View Single Post
Old 07-11-2013, 08:44 AM   #8
knc1
Embedded Cheerleader
knc1 ought to be getting tired of karma fortunes by now.knc1 ought to be getting tired of karma fortunes by now.knc1 ought to be getting tired of karma fortunes by now.knc1 ought to be getting tired of karma fortunes by now.knc1 ought to be getting tired of karma fortunes by now.knc1 ought to be getting tired of karma fortunes by now.knc1 ought to be getting tired of karma fortunes by now.knc1 ought to be getting tired of karma fortunes by now.knc1 ought to be getting tired of karma fortunes by now.knc1 ought to be getting tired of karma fortunes by now.knc1 ought to be getting tired of karma fortunes by now.
 
knc1's Avatar
 
Posts: 6,185
Karma: 5830426
Join Date: Feb 2012
Device: Too many.
Quote:
Originally Posted by thomass View Post
There is an KUAL extension for this purpose: here
Thanks!
I had forgotten that one.

Although it is only a partial solution, but better than no solution at all.

What that thread goes on to discuss is the Amazon firmware's "split brain" behavior and the problems that can cause.

1) Put book on Kindle (without a connection available to Amazon).
2) Files are generated on the Kindle
3) Delete the book, before the firmware can check if it is also present on your device's Cloud(let).
4) If the book is on your device's Cloud(let), the generated files will be retained on the Kindle so they do not have to be re-generated when (if) you re-load the book from your Cloud(let).
5) If the book is 100% device local (not on Cloud(let)), then the generated files should all be deleted also.
6) But in this scenario, the firmware has not been able to contact the device's Cloud(let) to determine if it is a truly a "device local" book.
7) Now the book is gone, locally, so it will not be checked for on the Cloud(let) even if a connection to Amazon becomes available.

I wrote above "a partial solution" because the Amazon firmware takes one path through the above cases of "split brain" behavior and the KUAL file prune button takes another path.
Neither is exactly correct in all situations.

---
"Split Brain" : Partial dataset storage in two (or more) locations, one or more of which may not always be available.
Not a "good idea" (tm) in dataset storage design.

Last edited by knc1; 07-11-2013 at 08:56 AM.
knc1 is offline   Reply With Quote