[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: lots of feedback on the test-list

On 12/03/2012 07:40 PM, Máirín Duffy wrote:
On Mon, 2012-12-03 at 13:44 -0800, Adam Williamson wrote:
All of them. So
there's a fairly strong burden on us, as the ones trying to Do Something
Different, to both make our Different thing worth the effort, and
minimize the effort just as much as we possibly can.
I know you didn't mean it this way, but I do want to be clear that we
didn't change the way we represent the custom part UI just to be
different. I think maybe some users are getting frustrated because they
are assuming this was the case, and I completely sympathize because I'd
be frustrated too if that was the driving goal here - thankfully it's

The idea was to enable folks who don't understand all the intricacies of
the technologies to end up with the final disk layout they wanted and
still be able to take advantage of the tech we can offer. I think no
matter your level of expertise in the specific tech, you likely
understand what you want your filesystem to look like afterwards, so the
idea was to start from that common ground and hand-hold them through
filling out the tech. Even some of the sysadmins I interviewed when
putting the mockups together couldn't remember what all the different
RAID levels meant, for example. :)

I think the biggest weakness, based on the feedback, is our ability to
handle pre-existing configurations that folks want to re-use. It wasn't
a priority for the design (and it maybe should have been - that is
absolutely my bad,) but is something we should support. Hopefully we can
improve that experience by learning more from folks running into issues
what exactly they were trying to do & how exactly their previous system
config was set up before they ran into issues.

But I do think we can improve things within the design itself as well.
Like I said on IRC, the conclusion isn't 'this layout isn't working,
ABORT ABORT ABORT', but the conclusion may be 'if we put our thinking
caps on we can probably come up with ways to refine the design so it's
more discoverable and less scary'.
Yep, I agree 100%. I don't think anybody is saying what we have is
perfect, and I will be the first to admit I don't think the design is
perfect either - you always have to iterate to get to a good design. I
am sure even Jonathan Ives and the other Apple designers iterate over
and over - because that's the design process. We have a good base design
and we need to simply continue refining it to make it better.

Another thought - I read a book called "the lean startup" recently - in short, it's about delivering small, incremental bits of a product to the customer - getting feedback as you go along - so that you wind up delivering *the right thing*. In the business world - the point of this is avoiding wasted time, since so many companies wind up building something that entirely winds up just not being what anyone wants - waste of time, waste of money, etc. In the open source world, we loosely refer to this as "release early, release often," though a lot of that is more along the lines of transparency and not always highlighted as "getting the right thing out the door."

There were, thanks to mo, tons of mockups and such posted - but I think there's still, in those early stages, a bit of non-reality in that mockups don't actually have the same feel as when you're actually installing your machine. And it's sort of difficult to say, "let's just give the first 2 steps" when part of what you want to be doing is making sure the full install actually works as well, and need to get to milestones like alpha/beta/etc. to even get it in front of people. But now that we have the base design, as pointed out, and now that there are significant numbers of folks who have actually had hands-on experience, I wonder if showing mockups (or even implementing) just one or two small bits at a time - so people can try them out, give feedback, and then move to the next few small bits after those are right - would be a good idea, rather than trying to incorporate ALL the remaining feedback at once?


Anaconda-devel-list mailing list
Anaconda-devel-list redhat com

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]