> > IDE drives using DMA have nearly all of the same performance advantages
> > as SCSI, except perhaps that you have to manually enable some of them
> > using hdparm.
>
> Parp. Benchmarking I've done says they are at best 65% of the speed of
> a good SCSI drive. Once they have proper command queueing then the reasons
> for most people using SCSI will vanish - but not yet.
In this scenario, IDE technology just reinvents the SCSI technology.
That's just time lost in my opinion, and such imaginary IDE drives
would be as sophisticated as SCSI ones and so should not be cheaper.
History is sometimes stupid, but it is reality.
Remember some company that recently reinvents "preemptive scheduling"
and "32 bit addressing".
Gerard.