How Usenet stores music, and what an NZB actually points at

Usenet predates the web and was built to pass text between servers. Music ended up there the way most things did: by being encoded into messages and posted.

Articles, groups and retention

A file on Usenet is not stored as a file. It is chopped into many articles, each a few hundred kilobytes, encoded as text, and posted to a newsgroup. Providers replicate those articles and keep them for a retention window that now runs to several thousand days.

Reassembling an album means locating every article that belongs to it and stitching them back in order. Doing that by hand would be unthinkable, which is why the NZB exists.

The NZB is a pointer file

An NZB is a small XML document listing exactly which articles make up a release, in which groups, in what order.

<file subject="Artist - Album [01/12] - track.flac">
  <groups><group>alt.binaries.sounds.flac</group></groups>
  <segments>
    <segment bytes="768000" number="1">part1of88@news</segment>
    <segment bytes="768000" number="2">part2of88@news</segment>
  </segments>
</file>

The NZB contains no audio, only addresses. In that respect it is the Usenet equivalent of an infohash: a compact description of where something already is.

Addressing one song

As with torrents, a release usually holds a whole album, so a single track needs a release identifier plus the index of the file inside it. Same shape as the torrent side, different network.

The access problem, and why it matters

Here is the practical difference between the two networks. Playing a torrent needs a client. Playing something from Usenet needs an account with a news provider, because the articles live on their servers and are not public.

That requirement is the single biggest reason Usenet music is less accessible than it should be, despite the catalog being larger and often higher quality. If the fetch happens server side, the requirement disappears for the listener: the audio arrives as ordinary HTTP, and a browser is enough.

Why bother with both

Torrents are strongest on what is popular, because health follows demand. Usenet does not care about demand at all: an article posted once sits there for years whether anyone wants it or not. That makes it far better for deep back catalog, obscure pressings, and the records nobody is seeding any more.

Indexing one network gives you a partial map. Indexing both is how you cover the music that only exists in one of them.