Error in docs about "file creation date"

[Wilhelm Sanke]

On 23 May 2007, at 09:17, Ian Wood wrote:

> >> Obviously this isn't a problem for images that have been 'saved as'  
> >> with the EXIF data stripped out because they will have a new valid  
> >> creation date being new files, but people do it to original files  
> >> as well (or copies if they know what's good for them...).
>
> Wilhelm - out of curiosity, what would you regard as the 'correct'  
> creation date for an image that has been opened, edited and saved to  
> a new file?
>
> Ian


First of all, I would like to have a correct creation date - the day the 
photo was taken - for all images whose contents remain unchanged 
although they may have been copied, transferred to another volume, or 
saved under they original or even a new name. As we have seen, this is 
not guaranteed, and the detailed-files function does not help much here.

For an "edited" image this is a question to what extent changes were 
made, e.g. if only basic properties like contrast, color balance etc. 
were edited then I would prefer to keep the original creation date. If 
you more or less produce a new picture on the basis of the original one, 
it seems obvious that one should choose the modification date as the new 
creation date.

The best alternative would be to have a choice: There are even tools 
that allow you to save the EXIF data before editing and then add them 
again after the image has been changed, like "Exifer" for Windows.-

Another remark concerning my request to add the EXIF creation date to 
the detailed files: I see that the detailed-files function also collects 
platform-specific information, where the information is simply left out 
for the corresponding item when it does not apply for a platform. The 
arguments of  Jim and Jaqueline need of course to be considered, 
especially the possible slowdown of data retrieval, but the extent of a 
decreased speed could be tested in practice.

Best regards,

Wilhelm Sanke
<http://www.sanke.org/MetaMedia>