Ron Stodden wrote:
> On Sat, 22 Jan 2000, Civileme wrote:
>
> > That is the purpose of a beta, to eliminate bugs with user
> > discovery.
>
> Yes, but the starting point for a beta is a completed alpha. Alpha testing
> is done by the developing orgainisation and includes testing both at the
> module level and module integration (system) levels. Testing is
> against the written specifications. During this process, all discovered
> bugs are corrected and retested. Only when no bugs known to, or
> discoverable by, the developer level exist can the alpha test be regarded as
> completed and passed. This is a pre-requisite for the beta test.
>
> Beta testing is done by outside users starting against the developing
> organisations written specifications, who approach with a different set of
> requirements and a different mindset to developers, and therein lies the
> purpose and value of a beta test. A product may well fail its beta test, in
> which case the developing organisation can be put right back to square one.
>
> People validly complain if any alpha-level (ie developer-discoverable) bugs
> are discovered in beta software - and Air 7.0 and its install program are
> liberally scattered with them, unfortunately.
>
> I'm talking about well-accepted (for decades) industry conventions, of which
> MandrakeSoft does not appear to be cogniscent. I am sure their funding
> organisations should be informed of that problem.
>
Oh now I wouldn't go so far as that. I do think that the development model they
follow is not the classical one, and the terms "alpha" and "beta" may be
misapplied from your perspective. But "Do it that way cause we've always done
it that way," is not an argument that will sway me to change a process.
>From my perspective, Hmmmm
1. Well I have some annoying bugs I don't have time to deal with. Not large but
nevertheless irritating.
2. Three days ago, a CPU did a HCF (halt and catch fire) on me because a fan
bearing died and the user didn't complain about the groaning noise. The disk
was so badly corrupted that fsck didn't even try to run on boot. I installed
the hard drive from the incinerated computer in another machine and ran e2fsck
-c -b (whatever) /dev/hdb1-11 on it. I then moved it to a new computer and lo
and behold... It worked. No reinstall, no loss of data. Everything checked.
3. That has to be one of my standards as a customer. Any system that won't do
that is out the window.
Anyway, I didn't write this to criticize your point of view. Far from it.
I wrote it to suggest we ask the folks at Mandrakesoft to invest a small amount
of time into preplanning the next release with some of us. Even a few chat
sessions online to plan the testing/standards/objectives, to imagineer it if you
will, would improve the relations between Mandrake and some of the testers and
users which seem to be becoming strained in the headlong rush to move product.
Quality of what is moved is equally important to getting it on the shelf.
Best regards,
Civileme
>
> --
>
> Regards,
>
> Ron. [AU] - sent by Linux.
