On Mon, Jan 02, 2006 at 06:14:09PM -0500, Benjamin Mesing wrote:
> However I have often heard complaints about broken dependencies and
> broken software in testing.
Sure. Testing isn't perfect, it won't guarantee your software is bug
free; and furthermore it changes daily, which means which bits are broken
will change from day to day too.
> From what I have heard, I would not like to
> go with testing for my system.
Hearsay isn't exactly the best thing to base your conclusions on,
especially when you're trying to promote them to a large group of people.
> Therefore I see two possible choices for
> the end user:
> 1. stable for a stable system which is totally sufficient for the
> average user (who has very little knowledge about computers and
> does not care if feature XYZ is already available...)
By and large, Debian isn't that great a solution for people with "very
little knowledge about computers", unless they've got someone looking
after them. If you look at all the people who run stable, I doubt you'll
find many who fit that description.
> I think one of the reason why testing is doing bad compared to unstable
> is, because serious bugs usually get fixed in unstable pretty fast,
We've got 600 RC bugs in unstable that are older than three months. There
are (an estimated) 450 or so in total in testing. I don't think it's remotely
fair to say that
Testing is very effective at minimising problems that can be automatically
measured; if you think testing has problems that can be done better, and you'd
like to see fixed, I'd suggest finding some way to measure them.
Creating a list of packages that need to be installable together would
probably be a good idea, although obvious answers ("everything optional
or above -- policy says so!") are likely not to work, because it relies
on unstable being more free of breakage than it in fact is.
Cheers,
aj
Attachment:
signature.asc
Description: Digital signature