[KimDaBa] About the renaming issue

[Daniel Berger]

Dario Spagnolo wrote:

>Jesper K. Pedersen said:
>  
>
>>One quick question from a quick read though while compiling.
>>If I split up the index.xml file, how would you the represent it in the xml
>>files when images has been moved in KimDaBa.
>>
>>Imagine I have this order in KimDaBa:
>>misc/img1 party-42/img12 misc/img2 party42/img42
>>
>>currently I get the order from the order in the XML file, but if I split
>>    
>>
>up one file per directory, I would loose that feature.
>  
>
>>Regarding the idea only of having md5 sums in the db (from Darios
>>    
>>
>email): I tried creating md5 sums for 892 images, which took approx 30
>secs. Are you really willing to wait 30 secs/1000 image when KimDaBa
>starts up?
>
>There is some first-use loading time anyway because of the thumbnail
>generation. But it's true that 0,03 seconds/image is a bit longish (what
>kind of computer is that ?), but I guess we could also wrap the md5
>function so that it only takes into account the first 10kb and the last
>10kb, no ? I guess it would still work perfectly fine for 99,999999% of
>situations.
>
Think about speed!
I've 11'600 images in KimDaBa, and my P-III 866 Notebook is getting to 
slow to have a good feeling...
That scan would take about 330 secs -> 5,5 mins!!! No, it would even 
take longer in cause of the slower fsb etc.
On my Home Box (2600 MHz) KimDaBa works very fluently now, and I don't 
hope it will get slower in future.
But I'm shure, that there are peoples with more images than me.