Interested in helping improve the Fedora desktop experience?

Máirín Duffy duffy at fedoraproject.org
Wed Oct 21 18:21:17 UTC 2009


On 10/21/2009 02:08 PM, Adam Williamson wrote:
> As Colin says, Fedora's general philosophy is and has historically been
> that it's more important to fix the bugs than to worry about temporary
> bodges and workarounds for the immediate experience. This is something
> of a function of Fedora's identity, and I'm not sure whether we want to
> change it significantly. When I was at Mandriva I was much more to the
> 'let's bodge around this to create the best possible user experience for
> this particular release' side of things, but that approach does have its
> penalties: the bodges pile up on top of each other like there's no
> tomorrow (ldetect-lst...grrk) and the person-hours that go into writing
> and maintaining and checking the bodges are time that could have gone
> elsewhere.

These are great points. We definitely don't want build bandaid upon
bandaid upon bandaid of cruft.

Generally speaking, no matter what the specifics are of the issue the
users might run into, do you think the following ideas make sense:

- having an area in get fedora, perhaps the proposed post-download
splash
(https://fedoraproject.org/w/uploads/f/fb/Getfpo-downloadfinishsplash_1.png),
perhaps even further up front, where as issues with the installation
process (and general experience post-install) are reported, they are
listed with workarounds right there

- similar to the above, but have the actual installer some how (i don't
know how yet) give you some fair warning or suggestion when things go
around (things like Colin's bugzilla idea)

So maybe we don't hard code nvidia, or maybe on a server somewhere
there's a listing based on reports that anaconda could try to connect to
and if it detects something... or maybe it's all documentation, there's
no HW detection involved. I don't know. It's hard to suggest specific
ideas to solve the problems before we've gone through the process of
completely combing it and identifying all the problems in play here.
> 
>> I bet if you ask around
>> we could come up with a decent set of heuristics for which hardware
>> devices cause the most angst, maybe maintain some sort of list and if
>> one of them is detected we give the users a fair warning. 
> 
> Well, yeah, we could. It would be very icky and give a lot of false
> positives. It would take constant ongoing maintenance. As I said, I'm
> not sure it's the kind of direction we want to move Fedora development
> in.

I just want to note that AFAIK no rigorous analysis of the user
experience has occurred so it might be a little premature to apply to
off-the-cuff ideas. :) Maybe what I threw out there isn't a great
solution - totally cool, at this stage it's good to pose lots of ideas
and be willing to throw ideas out - use the bad ideas as brain food to
come up with better ones rather than just say no no no :)

I think the install user experience is something that could be improved
using design thinking, so I just threw out that idea in an attempt to
give a specific example of how it may implicate actual interface changes
to the installer. (It may not, and that's fine too!)

>>> Graphics is especially tricky, because linking to a wiki page is
>>> obviously problematic if the driver is failing entirely, since you
>>> wouldn't be able to see our error page...
>>
>> Right, so in that case the implication might be we add some kind of
>> warning to the webpage so the user sees it even before starting the
>> install process. For other issues it might be more visible to place it
>> in the installer itself maybe. (These are just off-the-cuff ideas though!)
> 
> On this specific issue I think it's something we can improve. The
> 'common issues' page is sort of my baby now (it was around for F10 and
> earlier releases but I substantially expanded and revised the format and
> content for F11 and am working on it for F12 too). I maintained a
> similar page at Mandriva for each release (called the 'Errata' there)
> and found that by consistently documenting issues there and referring to
> it in public discussion and release documentation it became a well-known
> reference source which most users would refer to habitually. That's kind
> of my endgame for the common issues page. I'm already trying to ensure
> it's mentioned in release announcements - it was in the Fedora 12 Beta
> release announcement, for instance - and encourage people to add issues
> to it and refer to it when doing forum support, IRC support and so on.
> It would certainly be nice to have it referenced in the install process
> too.

Yep. I think it's an *excellent* improvement to the install user
experience right now that the page is even available. I think as part of
this project I'm proposing, we should look at how we can make it more
visible and useful for folks like my friend so he could have noticed the
workarounds before wasting so much time and suffering so much pain!

How did the similar page work in Mandriva? How did you folks broadcast
it to your user base then? Was it pulled into any apps or marketed in
any way we might be able to learn from?

~m




More information about the Fedora-desktop-list mailing list