Time recorded within text file?

SuperLab SuperLab 4 Support
1

Hi there,

I ran a standard face recognition paradigm back in April for my MSc dissertation. My supervisor has looked over the original text files and is having a problem understanding the time recorded within them when compared to the “created” and “modified” times. I must admit I’m a bit flummoxed too. The times in the text files are consistently between 10 and 15 minutes before the “created” and “modified” times, and the latter two are consistently the same.

Could anyone explain to me how this might happen? I was running superlab 4.x (couldn’t tell you the specific build number as I’m unable to get on to campus now).

Does superlab use a different clock to the OS (OSX)?

Any help would be greatly appreciated, thanks.

Chris

2

The time that is printed in the heading/intro section at the top of each output file is the time at which the experiment began running.

The create time that the operating system is showing to be associated with the output file is the time at which the experiment ended and the data was saved to disk storage.

You mention that the differences that you observed range between 10 and 15 minutes. To me, that indicates that your experiment typically took anywhere from 10 to 15 minutes to be completed. Hopefully you can confirm my hunch that your experiment would typically have a duration of 10-15 minutes. If so, then all these various timestamps make sense.

To be clear: SuperLab does not use a different clock than the OS. The two sets of timestamps that you are seeing were produced/timed off of the same clock. One timestamp was taken at the beginning of the experiment and the other was taken at the end.

Therefore, if you are interested in documenting the time at which each participant began viewing an experiment, then you would use the time printed in the into/heading of the text output file.

3

I see, thanks for that, it’s as I would have expected. The experiment did indeed take roughly between 10 and 15 minutes.

That explains the simple stuff, now I have another related query. My supervisor also notes that on his mac some of the ‘created’ and ‘modified’ times actually occur before the timestamp in the text files.

Furthermore, on my mac, some of the discrepancies between the timestamp and the ‘modified’ and ‘created’ times are upwards of 50 minutes (though I always see the timestamp in the text file as happening before the ‘created’ and modified’ times).

Now, having watched all my participants undertake the experiment, I know that no-one took longer than roughly 20 minutes.

To clarify, there are 4 macs I’m talking about here. 2 were used to run experiments, one is my supervisor’s and one is mine. The majority of the experiments were run on one mac, with the second mac being pressed in to service on the odd occasion that two participants turned up at once.

I have a hunch that as the majority of the time stamps look about right (between 10 and 15 minutes), they were run on the 1st mac.

Those that look odd to me (upwards of 50 minutes) may have been run on the second mac which had a different time on it.

Could it be the case that the ‘created’ and ‘modified’ times are altered by OS X depending on which mac you look at them? That is, say an experiment was run at 12pm GMT, but the OS clock is showing 1pm GMT. Now, superlab would put ‘1pm GMT’ in the text file and OS X would save the file as ‘created’ and ‘modified’ as 10-15 minutes after 1pm GMT
Say then, I moved the files to a second mac which displays the correct time, would the ‘created’ and ‘modified’ times change to show 10-15 minutes after 12pm GMT?

Apologies if I’ve made little sense, it’s just a little hard to figure out by myself!

Thanks again,
Chris

4

Save data to disk after each trial ?

There is an option to save data after each trial. If that is enabled, header rows are written as soon as the experiment begins (thereby creating the log file) and data is written to it after each trial (thereby modifying it). On the other hand, if that option is not enabled, the logfile is created and written out fully only at the end of the experiment.

This may possibly be part of the confusion. Look in Experiment > Options > Misc to see if the option is selected in the various experiments.

Greg Shenaut

5

[QUOTE=Xenon901;1950]I have a hunch that as the majority of the time stamps look about right (between 10 and 15 minutes), they were run on the 1st mac.

Those that look odd to me (upwards of 50 minutes) may have been run on the second mac which had a different time on it.

Could it be the case that the ‘created’ and ‘modified’ times are altered by OS X depending on which mac you look at them? That is, say an experiment was run at 12pm GMT, but the OS clock is showing 1pm GMT. Now, superlab would put ‘1pm GMT’ in the text file and OS X would save the file as ‘created’ and ‘modified’ as 10-15 minutes after 1pm GMT
Say then, I moved the files to a second mac which displays the correct time, would the ‘created’ and ‘modified’ times change to show 10-15 minutes after 12pm GMT?

Apologies if I’ve made little sense, it’s just a little hard to figure out by myself!
[/QUOTE]

Hi Chris.

Actually, whether you realize it or not, I think you have mostly provided the answer to your own question. So perhaps you are simply looking for someone to confirm your hunch?

The reason I say you answered your own question is that I do think that this is explained by the fact that the second mac has a different time. Actually, even the fact that the second mac has a different time might be “overkill” in terms of facts needed to solve this question. In some cases just the simple fact that there are two machines between which to transfer files is enough to explain a case like yours.

I am not an expert on the internals of the Mac operating system. However, Mac OS X shares much in common with the BSD Unix system. Unix systems typically associate three date values with each file. These three values are atime, ctime, and mtime.

Here are some links that might help clarify:

ctime_atime_mtime

Note that ctime is not the time of file creation

“stat” and “touch” are two tools that are available to examine and to alter these times.

Anyway… when transferring files from one machine to another, a lot of little things can go on “behind the scenes” that you might never notice or necessarily expect. Factors that will be in play during a file transfer:

File system on the source machine.
File system on the target machine.
User permission/User identity of the user performing the transfer.
Software used for the transfer.

File systems could include NTFS, FAT32, HFS, MFS, EXT2, or others.

Software being used could be anything from the command-line scp utility to the GUI “Finder” application to an email client being used to email the files back and forth.

Something that is probably a very important point to glean from any discussion of atime, ctime, and mtime, is this:

It is easy to tweak these times.

It is easy to tweak them accidentally and harmlessly, and it is also easy to tweak them deliberately. It wouldn’t be something that probably deserves too much concern. There’s no way for me even guess what various types of scanning software or backup software or any other factors might be operational on either of these two machines that might cause any apparent date-time weirdness. But I don’t see any cause for extraordinary concern.