"currently caching files from the start up to where you read"

[Harald Barth]

> Does this mean that if you open a file and seek to the middle of it
> then the arla afsd will have to read the first half of the file
> anyways?

As soon as you read from any place in the file (for example the
middle) everything from 0 to that position has to be fetched.
This is a bit of a pain for file formats that have the index at
the end (mp3, zip).

This has the side effect that your files can not be bigger than
the cache size either.

Tomas is working on block caching and we hope to raise enough funding
so that he can finish that project. If there is someone out there who
can support this (money, coding, whatever) contact Tomas or me.

Harald.