Expected velocity and volume of changes to a library's data

[pfarrell]

Starting a new thead because the drift is pretty big, and
I think this is a good topic.

The quick question is: How fast does how much of
a library change?

In [slim] Re: How to reliably/easily backup 100Gb of FLACs
On Sat, 2005-11-12 at 09:03 -0800, MrC wrote:
> representative of what will occur with music files. Since only
> metadata is being changed at the beginning of the file, binary diffs
> are fine.
> As Jim earlier pointed out, this is not constantly changing binary
> data in general - only the tags change, so all you need is the tag
> information and one set of music files.

Without doing any analysis, I would guess that the files in my library
are extraordinarily fixed. Nearly read-only from the start, and
definitely read-only after a month or two.

I'm guessing, but I would expect that 99% of the CDs' FLAC files
are never adjusted at all after I get the tags close.

This is clearly not like traditional IT backups, where
there are lots of changes to most of the files.

Seems to me that this should drive different tools and
criteria in the backup system.

Ideas?

--
Pat Farrell

http://www.pfarrell.com

--
Pat

http://www.pfarrell.com/music/slimserver/slimsoftware.html

[sben]

Expected velocity and volume of changes to a library's data

On Nov 12, 2005, at 9:28 AM, Pat Farrell wrote:

> Starting a new thead because the drift is pretty big, and
> I think this is a good topic.
>
> The quick question is: How fast does how much of
> a library change?
>
> Without doing any analysis, I would guess that the files in my library
> are extraordinarily fixed. Nearly read-only from the start, and
> definitely read-only after a month or two.
>
> I'm guessing, but I would expect that 99% of the CDs' FLAC files
> are never adjusted at all after I get the tags close.
>
> This is clearly not like traditional IT backups, where
> there are lots of changes to most of the files.
>
> Seems to me that this should drive different tools and
> criteria in the backup system.
>
> Ideas?

A different use case is the "slow creep" user. I want all the tracks
in my iTunes library to be rated; however, I don't have the patience to
sit down and rate them all. So, my iPod syncs with a "fifty random
unrated songs" list, which I listen to while commuting; I rate songs as
I hear them. Every couple of days, I get on the order of 50 songs
changed (slightly! just the ratings!). Most other tags are closer to
your use case: Now that I've got them the way I want them, they almost
never change.

Yet another use case is the user who hasn't yet figured out exactly how
they like their library tagged. This user will occasionally (probably
rarely) change tags across a significant fraction of the library (e.g.
changing how multiple artists are represented; changing how multidisc
albums are represented; adding composer information).

In all cases, I think there's going to be a gradual accumulation of new
files.

I suspect your use case is the most common. These other two are
getting closer to a "standard" backup scenario.

-- S. Ben Melhuish
sben (AT) pile (DOT) org