Remaining issues for the multi desktop DVD (was: Re: Dear Fedora board, please open your trac)
inode0 at gmail.com
Wed Jan 12 18:14:36 UTC 2011
2011/1/12 Máirín Duffy <duffy at fedoraproject.org>:
> On Wed, 2011-01-12 at 09:51 -0600, inode0 wrote:
>> Grandma would already be confronted by such a menu using any current
>> installation or live boot media. What you are saying to me is just
>> that all our current media sucks and this one would too.
> Actually she wouldn't be exposed to that menu at all, unless she pressed
> enter to interrupt the syslinux countdown screen. It countdowns from 30
> (I think it would be better to be 10 though, working on that for F15)
> and says hit any key to interrupt.
> (I say the above while at the same time respectfully pleading with all
> of my Fedora colleagues PLEASE PLEASE FOR THE LOVE OF SACRED BEEFY
> MIRACLES AND OTHER HOLY THINGS, LET'S ALL STOP USING GRANDMA AS AN
> EXAMPLE OF A NON-ADVANCED USER. </soapbox podium> :) )
Ok, but we could arrange this media to behave similarly for this sort
of user if it is really necessary to deal with such an edge case.
>> > I think the choice needs to be presented in GDM, with some opportunity
>> > to get help with it.
>> While this has some appeal to me for a unified installation DVD it has
>> no appeal as a unified live media arrangement as it would just
>> introduce a hassle on every boot at a different point in the process
>> and really would be difficult for some spins containing differences
>> other than the DE.
> How does it provide a hassle at every boot? I'm confused here. What I
> think would be a hassle is having to reboot just to try out a different
> desktop.... with a GDM-controlled selection you could simply log out of
> one and log back in, no reboot required. If you're trying out different
> things, having to reboot at every choice I think discourages it. Plus,
> for the use case of an ambassador at a show with only one or two laptops
> to demo things, having to stand there while the thing reboots might be a
> hassle, especially if you miss the syslinux timeout. Depending I am
> thinking it could be a difference of < 30 seconds to log-out / log-in to
> 3-4 minutes for reboot and keeping an eye on syslinux.
We are thinking about different use cases I think. For a running
installed system using the display manager is way way better and a
lovely approach, if only it existed.
For someone who boots a piece of live media they typically will only
play around with it for a while. The next time they use it they will
boot it anyway and maybe think, hey, I'll try this XFCE thing and see
that that is all about. If you aren't installing but are using a live
CD frequently and you want to use KDE it doesn't really matter much to
you at which point you make this choice.
For using a laptop to demo things at a booth I don't really want to
discuss non-existent options. This is better than what we have now and
this exists now. When something better exists I'll lead the charge to
change our laptop setup to use that instead.
>> > Note that in no way am I saying that this is not a good idea, simply
>> > the implementation could use work, and that implementation reflects
>> > (for better or worse) on the Fedora brand.
>> The reaction I would like to see from the board is one based on its
>> trust that these people probably know what they are talking about. How
>> can the board help them make it happen? Three areas have been
>> identified: releng, QA, and design. Whatever issues need to be
>> resolved in these three areas may or may not need the board's
>> involvement, but I think it is clear that encouragement from the board
>> to get this done would help get it done.
> I tried to engage and while I'd prefer to not say anything at all... the
> amount of intense negativity directed towards me as a result of trying
> to engage seriously frightened and completely dismayed me. We need more
> civility around this initiative. When negative energy is unleashed folks
> really need to stand against it and contain it, and not just sit back
> and let it run rampant over and over again. That is what happened last
> time, and I honestly had to set up filters to block some messages in
> order to remain productive.
> In short, the multispin initiative is a good idea and a damn shame it's
> been tainted with such negativity.
I'm sorry that happened.
>> Many of the interested parties in this discussion will be in Tempe
>> very soon. Why don't we just sit down in person there where I'm sure
>> we can do this more productively?
> [snip some great stuff]
>> I also carry around an old 8GB usb stick with this same aggregation of
>> live media on it. I can use it to show things to people just about any
>> time the opportunity arises. Do you think I ever carried around 10
>> live CDs?
> +1, multispin media is an *excellent* idea considering both the
> convenience of the form factor and the lessened environmental impact of
> the media itself, which will be more generally useful than any one
> single piece in the form we do today.
>> Looking at it from the perspective of pressed media the user is always
>> going to be stuck having to make a choice. If we press this in the
>> future there will be a choice made from a menu. Today there is a
>> choice of which of 10 physical pieces of plastic the person wants to
>> pick up and lug home (for those who are lucky enough to hit an event
>> that actually has options for them). If you don't want to make a
>> choice then boot it and get the default, that isn't hard or confusing.
>> If you do want to make a choice then make if from a simple menu, that
>> isn't hard or confusing either.
> Honestly I could live with a syslinux menu to choose between spins
> (although I still think it's far from ideal) to get this out the door
> sooner, but the problem is the menu as it was presented to the board
> some months ago was the squint-equivalent of:
> This kind of sloppiness is simply not acceptable and is avoided
> altogether using separate pressed media, because the sleeve designs of
> the media serve this purpose instead and are written out far more
> cleanly and clearly.
Ok, but what you looked at was a rough proof of concept and what we
have today is a slightly less rough proof of concept. I agree that the
old menu and even the new menu which is better are not examples of the
sort of menu I would like to see. I think if we sit in a room, get
clear on the use cases we are targeting (and those might change a bit
with some good discussion) the design team can help produce something
the project can be happy with. There was talk about dual arch and
mixed live and install media. We need to get very clear on what is
going to be included. What I've been using and what I am really happy
with includes only live media and we only produce 32-bit live media
now so some of the confusing menu issues vanish if that ends up being
the choice for content. Whatever is included people will get a nice
and simple tool to remix their own multi-media as well.
> http://lists.fedoraproject.org/pipermail/advisory-board/2010-October/009421.html for a list of
> suggestions for improvement (they were ignored before, I'm not sure why http://lists.fedoraproject.org
> /pipermail/advisory-board/2010-October/009510.html ), although since we've done some more thinking
> about syslinux's layout for F15 I'd revise it now to suggest the following:
> Welcome to Fedora 14! Select a desktop to try below:
> * Try Desktop
> * Try KDE Plasma Desktop
> * Try LXDE Desktop
> * Try XFCE Desktop
> * Verify media
> I would also suggest giving the new syslinux config & graphics I put
> together a shot to make the menu appear a lot nicer.... here's the files
> and instructions:
>> I hope the board will get behind this effort but I will continue using
>> it to promote Fedora whether that happens or not. This media enables
>> promoting Fedora in ways that don't exist currently.
> Certainly on paper having one more sales agent to help sell your product
> is helpful, but you have to consider whether it's worth it if said sales
> agent makes your product come off basement-produced rather than polished
> and professional. Let's not sell ourselves short when we do actually
> have the skills and professionalism to do it right...?
I completely agree. However, for some of the uses I've described even
a basement-produced product is better than the existing polished
product we have. It is only one menu that lacks polish and the person
I'm showing this to doesn't even need to see that menu, then it boots
right into what you see by booting the spin directly. Let's clean up
that extra menu layer though by all means so new users can boot it
themselves and we aren't unhappy with what they see.
More information about the advisory-board