EOMA68 crowdfunding campaign (the last few days) plus ways forward for Libre Computing

[Xavi Drudis Ferran]

El Mon, Aug 22, 2016 at 07:10:13PM +0200, Paul Boddie deia:
> Once again, I may have pledged for rewards in the campaign, but I'm not 
> explicitly endorsing it. Everyone should weigh up the different factors and 
> their own needs when supporting such efforts. However, it is worth noting that 
> there are only 4 days left until the campaign ends:
>

ACK, the RYF or money back guarantee helps (I don't usually see the
same hardware being offered with different software where only one
option is RYF _and_ the RYF option getting more than double the demand
than the next most popular option, that already says something of the
community-to-be). But it's already difficult to tell somebody else
what to spend their money on when you have used the device for a few
months, in any crowdfunding it is wiser to just spread the word and
let people make their own minds about it.
 
> There has been a range of responses from enthusiasm about the modular 
> computing concept and support for hardware that fully supports Free Software 
> through to criticism about the performance and the ethical record of the 
> system-on-a-chip vendor that happens to be used for the first card. Reading 
> some of the responses, I note that maybe there is some kind of law of public 
> discourse that the more severe the criticism, the less constructive it tends 
> to be.
>

In some case it's been more constructive than others. Funny thing is
that when an evil corporation launches another evil product, you can
hardly hear some boohs somewhere, not because everybody likes it, but
because everybody expected something of the kind. But when someone
says to be trying to sell something good in an ethical way, everybody
instantly disbelieves it so they're going to point out any possible
objection to prove it is not perfect. That's useful to reinforce their
inductive reasoning and saves them the trouble of having to apply
deduction to every news piece they come across. In other words, they
know these things don't happen, so they simply find confirmation of
their knowledge and share it.

And to be honest, when something looks too good to be true it usually
isn't true. The lack of features/performance and reputation of the SOC
vendor, among other things, is what makes it not too good to be true,
so somewhat likely to eventually be true. 
 
> * Whether EOMA68 (and its siblings) or something similar (see next point) has 
> a role to play in this, perhaps only as a step towards the eventual goal of a 
> set of platforms we can rely on
>

I don't know if the EOMA-68 spec has much future or too little. But I
think it has the potential to start (with others) a certain market
niche for freedom respecting hardware. Manufacturing is a high volume
thing, so it is very difficult to compete. They're aiming at a product
that does not exist, so it does not have competitors already (the
general concept is similar to phoneblocks, or even many dual
Computer-on-module and peripheral mainboard, but the standarisation
idea is not really available anywhere else). Incumbents could enter
the market if they find a business there, but Luke is trying to define
its market in a way that is less profitable long term for incumbents
than their already existing markets, so is somewhat less likely.

Accordingly, the market should also be little profitable for Luke. 
But then it's a curious moment to start it: 

- There's a certain saturation in the conventional market (first 
PC sales went down, then laptops, then tablets... ) 

- There's some slow down in moore law. 

- There have been difficulty in finding new needs for the ever
higher computing power of new devices, so "good enough" computing
has a better chance than historically.

- There's unprecedented hindrances and antifeatures in most consumer
available hardware (up to now you could more or less find some
hardware to run with free software, now most software is designed to
try to stop you from loading any free software or at least to stop you
running 100% free software). Up to now we could use mainstream
hardware, pay the same than anyone else for our computers, but use
them in more or less freedom.  But now those times are ending. Soon
you'll have to go for libre hardware, free software, decentralized or
P2P services and open content or go for closed hardware, proprietary
software, centralized services and DRMed content. The first market
may be small or non-existing depending on what we can do, but I find
very difficult to prevent the second. I imagine unlikely the prices
to be the same for both. 

For me it is difficult to foresee whether EOMA-68 will fly, if it will
last lots of time or be replaced by something else (hopefully as good 
or better), an evolution of the concept like EOMA-200 or whatever or 
something completely different. But I think it is already helping by
discovering some demand for something like it. 1414 pledges.

> * Whether there are other ongoing efforts in this area (not one-offs) that 
> just need more attention than they have been getting
>

A part from what you list there are efforts around minifree, libreboot and 
the like. That is, people can get fed up of buying new faster hardware 
with more antifeatures, and start recycling older hardware from the times 
when there were less and weaker antifeatures. This is intuitively a lot 
of effort and eventually will run out of usable hardware, but has several
advantadges: 

- you reap the benefits of mass production (with some discount since the 
hardware is old). 

- you boycott current offerings. Instead of debating whether to buy a device
with chips from an evil vendor or from another vendor that is evil in some
other way, you buy second hand from another consumer and no new sales for
the companies causing the problems. 

- you get down of the treadmill instead of running ever faster. In as
much as we struggle for computers as fast as mainstream we will be
allowing society to keep pushing ever weightier content at us. At some
point you simply have to tell your friends to not send you 12MB fotos
of their holidays, tell your TV station that you don't really need 4K and
will opt for lower res, tell your supermarket that you don't really need 
cloud connected mass-spectrometers in microchips inside your toilet
paper, etc.

- off-topic benefits like waste reduction, less resource depletion, etc. 
 
> * Whether or when libre CPUs and SoCs might be viable choices for such 
> platforms
>

Hopefully one day they will. I don't know much about chip
manufacturing but it looks like this needs the niche market bigger
than it is now (or a chip that is attractive in different
markets). I'd love to be wrong.  In any case trying it is very useful
in itself, regardless of whether it has more or less success. And
those putting money on it probably know their chances better than I
do.

> * Whether some existing products are good-enough choices for hardware 
> platforms, even if they are not libre hardware or employ encumbered 
> microarchitectures (which a lot of established products do, of course)
> 

I don't know all existing products, but my experience is that almost
none of the products you can easily buy new are useful. Interesting
things are few and far between, in crowdfunding experiments, to be
bought from the other side of the planet during certain time window,
ever uncertain of their ability to sustain a community that keeps the
thing alive... And of course if you get free software you don't get
fair trade, or local economy, anonymous shopping, or your preferred
payment options, or environmental sustainability or ...

I like the pyra-handheld.com though. It's a pity that they don't have a
RYF version, but with a wifi USB dongle you could use it in
freedom. It might eventually get 2d acceleration, but I don't know
about video acceleration. 

> There's an interesting summary of processor suitability done for the criteria 
> of EOMA68 that some might find interesting:
> 
> https://www.crowdsupply.com/eoma68/micro-desktop/updates/picking-a-processor
>
> "It should also be pretty clear that there is literally not a single processor 
> that checks every single box! As in, there is not a single processor in the 
> world that is eco-conscious, respects software freedom, is ethical and 
> accessible. This is a pretty insane situation to be in, in the year 2016."

If the campaign failed, but the effort had spread awareness of this kind 
of obscure (to the general population) oligopoly and depressing products,
it would already have been useful.

I was in fact surprised that so many people knew Allwinner as GPL violators.
I'm also finding more people aware of secure boot, signed bootloaders and
so on. So the word slowly spreads. 
 
> 
> There needs to be a constructive debate about incrementally improving this 
> situation. Instead of "I hate that processor" or "wait for my radical SoC I've 
> just started designing", people need to help find products that uphold 
> software freedom and privacy while also being usable (obtainable, for the most 
> part) for small libre hardware projects. And there needs to be an appreciation 
> that this work is not meant to create the "toy of the month" - a gadget that 
> is fun for a while and then stashed away somewhere - but instead to build an 
> environment where we shouldn't be constantly needing to urgently figure out 
> what kind of hardware we can use that uphold our values.
> 

Well some kind of wiki to document each chip and it's problems for low
volume RYF libre hardware (MOQ, GPL violations, proprietary drivers,
NDAs, etc.) would be something.

Just one quick note: for people hating Allwinner, there's the option
to get an EOMA-68 laptop with a pass-through card. This does not work
as a computer but can be plugged to something you have and use the
"laptop" as a USB keyboard and mouse and HDMI screen. So you can plug
your preferred SBC, HDMI stick, mobile phone or add a second screen to
your laptop. Then you may use one less CPU and still enjoy a laptop
form factor. And upgrade to a full cableless laptop if one day they
sell a computer card with a CPU you don't boycott.

The campaign is now accelerating at 75% of funds and 95% of days, it
must have been foresseen for 250 cards and 250 laptops. But the
creator has stated that it would be viable with other quantities.
Since there are already 487 (+ 40?) cards and some of the critical
laptop components are also required for other products (like
micro-desktop or breakout board) it seems the savings in one side
could compensate the loses in the other if they get at least 100
orders for laptops (they're now at 97). So requesting a laptop kit may
be more useful to the campaign (but more expensive for the backer)
than requesting something else. In theory they shouldn't get any money
if they don't reach 100% of pledges (36775$ left in 3 days,
difficult), but in practice if the requested items can be produced
with the pledged money, something should be possible to arrange.

I don't have an issue with people hating a processor. I think most new
processors have earned hate, and hate is fine as long as you redirect it
to good endeavours. I can understand boycotting Allwinner. And if
then they boycott all the rest much the better. If they boycott
Allwinner and buy Intel then I fail to understand.

I mean they stopped selling computers and they only sell traps. Fine,
we stop buying computers. That'd be sane. Not always easy, but sane.
At least saner that this addiction to ever more performance and 
features, at whatever privacy and control prices. 

> So, does anyone have any opinion about the kinds of projects (most likely 
> being undertaken already) that need our attention or support? 

I don't know. RISC-V ? 

> How do you envisage a sustainable computing platform? 

I guess I'd need to envisage multiple sustainable fabs. That's the
hard part. I think Open Source Ecology has not got there yet. 

> And since all discussions inevitably lament how much memory Firefox
> uses these days, how do you envisage a less demanding form of
> computing being extended to online services?

Err... an army of anti-sect therapists ?
 
> Sorry for the long message!
>

long ?