In reality, Debian unstable is sort of like Red Hat 8.0 or 9.0. Debian tends to lag their "stable" title behind. Many people use unstable with minor or no issues.

Check this link for more info though.

http://www.debian.org/releases/

Interim

Debian unstable only accepts stable software releases (no "in development" type of stuff, for the most part). Once Debian has decided that a package doesn''t cause too many problems, works everywhere with the desired results, and is properly packaged (there are strict guidelines for this): it makes its way to Debian testing. So, Debian unstable is like other distributions'' stable releases: it''s all stable software, but it isn''t tested all that much yet.

I run Debian unstable on my desktop. The only problem you might find is when a large change is being made to Debian, because it happens in unstable over a couple days (or a week or two, for large collections of packages: like when Debian''s KDE was moving to 3.x from 2.x).

For example: certain packages couldn''t be installed for a day or two when the C++ ABI shift was taking place since something they rely on had updated and they hadn''t, or vice versa. This isn''t that big of a deal (especially since nothing really large is happening right now to Debian), it''s just an inconvenience if the package you want to install happens to be broken today; you can get an earlier release from testing or the Debian archives, or wait a day or two for it to be fixed.

"unstable" will always have software that is at least as updated as, or more updated than, software in stable. Eventually, the current "unstable" branch is renamed "stable", the old "stable" branch is thrown away, and a new "unstable" branch is created (copied over from the new stable, mostly). "unstable" and "stable" are different versions of debian, not categorizations of packages. Thus, you should only have unstable sources.

How appropriate. You fight like a cow.

quote:Original post by -vic-
One more thing: in my sources.list i have only respositories for unstable. I didn''t want to put the stable/testing repositories because i think the mix of them will give me some trouble, which i don''t want, so i use only the unstable repositories.

I only keep one official respository in my sources.list at a time, but mostly because the official repositories'' package lists are pretty large (and when I update everyday, it just adds up). There''s little reason to add additional repositories to your sources.list (except for maybe Debian''s experimental repository: but those packages normally aren''t in unstable for a good reason ).

There are methods of telling apt which repository to prefer for multiple copies of a file, if you did want to have more than one repository in your sources.list file.

quote:Original post by Sneftel
Eventually, the current "unstable" branch is renamed "stable", the old "stable" branch is thrown away, and a new "unstable" branch is created (copied over from the new stable, mostly). "unstable" and "stable" are different versions of debian, not categorizations of packages. Thus, you should only have unstable sources.

Unstable never becomes stable (it''s always "Sid"), but testing does (testing is "Sarge" now, it used to be "Woody" among other names; stable is currently "Woody"). Although, if you consider the piece-by-piece transfer of unstable to testing, I guess you could say unstable becomes stable over time .

quote:Original post by Null and Void
Unstable never becomes stable (it''s always "Sid"), but testing does (testing is "Sarge" now, it used to be "Woody" among other names; stable is currently "Woody").

Not true. I remember when Potato was unstable. Then Potato was stable, and Woody was unstable. Codenames remain with a particular release, whereas designations such as "stable" and "unstable" are sliding.

How appropriate. You fight like a cow.

quote:Original post by Sneftel
Not true. I remember when Potato was unstable. Then Potato was stable, and Woody was unstable. Codenames remain with a particular release, whereas designations such as "stable" and "unstable" are sliding.

I can''t comment on what was done historically, but the next stable release will be ''sarge'', which is currently ''testing''. ''sid'' is ''unstable'' and will be the stable release after next.

True, the methodology has changed. My point was that a particular codename (Sid, Woody, Hamm, etc) is not indelibly tied to a classification (stable, testing, unstable), which NaV had seemed to imply in his last paragraph.

How appropriate. You fight like a cow.