LTS and release methodology

[Pär Lidén]

2008/7/10 Matt Zimmerman <mdz at ubuntu.com>:

> It is already documented, but most people file and respond to bugs without
> reading the documentation (and why should they have to?).  Perhaps the only
> way to track regressions more accurately would be to represent them as
> first-class data in Launchpad.  This is tricky, though, as we've learned
> that the more visible a knob is, the more it is turned, regardless of
> whether it is appropriate or not.  If the data is too noisy, it becomes
> useless (this is why Importance is controlled).


Hmm, yes, I see. It was something like that first-class data thing I meant,
but I understand that I can be quite tricky.


I agree that we need to communicate better around our release activity,
> particularly with LTS.


Good!


> I by no means imply that 8.04 was not ready to use.  There were some flaws,
> which there always are in software releases, and in my view we addressed
> them appropriately with updates and 8.04.1.


Yes, at least for me, every problem of importance I had with the original
8.04 has been fixed now. I'm happy for that, and am grateful to the
developers about it. Maybe I'm repeating my point, and maybe you already
agreed to this, but IMHO there should have been some message somewhere that
the software has not the same level of stability as for example 6.06, and
those wanting that stability should wait till 8.04.1.


> To you, "LTS" may mean "so stable", but to another, it means that problems
> are actively fixed (which implies some change and therefore instability)
> even after release.


Hmm, from what I've understood from the official texts, the LTS are supposed
to be more stable. From the release announcement for Hardy [1]:

Ubuntu 8.04 LTS Desktop Edition features incremental improvements to
familiar applications, with an emphasis on stability for this second Ubuntu
long-term support release

That sounds to me that the LTS should be more stable than other releases
("emphasis on stability"). Maybe I'm reading it wrong; perhaps it should be
understood as "stability in the long term", not neccessarily stability right
when it's released, but over time instead.  I've also understood stability
in these texts as "low bug-count", or something similar. Maybe I've also
misunderstood that. If so, I'm probably not the only one to misunderstand
such statements, and it would be wise to consider making them more clear.

> One thing it can never mean is that there are no bugs in it, for that is a
practical impossibility.

Yes, I definitely understand that, and if the ambition was zero known bugs,
it would take many years before it could ever be release (as the newest
Emacs for example). Maybe a complex product such as an OS could never be
released if that was the ambition.

Thanks for taking your time to answer my questions. I hope my thoughts could
be of some benefit for Ubuntu releases, and Ubuntu users, in the future. And
thanks for the effort you all developers are putting in making Ubuntu the
best Linux distro for my needs (except perhaps stability sometimes).

Regards
Pär

[1] https://lists.ubuntu.com/archives/ubuntu-announce/2008-April/000110.html
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.ubuntu.com/archives/ubuntu-devel-discuss/attachments/20080710/256d57d0/attachment.html>