|
zajac
|
|
|
Group: Forum Members
Posts: 28,
Visits: 98
|
Just wondering if this is normal or if I'm doing something wrong. I created a batch file that seems to miss some tasks. I organise tasks into individual folders, so it look like this
<batch> /directory = "C:\HAL_Tasks\OMO_Letter" /directory = "C:\HAL_Tasks\Visual_RT" /directory = "C:\HAL_Tasks\Visual_RT_4" </batch>
After messing around for a long time, I noted that the OMO_Letter task is an .iqx file. The other RT tasks were .exp and they were working fine with the batch code, but the batch wouldn't run the first OMO task.
I simply changed the OMO_Letter task extension from .iqx to .exp. Straight away, it now runs all scripts no problem. Why is this? Do I have to go through and rename all of my tasks to .exp to make them compatible with the batch element? I am running windows 8 and V4.0.4.0 64bit
|
|
|
|
|
Dave
|
|
|
Group: Administrators
Posts: 13K,
Visits: 108K
|
You ought to include the file extensions in your <batch> syntax (this has always been the case):
<batch> / file = "C:\myfolder\myscript. iqx" [...] </batch>
If you are using Inquisit 4, you should save your files in the *.iqx format. To convert an older *.exp, open it in Inquisit 4 and then save it as *.iqx via File -> Save As...
|
|
|
|
|
zajac
|
|
|
Group: Forum Members
Posts: 28,
Visits: 98
|
Dave - sorry to bring this up to the top again, but I'm experiencing some more problems with a batch file and Windows 8.
1: Sporadically, ctrl+Q does not work at all. It isn't due to a trial taking place etc. The experiment just keeps running. I have to ctrl+alt+del and kill the whole program. I can't predict in which script it is doing this because it appears relatively random and I can't see how a specific script would affect this anyway.
2: I've had some timing issue with a specific task when it is run via batch. One time it got half way through this task and just quit out and moved on. The second time it barely started the testing phase before it quit out and moved on. It runs fine when I execute the script individually.
ALSO - in regards to this script I just mentioned, I have the test block timeout set to 1 minute (60000). I have timed this quite a few times with a stop watch. It appears that it times out much closer to 50 seconds than 60. How is this possible?
I've attached the suspect script in case there is something wrong with it, but it doesn't seem that way. Is there a way to store a time variable in the data output to check the timing afterwards via calculation etc?
|
|
|
|
|
zajac
|
|
|
Group: Forum Members
Posts: 28,
Visits: 98
|
I've also just timed this script a few times. The test block should run for 90 seconds. It times out @ 1min20sec (or around 80 seconds only). I'm starting to have some real concerns about the timing accuracy?
|
|
|
|
|
Dave
|
|
|
Group: Administrators
Posts: 13K,
Visits: 108K
|
Sorry, but I'm not seeing any of these problems happening in my tests. Is your Inquisit installation on that machine up to date?
|
|
|
|
|
zajac
|
|
|
Group: Forum Members
Posts: 28,
Visits: 98
|
Hi Dave. Yes, lab machines were installed from the latest download file yesterday. I'll do some timing investigations with Win 8 today. Yesterday I was timing on a Win 7 laptop.
Re the ctrl+q issue, any ideas? The testing account on those machines is not an admin account, but I don't think this would have any impact?
Is there a way of storing time in milliseconds in the summary data file?
|
|
|
|
|
Dave
|
|
|
Group: Administrators
Posts: 13K,
Visits: 108K
|
No, I don't have any spontaneous idea about the random CTRL+Q malfunction. Regarding the timing, you have several elapsedtime etc. properties at your disposal which you can log to the data file. See the documentation / language reference for details:
elapsedtime property The amount of time since the script or experimental event began or a clock was started. Applies to <block> <clock> <expt> <likert> <openended> <script> <survey> <surveypage> <trial>
Reviewing the technical logfile (Tools -> View Log File...) may also be worth a shot.
|
|
|
|
|
zajac
|
|
|
Group: Forum Members
Posts: 28,
Visits: 98
|
Ok, so I've been doing some more testing on the lab machines. My RA and I have timed the Finding As and Number Comparisons tasks multiple times. Finding As is a 1 minute task. Number Comparisons is a 90 second task. For both tasks, I created a test block .elapsedtime value which is printed to the summary file. I also have the main data file printing a 'test.elapsedtime' variable for every trial. In the both the summary and main data file, the millisecond count comes out within @10ms of the timeout attribute. However, I can confirm that when time with a stopwatch (multiple times mind you), they are not agreeing. Finding As times out according to the stopwatch at @ 50 seconds. Number comparisons times out @80 seconds. Now accounting for some motor delay in starting and stopping the stopwatch, let's say there was a realistic difference of a few seconds maximum. The ten second difference I'm reporting isn't accounting by any stopwatch start/stop variability. We've even timed with two separate stopwatches and they are in agreement with each other.
Dave - how did you time these tasks? Did you just print out and elapsed time element? Or did you manually time them?
I'll note that I have a 2 minute task (120ms). This according to a print out elapsed time variable is timing out @120ms as required. Interestingly, the stopwatch has it at pretty much the same time. So it agrees with Inquisit with this longer task.
|
|
|
|
|
zajac
|
|
|
Group: Forum Members
Posts: 28,
Visits: 98
|
Oh, and log file doesn't report any errors etc. It always says the scripts were executed successfully.
|
|
|
|
|
Dave
|
|
|
Group: Administrators
Posts: 13K,
Visits: 108K
|
> how did you time these tasks? Did you just print out and elapsed time element? Or did you manually time them?
I did both. Timing checks out okay. Did you account for your instruction pages' display time? That's part of the timeout.
|
|
|
|