>Thanks Thomas, looks very interesting indeed!

It got the best results in the german c't magazine, which is still testing on technical grounds - after we already decided on this as our defragger <g>.

>I'm trying the defrag by name on my data drive right now, as a first go-round.
If the disk has less than 20% free space and is heavily defragged with LARGE files I go for a space run before defragging by name or date - but if there is enough free space I do it in one run.
Be sure to decide *for each partition* which is the best strategy. And take the time to familiarize yourself with defragging system files either at boot time or in multiboot systems. One of the worst things that can happen is a heavily fragmented swap disk and/or registry.
>
>It sure is nice to have options for defragging. I have the impression that the standard defragger simpy does it by first-fragmented-to-first-free, or maybe even some simple algorithm to have itself run as fast as possible.
>
>This surely is an area where VISTA could make huge improvements with virtually no user impact, by having some kind of smart allocation of file extensions and carefully sprinkling free areas for that purpose. Some user options to control that kind of stuff would be even better.

Agreed - my Q'D way of reserving whole partitions is not for the masses. But MS strategy is not for speed but for gadgets: how on earth can it happen indexing is mostly set to "on" on systems I access the first time ? How often do you search something on your disk and need it so fast that it makes "sense" to slow every write operation ? Teach users to think about useful directory structures instead of slowing down the whole machine!

regards

thomas