On Friday 12 October 2012 11:59:56 Rohan Beckles wrote: > Hi Gael -- > > Thank you for your response. Yes, I'm aware of Kwooty - in fact I use it. > :) Unfortunately, Kwooty depends on NZB websites/search engines, and those > services don't index everything which is why I'd like to develop a more > traditional NNTP robot which indexes the headers on the local machine. I > can then search them, and queue items for download. > > However, you're right about sharing code. I'll be looking at Kwooty, KNode > and KLibido for inspiration - this will be my first KDE project. :) > However, the pressing issue is whether Akonadi can handle a full header > pull from a big binary newsgroup. We're looking at around 5 billion > headers from the biggest group. :)
>From my experience with other projects (Massif-Visualizer, KDevelop, Kate, ...), you usually always hit unexpected bottlenecks when working with huge data sets that have not yet been tested. That said, it is usually also quite easy to run a profiler on such test cases to detect these bottlenecks and fix them, which would also improve the "normal" usecase. That said, I cannot give you an educated response to whether Akonadi will hold up or not. I'd say: Try it out and improve the codebase until it works :) Cheers -- Milian Wolff m...@milianw.de http://milianw.de
signature.asc
Description: This is a digitally signed message part.
>> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<
