On Mon, Oct 04, 2010 at 01:28:00PM +0200, Julien Vehent wrote: > On Fri, 1 Oct 2010 07:35:05 -0700, Bradley Giesbrecht > <bradley.giesbre...@gmail.com> wrote: > > On Oct 1, 2010, at 1:07 AM, Tom Hendrikx wrote: > > > >> On 01/10/10 08:50, Eric Bischoff wrote: > >>>> I'd be more pessimistic on this figure. The hash driver seems to > >>>> be broken > >>>> and I do not think another release should be done while it isn't > >>>> fixed. > >>>> > > Hi there, > > I just got news from the editor of the french gnu linux magazine, and the > november issue will feature an article on Dspam 3.9.1. > However, since I didn't know about the issues with the hash driver before > sending the article for publication, the described set up will be based on > the hash driver. > > Therefore, I can't but agree that the hash driver issue should be fixed > before a stable version is released. > > Also, somewhere in the documentation is stated that the hash driver is the > only fast enough driver to use SPBH. Is that still true ? > > > Thanks, > > Julien Vehent >
Hi Julien, I have been looking at the this exact question. Going on a sheer token count, SBPH generates way more tokens than the CHAIN. We are still running 3.6.8 and it was not allowed in the code to use Markov-anything with a non-hash driver. Version 3.9.0 does allow it. For heavy use, in would not want to use SBPH but I think that the increased performance of OSB, even relative to SBPH, plus the marked decrease in tokens relative to SBPH means that it would work just fine. Of course, you should do some testing to ensure that it will meet your needs. We are planning to run a new 3.9.x with OSB in parallel with our existing 3.6.8 version with CHAIN until we are happy with the results. Regards, Ken ------------------------------------------------------------------------------ Virtualization is moving to the mainstream and overtaking non-virtualized environment for deploying applications. Does it make network security easier or more difficult to achieve? Read this whitepaper to separate the two and get a better understanding. http://p.sf.net/sfu/hp-phase2-d2d _______________________________________________ Dspam-devel mailing list Dspam-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dspam-devel
