[SoaS] Planning towards SoaS v3
pbrobinson at gmail.com
Sat Jan 23 10:17:16 EST 2010
>> Can you say more about how and why the boot helper is going to be buried?
> >> I've got a ton of older hardware that works just fine once its booted,
> >> some of it zips, but doesn't have the capability to boot from USB.
> > When was the hardware released? What's the details of the hardware? Just
> > about anything released in the last 10 years should be bootable using USB
> > although it might need a Bios upgrade. The lowest level of hardware CPU
> > support in Fedora is i686 and the vast majority of that hardware supports
> > booting usb. Upstream supports a much wider variety of hardware with a
> > lot less of the issues we see.
> In theory this may be true. In practice, it is not reasonable to
> expect that (a) a child or teacher will be able to update the BIOS or
> (b) the child or teacher would have permission to update the BIOS.
> Most machines in schools, libraries, cafes, etc. are locked down. One
> of the beauties of SoaS is that it lets the learner/teacher do an end
> run around such controls.
I've found that most devices in Libraries are locked down not to boot off
anything other than main hard disk. So if they can't boot of USB its
unlikely they'll be able to boot off the CD either so you still have the
same issue. In the case of the child or teacher not being able to upgrade
the BIOS, I wouldn't expect them to in that situation, in most of those
cases it would be a case of opening a dialog with which ever department
supports the devices. And whether its booting of the CD or USB the ultimate
eventuality is that its to run a 3rd party OS, and all of that comes back to
the policy of that particular institution.
> > So in short its not going to disappear but as it seems to cause so many
> > support issues on the sugar lists I would like to see a concerted effort
> > find out why its needed, fix those problems and to generally de-emphasize
> > the boot helper CD as the fix to every boot problem.
> Is that really true? I would love to see some data to back that up.
> Not being able to boot causes "so many support issues", but what
> percentage of those issues are due to the CD?
Well its seems to take up quite a bit of the discussion on the soas list and
it seems to be most of what I read about from those lists. So in the very
least my perception is that its a problem. The area of focus for SoaS is
much smaller than that of Fedora yet this particular issue seems to take up
a lot more of people's time.
> We need to also consider it if (a) there aren't going to be new
> support issues that arise due to the BIOS problems people will
> experience and (b) whether the CD support problem isn't primarily due
> to a mismatch between the CD image and the USB image, which could be
> perhaps remedied in some other way, or at least more readily
> identifiable as the problem to the end user.
My point is that my belief is that most of the issues we have with not being
able to boot off USB are bugs (some minor some more so) that need to be
addressed to ultimately give the user a better experience. A CD to boot a
USB drive isn't a good experience compared to just a USB key. You can't
shove a CD in your pocket, its more delicate and it ultimately doesn't give
the user the experience that SoaS is suppose to give. A portable Sugar
experience you can take everywhere.
> Maybe a few years from now, once the MS campaign in support of LiveUSB
> boots takes off, there will be more industry/marketing support for
> updating the BIOS, but for now, to the average user, it is not
> something that they are comfortable with.
> > For example I've seen you have issues booting Apple devices with the
> > Every single Intel based Mac uses EFI to boot and hence should work out
> > the box with SoaS and shouldn't require the helper CD. If it doesn't work
> > need to work out why and fix the bug, not plaster it over with an ugly
> > In general it makes people's experience of using Sugar on a Stick less
> > pleasant and hence less likely to keep on using it. Its a constant
> > and problem on the lists as well.
> Not being able to boot does make the experience less pleasant. But
> removing boot options is not the obvious fix, IMHO.
Maybe I could have worded it better. It seems to be the fix to any boot
issues is to use the boot helper CD so I would like to see it de-emphasised
as the fix to everything so that in the very least we can try and fix the
issues rather than constantly plastering over them. If we don't find out
about the issues we can't fix them so we will never get to the case were it
can go away. The Mac case above is a perfect example. When in actual fact
rather than working around it with a CD a bug should have been filed and
assistance from the EFI people upstream sought. It seems that the boot
helper CD is the first fix to any of the issues rather than a last solution
like it should be. Another case is the fact that both of the last SoaS
releases haven't booted off a USB key on my eeePC 901..... no option for a
boot cd there. I think we'll remove the vast majority of issues with boot by
fixing a couple of bugs. So if we focus on that for a release and get some
test cases (Caroline apparently has some).
The major issue that I see from my perspective is that hacks seem to be
fixed with other hacks rather than putting a proper solution in place, and
people that try to assist in getting proper fixes in place get shot down and
end up throwing in the towel and going off and working on something else
because its easier. I've spoken to a number of people that have all
expressed this concern. So I'm not proposing (and it should have been
something like "work towards making the boot helper CD dead so it can be
buried" ) to kill to boot helper CD, I'm wanting to try and get some stuff
in place to fix the issues like that on the Mac's that are causing the need
for boot helper CDs where there doesn't need to be any.
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the SoaS