Since I’m trying to replicate my rat study in humans, I’m using some backward masking to make the task harder. And all of the masks are loaded at the beginning of the protocol. So far, so good, the experiment is working. But then I decide to change my masks, I rename the old folder “Masks_old” and make a new one with the name “Mask" (so I don’t have to touch the protocol). But it keeps showing me the old masks. OK, MWorks seems confused, maybe because both folders are in the same folder together so I move the old folder to a different location. Still no luck. I try renaming the new masks folder and changing the protocol, but then it won’t load the images (reports “probably bug”). I assume that there is still some cache where MWorks looks for the images but I can’t find it (tried in \Documents\MWorks with hidden files on, but no luck). Perhaps I’ll find an answer by the time you reply (since it seems this is how this dynamic works 
), but so far I had no success.
Vladimir
I am unable to reproduce the problem you describe.
I have attached my test experiment. I run it once, and I see “1”, “2”, “3” on the display. I close the experiment, rename images to images_old, rename other_images to images, reload and re-run the experiment, and I see “4”, “5”, “6” on the display, as expected.
Can you provide an example that demonstrates the issue you’re having?
Regards,
Chris Stawarz
Attachment: folder_rename.zip (74.5 KB)
A lot of the time when I deal with MWorks I encounter the phenomenon similar to the “singing frog”.
Some changes that I’ve tried one day and they didn’t work, the other day work just fine.
Sometimes it makes me doubt my sanity.
Anyway, the protocol is working now (even though I’m not sure why, I’t might have something to do with changing the picture indexing form 1 instead of 0, although in the previous experiment it worked just fine with indexing starting from zero), so if there are any new problems I’ll let you know.
Thank you for your time.
Vladimir
Scratch the previous message. It works fine when I use my laptop as both client and server but it doesn’t work when I try it using Mac mini as a client.
For some reason in needs to be fixed locally on the server computer using the MWorks editor (not the text editor or XML code) and then it works.
I really have no idea why is this happening but I might be the only one with this problem.
Now it’s working.
Vladimir
It works fine when I use my laptop as both client and server but it doesn’t work when I try it using Mac mini as a client.
I had assumed your masks were standard image stimuli, but that invalid value ... for attribute "path" error isn’t something the standard image stimulus would raise. What type of stimulus are these masks – something custom your lab uses?
For some reason in needs to be fixed locally on the server computer using the MWorks editor (not the text editor or XML code) and then it works.
You mean you need to modify the path to the images folder in MWEditor, and making an identical change in a text editor causes problems? If so, then it might be a text encoding issue (e.g. the XML declares itself to be UTF-8, but the text editor inserts byte sequences that aren’t valid in UTF-8). But I wouldn’t expect that problem to result in the error you’re seeing.
Chris
I’m also attending a seminar right now, and I needed some time to document the problem properly.
Find the description in the attachment.
Best
Vladimir
Attachment: MWorks_problem.pdf (680 KB)
Thanks for the additional details. Could you also send me the various files involved? Ideally, I’d like to have an experiment file that doesn’t load, an apparently identical experiment file that does load, and the associated image files. If I can reproduce the problem myself, then I should be able to diagnose and fix it.
Thanks,
Chris
I think that it won’t be a problem but I need a confirmation from my boss for this.
He is on vacation and might not be reachable before he comes back so I will contact you when I hear from him.
Vladimir
OK, the boss is back and I have a green light.
In the .zip file you will find everything that you need to replicate the experiment.
Let me give you all the additional data that I can think of:
Computer 1: MacBook Pro, OS X 10.9.5, MWorks 0.4.4
Computer 2: Mac mini, OS X 10.6.8, MWorks 0.4.5
Experiments are stored in /Users/labuser/Development/zoccolan_vnsl/behavioral_projects/behaviour_Vlad/experiments/ (you should make this one yourself or change the paths in the protocols). In the second one you should delete "/Volumes/OSX Maverick/“ (inserted by editor after local edition I’ve already described).
The first protocol that I’ve told you about is “Humans_1” - works well on MacBook (as both server and client), doesn’t work in the MacBook (client) - Mac mini (server) configuration.
The second protocol (Humans_2) works (path seems to be the same).
Also, for whatever reason, the edited protocol is almost 20KB smaller.
Since I have a working protocol now, you don’t have to rush to figure this one out (it might me useful for some further versions if you find that it’s a bug).
Right now I have a more pressing issue.
The “setup_variables” file should be the one defining the external parameters (monitor size, number of monitors), and there’s even an example file for two monitors.
Yet, I’m not able to get this to work (either by attaching an external monitor to my MacBook - MWork crashes, or by attaching two monitors to Mac mini).
If I attach two monitors to Mac mini I will get the stimuli presented on one of them (full screen), but I can’s get the mirror screen on the other one.
Any suggestion is welcome.
Vladimir
Hi Vladimir,
Thanks for sending all the files. I’m still looking things over, but one detail caught my eye immediately:
Computer 1: MacBook Pro, OS X 10.9.5, MWorks 0.4.4
Computer 2: Mac mini, OS X 10.6.8, MWorks 0.4.5
There’s no guarantee of compatibility between clients and servers from different MWorks versions, and attempts to mix versions often lead to strange errors. Can you try installing 0.4.5 on your MacBook and see if that changes anything?
In the meantime, I’ll continue my investigation using the files you provided.
Cheers,
Chris
Hi Chris,
just lake you’ve said, there is no guarantee that different versions will work together and that’s also true for the protocols written for different versions.
Since I don’t have time to change my protocols now, I’ll stick to this and maybe change it later if there are some problems in the future.
Vladimir
