Re: Volume Sets

On Sun, 13 Apr 2003, Tyler Dauwalder wrote:

I just realized that I haven't replied to this mail. Not that I have
much of substance to say, though. :-)

> Lest things quiet down too much around here, I thought I'd let you
> know what I discovered about volume sets.
>
> Firstly, none of the multi-CD sets I looked at (MS VS.net, MS VS 6.0,
> MS Office 2002, Riven, Myst) were true volume sets a la the
> standards. In the case of Microsoft's CDs, they *did* set the volume
> set identifiers on each CD, but to unique values that matched the
> corresponding regular volume identifiers (i.e. MSVSNET1, MSVSNET2,
> ...). The latter two sets mentioned didn't even bother setting the
> volume set identifiers.
>
> So, I guess there's no need for us to worry about supporting true
> volume sets, as it would appear that no one else does either. Perhaps
> someday we might want to look into it...

At least for multiple CDs, I see little benefit.

> Lastly, as to multiple volumes on a multisession CD, I discovered an
> interesting bit in the UDF specs that specifically states that only
> the volume in the "first recorded track of the last existent session"
> shall be presented. It also notes that these are the same semantics
> as those used by iso9660.
>
> So, it sounds to me like we ought to present only the last iso9660
> and/or UDF volume on a multi-session CD by default, but make any
> others available (since I think that's a very handy feature to have).
>  I don't know that this is anything we need to offer in the Disk
> Device API though; Tracker could just figure things out on a device
> by device basis. Also, in the case of other filesystems (like the R5
> CD with a BFS image), I think we should just present them all like is
> currently done.

I don't like the idea of the Tracker needing to undertake special
examination of devices. How about the (my :-) idea to reflect those
relationships of partitions in the Disk Device API by methods like
BPartition *BPartition::Supercedes()/IsSupercededBy()?

CU, Ingo