Error in docs about "file creation date"

[Alex Tweedly]

Wilhelm Sanke wrote:
> On Wed, 23 May 2007,Richard Gaskin <ambassador at fourthworld.com> wrote:
>
>> Right now Rev uses the OS APIs to get the file info, so it's about as
>> efficient as it can be.
>>
>> I can see the benefit of using EXIF, but since this requires additional
>> work and processing time, and only applies to a subset of a very small
>> number of file tyes, this sounds like an excellent candidate for an
>> external.
>>
>> Anyone here interested in writing it?
>>
>> -- Richard Gaskin Fourth World Media Corporation
>
>
>
> I do not know if he is interested and would have the time, but I think 
> Alex Tweedly, author of the wonderful "libEXIF" stack, would be the 
> ideal candidate.
>
I'd certainly be interested. I *should* have time.
> I have tried to "tweak" his stack a little to retrieve the "original 
> dates" from all image files of a folder. Execution time per image file 
> is from 13 to 40 milliseconds depending on what amount of date I get 
> from the "original date" line of Alex's variable "alltags" (i.e. only 
> the "original date" or with filename and/or seconds etc.).
>
> I left the stack script intact and edited only the "test" button to 
> get only the original-date lines. This means that the stack - as 
> before - first retrieves *all* EXIF data in the stack script before 
> only the date line is accessed from the changed button script.
>
> Would Alex rewrite his stack script to concentrate on the "original 
> date", I believe the execution speed per image file would be only 
> about 5 milliseconds or even less. An external based on his algorithms 
> would surely be astonishingly fast.
>
I'll have a look at it, probably over the weekend. I haven't actually 
looked at the EXIF stack for quite a while, so it will be interesting to 
see just how poor my code documentation was :-)

-- Alex.