% complete in C#. As soon as I check that it does the right things, I'll release the C# file and the DLL. == Triple to Quadruple trial type == Create a new trial type to do a quadruple of scenes instead of a triple. Pattern of TFTF, FTFT, or TFFT (where T = target soundfile and F = filler soundfile). '''I need to look at the current experiment to see how difficult this will be''' == Componentize the experiment == Where possible it'd be nice to make scenes we want to use in multiple experiments into components. This would allow for less copy and pasting, and mean that when the component is updated all scripts that use it (assuming they get the new component source or dll) get the fixes/changes for free. = Source code = |
| Line 21: | Line 39: |
| hg clone ssh://username@slate.hlp.rochester.edu:/Users/awatts/ExBuilder_stuff/Florian/Resu1 Resu1 | hg clone ssh://username@slate.hlp.rochester.edu//Users/awatts/ExBuilder_stuff/Florian/Resu1 Resu1 |
Resu1 (Version 2)
A number of changes need to be made to the Resu1 experiment before it is taken back in the field. Many of them need to be described in enough detail to be implemented. All of them need to be ranked by how important they are to be completed before the next trip.
1. Chooser
- Create a chooser scene so multiple experiments using the same scene-types can be selected from. (e.g. Resu1 or Katrina's new experiment).
This needs to be explicated. I have little clue what Florian means. It might make more sense to try to componentize scenes, but if they have any outside reference (say, to a trial list) that's difficult.
2. File Checker
- Check all files in the list to be sure they actually exist.
This is non-trivial and I'm pretty sure no experiment software does this because it's more trouble than it's worth. We run through all the lists before subjects do the experiments for a reason. 
We could write a separate script that runs through lists of a known format that could be run on each computer that will be running the experiment.
3. Config file
- Make a bunch of things configurable by a delim list.
Easy to do. We just need to decide which things should be configurable this way.
4. List Autobalancer
Create a part to do list autobalancing, so the experimenter doesn't have to choose. This is 75% complete in C#. As soon as I check that it does the right things, I'll release the C# file and the DLL.
5. Triple to Quadruple trial type
- Create a new trial type to do a quadruple of scenes instead of a triple. Pattern of TFTF, FTFT, or TFFT (where T = target soundfile and F = filler soundfile).
I need to look at the current experiment to see how difficult this will be
6. Componentize the experiment
Where possible it'd be nice to make scenes we want to use in multiple experiments into components. This would allow for less copy and pasting, and mean that when the component is updated all scripts that use it (assuming they get the new component source or dll) get the fixes/changes for free.
Source code
The ExBuilder files are available via Mercurial if you have an account on the HLP system:
hg clone ~awatts/ExBuilder_stuff/Florian/Resu1 Resu1
locally or
hg clone ssh://username@slate.hlp.rochester.edu//Users/awatts/ExBuilder_stuff/Florian/Resu1 Resu1
remotely.