On 2006-06-13T06:03:30, Alan Robertson <[EMAIL PROTECTED]> wrote:
> It is highly non-intuitive, to say the least.
Agreed.
> I would estimate that 80% of the people who used it would never figure
> it out without being told, or encountering significant frustration.
> IMHO, that's an 80% failure rate for the GUI. That seems pretty
> horrible to me. Since the the GUI is _supposed_ to be easy to use,
> that's not too good...
75% of all statistics are entirely made up though. Our testers figured
it out fine, even though they admittedly didn't like it. (Neither did
I.)
And ultimately, we can only cater so much to people who don't read the
documentation, or don't at least have a basic understanding of how
clusters and/or our software works. You know what they say about idiot
proof software, and I'd rather not go down that path. As simple as
possible, but no simpler.
> >I understand. My concern is that now this change is in CVS a "works for
> >me" mindset will be adopted.
>
> I understand that concern. But I think that would imply that your
> proposal was superior. And, it clearly isn't. Sorry.
>
> You are making a proposal which is a huge amount of new work, incurs
> great risk, which in the end will be inferior, and you wonder why we're
> not jumping at the chance to implement it.
Well, the GUI team works for you, so it's ultimately your decision. And,
clearly, GUI useability is all about opinion and personal taste (unless
you get in real GUI research, which is quite costly, and the videos tend
to be very painful for the developers to watch ;-), and you tend to have
very strong opinions - so it's mostly a mood point to argue about. It is
NOT something which we can back with hard facts or statistics, so let's
not pretend we could.
Still, let me voice my opinion one final time; if you think I'm not
adding anything new, feel free to not reply:
I do wonder whether the current way of creating clones or m/s resources
is intuitive to anyone who has, in fact, read the documentation, or even
to anyone who hasn't. (It actually confuses the heck out of me.
Primitive? I don't want a primitive!) Or whether it would actually allow
you to create a cloned group sanely.
And, shouldn't a group creation workflow lead the user to a meaningful
result, which would seem to imply that there are resources within the
group?
A two-stage work flow for creating complex resource containers, which is
consistent and cascadable (thus actually re-using several screens in the
GUI, which strikes me as good) would solve these problems in an obvious
manner. Or so I believe.
The GUI is supposed to make the system easier to use, that is true. But
the way how it models the configuration also shouldn't be too far
removed from how the system is really configured. (We can see where that
leads looking at how clones are represented right now.)
But, that's just Andrew's and my opinion. It's your call, you're the
boss when it comes to the GUI.
While we are at it, I personally don't believe that calling matters of
taste or opinion "inferior" or "superior" is something which is likely
to lead to amiable and friendly atmosphere, btw ;-) I will grant that
you don't think it is superior (just as I prefer my own), that's
alright, but saying that it _will_ end up being inferior is making a
non-subjective statement.
Sincerely,
Lars
--
High Availability & Clustering
SUSE Labs, Research and Development
SUSE LINUX Products GmbH - A Novell Business -- Charles Darwin
"Ignorance more frequently begets confidence than does knowledge"
_______________________________________________________
Linux-HA-Dev: [email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha-dev
Home Page: http://linux-ha.org/
