MobileRead Forums - View Single Post - IF you could design the user interface for the next Kobo Firmware

TechniSol · 06-10-2013, 05:58 AM

Theonna,

I'm all for improvements, but I think it unrealistic to expect a major paradigm shift in the core of the software.

The database is a rational approach in that the information must only be collected once, not every time the system is booted, or files are looked for. This should result in less time overall for various tasks to be performed as in theory the information must only be collected once and only updated as changes occur.

The real questions when one runs into over long pauses are what are they doing, and why is it taking so long? As fast as the processor is, one must wonder what takes so long for large shelves to be displayed. My guess is they are processing all of something rather than just enough to start displaying the first page and continuing to process in the background while allowing the relatively feeble human to see and injest that first page. Even a slow system should stay ahead of a human as long as there is no ability to jump far ahead, and even then should only bog down for the time to process one page and display it.

With such an approach the lag time becomes shelf length independent and relates directly to the processing time necessary only to process/display one page of the shelf. I have few shelves, and only enough items on them to require a few pages to display the entire shelf.(OK, one with 99 items, but it's quick too) My shelves display quickly. I think they would only require a bit of re-coding to achieve this(if they don't have it already), as long as it is possible to run a second task in the background to continue whatever processing they are consuming so much time with. They might even find the most efficient use of that task would be to only process a few pages ahead, rather than the whole shelf if that would keep up well ahead for most casual browsing and minimize power requirements. I would think an algorithm based on percentage would be most efficient so that as shelves grow ever larger the same percentage of pages are processed ahead by the background task rather than a set number of pages.

06-10-2013, 05:58 AM	#67
TechniSol GranPohbah-Fezzes r cool! Posts: 1,056 Karma: 3151024 Join Date: Jul 2010 Device: Nook STRs, Kobo Touch, Kobo Glo	Theonna, I'm all for improvements, but I think it unrealistic to expect a major paradigm shift in the core of the software. The database is a rational approach in that the information must only be collected once, not every time the system is booted, or files are looked for. This should result in less time overall for various tasks to be performed as in theory the information must only be collected once and only updated as changes occur. The real questions when one runs into over long pauses are what are they doing, and why is it taking so long? As fast as the processor is, one must wonder what takes so long for large shelves to be displayed. My guess is they are processing all of something rather than just enough to start displaying the first page and continuing to process in the background while allowing the relatively feeble human to see and injest that first page. Even a slow system should stay ahead of a human as long as there is no ability to jump far ahead, and even then should only bog down for the time to process one page and display it. With such an approach the lag time becomes shelf length independent and relates directly to the processing time necessary only to process/display one page of the shelf. I have few shelves, and only enough items on them to require a few pages to display the entire shelf.(OK, one with 99 items, but it's quick too) My shelves display quickly. I think they would only require a bit of re-coding to achieve this(if they don't have it already), as long as it is possible to run a second task in the background to continue whatever processing they are consuming so much time with. They might even find the most efficient use of that task would be to only process a few pages ahead, rather than the whole shelf if that would keep up well ahead for most casual browsing and minimize power requirements. I would think an algorithm based on percentage would be most efficient so that as shelves grow ever larger the same percentage of pages are processed ahead by the background task rather than a set number of pages. Last edited by TechniSol; 06-10-2013 at 06:09 AM.