Maintainers Responsibility (was alpha/beta software in Fedora 8)

[Toshio Kuratomi]

Bastien Nocera wrote:
> On Thu, 2007-11-29 at 09:43 +0100, Nils Philippsen wrote:
>> On Wed, 2007-11-28 at 11:31 -0500, Matthias Clasen wrote:
>>
>>> The importance of individual packages is pretty dependent on the use
>>> case. For most users of a Fedora desktop, it is probably far more
>>> disastrous if the email client or web browser crashes, than if some dns
>>> server the haven't installed has some bugs.
>> This assumes that Fedora is "just" a desktop, while it is much more --
>> the desktop is "just" a part of it. A significant one, surely, but a
>> part. From my POV, Adam was okay to put in that version of bind because
>> he tested it and found no problems, and most definitely not because the
>> bind package is somehow less important than say the gnome-panel.
> 
> Read Matthias' reply again. He made no such assumption, you did: "For
> most users of a Fedora desktop [...]"
> 
It really depends... If DNS crashes and Fedora is the DNS server for 
your subnet your end users might not be able to get to any network sites 
whereas if your email client crashes, well at least you have a browser 
for web mail.

On the flip side, If DNS has a few bugs that only hit in obscure cases, 
how is that any more or less disastrous than if evolution only has a few 
bugs that only hit in obscure cases?

So my real issue is with the naivete of the post Matthias was replying 
to -- trying to judge importance of a bug based on whether the app is 
for the "desktop" vs "server" is a false dichotomy.  You have to judge 
based on what happens when the bug occurs, how it is provoked, how many 
users (rather than computers) are affected when the bug occurs, etc. 
"desktop" and "server" might help you determine these answers but it 
isn't an answer in and of itself.

-Toshio