[Sugar-devel] Maintaining the sugar-live-build

[James Cameron]

I'm used to it.  OLPC faced this same type of deprecation threshold
back in 2014 and 2015 with loss of AMD Geode instructions used in the
XO-1 laptop.  It wasn't only the kernel, it was also the WebKit
JavaScript JIT compiler.  There was nothing in compiler run-time that
did check if the instructions were available, and so the Browse
activity did stop with a segmentation fault or invalid instruction.

Sugar Live Build was made to (a) demonstrate the Live Build
technology, (b) support students that Google was bringing us, and (c)
provide a minimum viable environment for development in a VM accessed
over SSH.  The degree of difficulty in using it was a function of the
feedback received; and we didn't receive useful feedback from people
who could not write an ISO image to a USB drive.

Yes, Sugar is becoming almost unusable by the majority's standards,
because those standards are also changing over time, and Sugar is not
changing with them.

Sugarizer provides a system that meets the majority's standards; just
install the app.

The majority is always right.  That's what it means to be in the
majority.  Like Martin, we should look for opportunities to serve the
minorities.  I'm glad Martin is taking on this duty for 32-bit Intel
systems.

On Mon, Oct 26, 2020 at 04:06:12PM +0100, Martin Guy wrote:
> On 26/10/2020, Chihurumnaya Ibiam <ibiamchihurumnaya at gmail.com> wrote:
> >> Fedora Soas boots into a lightdm login screen without ever telling you
> >> that the required username is "liveuser".
> > Yes there's a discussion for the bug here [...]
> 
> Thanks for that
> 
> > I don't know how to build SOAS images yet but you
> > can find F29 SOAS here
> > <https://archive.fedoraproject.org/pub/archive/fedora-secondary/releases/29/Spins/i386/iso/>,
> > which supports 32-bit as you said Fedora no longer supports 32-bit.
> 
> Yes, mandating 64 bit systems is a disappointing trend. Ubuntu's
> last32-bit LTS was in 2016, Redhat in
> Me, I think the CPu manufacturers are bribing the distro maintainers
> to make old hardware unusable to encourage people to buy new, but then
> I'm old and cynical :) It's more likely that frontline developers are
> surrounded by monster machines and don't believe anyone has crappy old
> rubbish any more.
>   I had similar battles with the Debian ARM crowd, who were anxious to
> srop support of armv4 in favour of armv5, just to be able to use the
> Count Leading Zeros instruction, but thereby wiping out the cheapest
> and most needy target machines. Similarly, the first Debian ARM
> hard-float port targetted VFPv7, and as a result the Raspberry Pi
> project had to no option but to make an entire new Debian repository
> and recompile everything for VFPv6, a huge waste of time caused by one
> silly decision... and this from developers of the "Universal Operating
> System"!
>   In the case of Sugar, enabling the widest deployment on the cheapest
> machines with the least user effort seems to achieve its goals more
> than giving a few percent of extra speed to power users.
>   However, providing both 32- and 64- bit images is an acceptable compromise.
> 
> > I think that if you want to make 32-bit images it'll be for your own use
> 
> See above. When *you* live in a poor area and people keep bringing you
> ancient laptops that they've found in the rubbish, maybe you'll think
> differently.
> 
> I'm a seasoned developer and building my own images is fairly
> straightforward, but most people don't know how to write an ISO image
> to a USB stick; maybe to a CD if they are techies.
> 
> > I agree with James re [not, I presume] including more activities
> 
> Yes, SLB is written by and for developers and for advanced users, it
> seems to me; instead of being a distro ni itself but a meta-distro: a
> kit for creating custom distributions. In this light, keeping it
> minimal does make sense.
> 
> What I was looking for was the easiest way to put a well-furnished
> Sugar on an old laptop with the least effort, to be able to evaluate
> it as a solution to offer to normal people, to make the luminary work
> of Papert, Negroponte et al more of a social reality instead of having
> every UI be, essentially, a worse version of Xerox PARC in the 70's.
> 
> What I found is that nothing works. I'd expected something of the
> quality of the Morphix Combined Gamer live CD (2004!); instead I found
> only one that had promise, and that needed more effort from me than I
> had expected. Imagine the experience of a normal user trying to
> achieve my goal: disappointment and abandon of Sugar as unusable,
> which goes against the goal of its diffusion.
> 
> However, making available well.furnished images worldwide with the
> publicity for them that sugarlabs has in this field accomplished my
> goal, not in my area but worldwide, and that's worth the relatively
> small amount of work that seems necessary to achieve that.
> 
> As well as the honour, of course!
> 
> Does that make any sense?
> 
>    M
> _______________________________________________
> Sugar-devel mailing list
> Sugar-devel at lists.sugarlabs.org
> http://lists.sugarlabs.org/listinfo/sugar-devel

-- 
James Cameron
http://quozl.netrek.org/