Expectation Management for Test Releases

[Miguel Pereira Leite]

Everyone that tries to install a OS must know a bit... even more true
when it comes to Linux. A simple MS office user don't try it.. So, it
also knows what "beta" means.... There is no escuse for those who say
that only seen the disclaimer *after*...

Sorry for my english

A Ter, 2004-04-20 às 21:47, Keith Lofstrom escreveu:

> (cc. tfox at redhat.com, Fedora documentation leader)
> 
> Keith wrote to the fedora-test-list:
> > Let's work on those disclaimers, okay?
> 
> Alan Cox replied:
> > It has them, when it boots. And nobody knew it didnt boot on some systems
> > until they tried it - thats what testing found.
> 
> Here is the disclaimer, AFTER booting "linux text" (graphic is the same):
> 
> 
> --------------- Warning!  This is pre-release software ! ----------------
> 
>      Thank you for downloading this pre-release of Fedora Core.
> 
>      This is not a final release and is not intended for use on
>      production systems.  The purpose of this release is to collect
>      feedback from testers, and is not suitable for day to day 
>      usage.
> 
>      To report feedback, please visit:
>   
>         http://bugzilla.redhat.com/bugzilla
>   
>      and file a report against 'Fedora Core'.
> 
>                              [ Exit ]    [ Install Anyway ]
> --------------------------------------------------------------------------
> 
> That's it.  That's all.  It says nothing about "updates don't work" or
> "box of broken glass" or "test releases eat babies".  And to get to
> that screen, a hell of a lot of code has to work - at least booting
> and anaconda and so forth.  If the installer isn't working, this message
> is inaccessable - in effect, no disclamer is provided anywhere in the
> process, to counterbalance the invitation to participate in testing on
> the web pages.  No suggestion is made to hunt through thousands of list
> messages to perform an attitude check before joining in the fun.  And a
> potential tester has to spend the hours to days necessary to download
> the ISOs before this message is even potentially accessable (this posting
> might help advanced Googlers find it).
> 
> I am asking to place information on the download page, warning people
> before they go through the work of downloading and attempting to help with
> testing that they are not invited to participate unless they understand
> certain things and meet some qualifications.  Potential participants also
> need to be warned that 30 years of experience testing software in other
> venues does not necessarily apply here;  the rules for Fedora are different.
> 
> The Boeing gentleman I mentioned wasn't trying to use Fedora on a
> "production system"; he was attempting to offer his help in testing, as
> he did professionally for his company for internal and external software.
> He was attempting to relate his knowledge of hi-rel software testing to
> what is happening here, and was disappointed about the waste of his
> efforts.  If people like that are not welcome here, at least let them
> know *why* in a kind way, because their professional impression of us
> matters to our own future.
> 
> I can rewrite the webpage at http://fedora.redhat.com/download/test.html
> to include some of these rules - the HTML style is easy to emulate .
> I am open to suggestions as to what rules I should offer.  Otherwise, I
> can extract some of the verbiage from the mailing list (do the authors want
> to be attributed or anonymous?).  If some of the attitudes on the mailing
> list are made obvious to potential testers before downloading then there
> will be fewer downloads (freeing up the mirrors for the people that need
> them ) and fewer unqualified testers annoying you with questions.   In the
> longer term, there will be less hostility and more acceptance of Linux.
> 
> If you believe restricting testers with accessable warnings will result
> in fewer interruptions and a better general release, here is a chance to
> test that idea.  If you believe in restricting testers through surprises
> and censure, be prepared for lifelong professional consequences. 
> The web does not forget.
> 
> Keith
> 
> -- 
> Keith Lofstrom           keithl at ieee.org         Voice (503)-520-1993
> KLIC --- Keith Lofstrom Integrated Circuits --- "Your Ideas in Silicon"
> Design Contracting in Bipolar and CMOS - Analog, Digital, and Scan ICs

-- 
Miguel Pereira Leite <mpleite at ppl-ti.com>
Proença & Pereira Leite - Tecnologias de Informação, Lda
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/fedora-test-list/attachments/20040420/78fbb7c6/attachment.htm>