[Sugar-devel] [IAEP] SLOBs Position on SoaS

[Douglas McClendon]

Douglas McClendon wrote:
> Bill Bogstad wrote:

>>
>> I also don't think we can leave Sugar LiveUSB to any distribution.
>> My impression is that both LiveCD and LiveUSB Linux distributions are
>> essentially gimmicks for all of them. 
> 
> I generally agree with the rest of your sentiments in this mail, but as 
> the now quasi-official 'Godfather of Fedora LiveCD' I have to respond to 
> this 'gimmick' claim.
> 
> It should be noted that historically the linux LiveCD (and the alternate 
> more usable but less compatable LiveUSB form) solved this problem for 
> linux distributions, and still solves it to this day-

I should also add for completeness the use of LiveCD/USB for rescue 
purposes, extended to general stateless utility with most features of 
the traditional OS.  I.e. boot, do whatever, use whatever, and know that 
the next boot will be identical (or can be if you reset the persistent 
overlay).

And of course portability is another counter to the gimmick tag.  Or 
rather, LiveCD/USB is what simultaneously required and pushed the 
distros down the path of making dynamic boot time hardware detection and 
autoconfiguration work much better than before the LiveCD era.  I.e. 
what might enable a traditional install to external usb rotating disk to 
be even remotely usable.  I.e. a very historic distro install to 
external usb rotating disk would be tied more or less to the hardware it 
was installaed on, and would not work well/at-all if moved to a 
different system.  But with the natural progression of LiveCD 
requirments of dynamic boot time hardware autodetections, leading to 
that just being a generally good idea, now it is more feasible to use an 
installed to external usb drive distro as a portable thing that can move 
from across different PC HW configurations.

An interesting point is that _if_ 5 years ago external usb flash drives 
were in fact performancewise on par with rotating disks, the LiveUSB 
case probably would not have evolved as it has.  I.e. you would just put 
a hardware agnostic install in an image and distribute that, possibly 
with some built in recovery subpartition.  While that may be where 
things are going, it is worth noting that much of the work involved 
getting that hardware agnosticism to work.  And the possibly transient 
LiveUSB implementation did perhaps play a key enabling role for the 
long-term, as it was a useful place for that hardware agnosticism to 
take root.

In the same vein, I see the current manifestation of soas and fedora 
liveusb as possibly transient in the long term, however their 
functionality and use in the short term, may be what drives the 
development of whatever obsoletes them.  I.e. still quite relevant and 
important.  Maybe this is analogous to how the XO-1 can still take 
credit for sugar.  I.e. even though it may have been a transient and 
obsoleted platform, it was necessary and critical for the overall 
lifecycle of the sugar software.

Personally, and this is the perspective of an interested but 
non-real-user of sugar, it seems like the soas that Sebastian is doing 
has been critical for sugar's long term lifecycle, and even in its 
current form will be for at least the next year or two.  And beyond that 
if btrfs as the default rootfs or some other technilogical progression 
causes significant changes, the work done for soas in its current form 
will probably live on, just in an evolved/morphed form.

Or I could be completely wrong, I speak for no one but myself and 
probably do way too much of that.  $0.02...

-dmc


> 
> Linux support for a wide variety of hardware is excellent, but on 
> average lags by months and years, and even widely varies from 
> distribution to distribution based on a bazaar of differing priorities.
> 
> People used to sit down with several redhat install CDs, spend a long 
> time, feeding CDs, possibly fighting partitioning nightmares, only to 
> end up with a linux install that didn't (yet) support some favorite 
> piece of hardware.  Then, reading forums, the user might try again with 
> a different distro, and just repeat until either giving up in 
> frustration, or finding an acceptable result after hours or days of futzing.
> 
> LiveCDs gloriously solved this problem.  Users could completely non 
> destructively try out the entire OS, albeit quite slow (or not so slow 
> with LiveUSB), and trivially discover whether or not this particular 
> release of this particular distro was suitable enough to justify 
> installing permanently.
> 
> Then the LiveOS installer came into play, which made things even much 
> more convenient.
> 
> Then my rebootless LiveOS installer came into play, which still hasn't 
> caught any great press or reviews, but I still think will when enough 
> people try it.  But that is entirely beside the point of this historic 
> background discussion.
> 
> I also implemented LiveUSB persistence for Fedora, which apparently 
> quite a few people actually use, even though I'll be the first to agree 
> with your below words, that it is probably not, and may never be a 
> solution for millions of users, _in and of itself, without installation_.
> 
> To me, I view LiveUSB, with or without persistence, as primarily an 
> _installation medium_.  It is something you can carry on your keychain, 
> walk up to a random PC configuration, use to test compatability, and 
> then at your option, permanently install to the local system disk, or an 
> external usb (rotating) disk, or even an external usb flash disk. 
> Though as mentioned in a prior mail, I don't have enough personal 
> experience with the installed to flash case.  I hear anecdotes that it 
> doesn't work too well with old crappy flash, and that features of btrfs 
> in the pipeline should make it work much better in the future.  I'm sure 
> there are plenty of experts here trailblazing that particular bleeding 
> edge, and I look forward to benefiting from their work in the future.
> 
> In summary - LiveUSB == primarily trial and installation medium.  I.e. 
> perhaps the thing that _generates_ 
> 'installed-normal-nonlive-fedora-on-a-stick' on sticks whose flash is 
> performancewise on par or better than a usb rotating disk.
> 
> $0.02...
> 
> -dmc
> 
> 
> 
>    Does anybody other then SoaS
>> use (or hope to use) Live environments for regular operations for
>> thousands if not millions of users?  Given the above, I conclude that
>> SoaS really needs to be something that SugarLabs supports.  That
>> doesn't mean that Sugar should be tied to SoaS, just that it really is
>> a fundamental
>> part of changing education.
>>
>> Bill Bogstad
>> _______________________________________________
>> Sugar-devel mailing list
>> Sugar-devel at lists.sugarlabs.org
>> http://lists.sugarlabs.org/listinfo/sugar-devel
> 
> _______________________________________________
> Sugar-devel mailing list
> Sugar-devel at lists.sugarlabs.org
> http://lists.sugarlabs.org/listinfo/sugar-devel