Hallo Chris,
We had a long meeting today, and Stefan ask for two little features… I’m not sure if it will possible to include in version 0.6? He ask for the feature to save used sounds and pictures to the datafile… what do you think about this feature?
Lieben Gruß aus Göttingen
Ralf
Hi Ralf,
We could add this, but I’m uncertain about many of the details.
For one, I think it probably shouldn’t be done by default, since not everyone will want it (particularly those using large image sets). Hence, we’d need to add a configuration option somewhere. Would that option be set separately for each experiment, or should it be global and apply to all experiments?
Also, while it’s easy to inject file data in to the event stream (by wrapping the binary data in a string value), this has some undesirable complications. We’d probably have to add a new system variable to hold the data (similar to the #loadedExperiment variable that’s already in the nightly build). The trouble is that variable values are announced whenever a client connects or an event file is opened, so you’d most likely end up with multiple copies of the files in your data. I’m not sure how we’d avoid this.
Are you familiar with the file hashing that MWorks does on image files? While it’s not the same as having the whole image in your event stream, it does provide a robust way of identifying the images used in a given run. We could (and probably should) do this for sound files, too.
Cheers,
Chris