Fairphone lessons (Re: Shiftphones details)

[Paul Boddie]

On Tuesday 14. May 2019 08.52.29 Bernhard E. Reiter wrote:
> Am Freitag 10 Mai 2019 17:06:40 schrieb Paul Boddie:
> > it surprises me that they have not managed to
> > attract broader publicity.
> 
> Shiftphones seem to focus on Germany (or German speaking companies).
> It is a significant invest to create text and give support in a second
> language.

Agreed. But it still surprises me that a wider audience were not aware of it, 
not through any efforts of the company targeting other countries and audiences 
(or not doing so), but simply because word does get around.

> > My understanding (and recollection) is that Fairphone fell into the same
> > "original design manufacturer" trap that lots of people do. Now, most
> > vendors do not care about the lack of longevity of the thing that they
> > have procured: they can always sell or give an unhappy customer the next
> > thing coming out of the factory.
> 
> In my perception Fairphone aimed for an improvement in longevity of their
> product and were successful. The Fairphone 2 was produced and on sale for
> about 30 something month.

Fairphone 2 is still supported with software updates, as I understand it. This 
is obviously a good thing and it may even be a notable thing in the mobile 
industry and in consumer electronics where the manufacturers have an incentive 
to sell customers a new model rather than support existing ones.

I was looking at update and source code availability for certain phone 
manufacturers only yesterday, and despite various practical challenges, it did 
seem to be possible to get updates for older models from certain 
manufacturers. For example, bq models seem to have relatively recent source 
code and firmware updates.

However, the aim should be for indefinite support: that is, support for the 
software should continue until nobody is practically able or willing to 
produce updates. Economic factors play a part here, clearly, because if there 
is continual churn in the code, lots of work is needed to prepare, test and 
deploy updates. Here, something could very easily be said about software 
engineering (or lack of it) causing such labour-intensive processes and 
discouraging sustained support of deployed software.

But other aspects, like a lack of standardisation of the hardware (with a 
tendency for each new product to be different and special and thus merit a 
completely new software effort) and proprietary/secretive hardware that only 
the manufacturer is able to support (with no incentive to do so once newer 
products are available), undermine or defeat any independent efforts to 
support software. Despite Free Software being used, end-users are being denied 
control by selfish interests.

And let us not forget that some manufacturers simply deny end-users the right 
to exercise the privileges granted in the Free Software licensing used in 
those manufacturers' products. Such behaviour is an affront to our principles 
and what organisations like the FSFE stand for, and yet such behaviour was 
practically excused within the Linux kernel development community, especially 
in the upper levels of it, because we should supposedly be happy that Free 
Software is being widely used. Again, what good is Free Software if the end-
user never gets all the promised benefits?

> > a Free Software initiative would have encountered software sustainability
> > issues at the first hurdle, giving them the opportunity to back up and
> > choose a different approach.
> 
> When trying to get a product out of the doors, you face a large number of
> small and larger decisions. First of all, the product has to "work" for the
> expected usage. Fairphone 1 was good in this regard, but Fairphone 2 a bit
> less so. Backing up and taking more time may have not been possible, without
> risking to not have a product at all. Which would have been the worst
> result. So to me your criticism is too harsh. After all they produced two
> phones that were significant steps forward.

I think my criticism is harsh, but it demonstrates that Fairphone were not "a 
Free Software initiative" because they did not give the issue of the software 
the priority it deserved, at least for the first product. And due to the way 
software and services are being developed and delivered nowadays, software 
viability has probably become the primary limit to product longevity (perhaps 
alongside battery degradation and other "repairability" issues that Fairphone 
have also confronted, to their credit).

> If we had more manufactures trying to go in the Fairphone direction, it
> would foster much more Free Software usages on mobile devices. It is fine
> to point out how they could do better, but I think we should even more
> applaude them for the advances.

Yes, I recognise their achievements. And they have improved with regard to the 
software, meaning that I look forward to what they produce next.

> > Naturally, the whole mobile industry suffers from these issues, too: it is
> > like the Wintel upgrade treadmill turbo-upgraded for the 21st century. As
> > software practitioners, we should be looking to offer real solutions for
> > this.
> 
> I agree, thought we first must understand the real reasons behind fast
> upgrades. Some customers are very happy about a new model each year and
> they'll buy it.

I am sure they are. These are presumably the same apologists for phone 
manufacturers trying to cut warranty terms where I live: people who openly 
said that they bought a new phone every six months, that longer warranties 
would make phones more expensive, and that nobody needed them anyway 
(presumably because at six months, they would sell their phone to some hapless 
buyer or fake up some kind of insurance claim).

The problem is that we have to share a planet with idiots like this, with 
their behaviour validating the destructive and wasteful actions of 
corporations who are not being held responsible for the consequences of their 
"need" to make money. With many other technological phenomena, the functional 
reasons for upgrading diminish over time as everything becomes "good enough". 
Sadly, in the spirit of the Wintel upgrade treadmill, the need to upgrade is 
seemingly driven by wasteful software and services.

So, we just have a bunch of different people enabling each other's destructive 
behaviour while everyone else is made to think that the problem is too 
complicated to be solved. I don't think any of us with a conscience should 
allow this situation to persist, particularly as we are fully aware of things 
that can be components in an eventual solution.

> > Why shouldn't my next phone be usable, even in a modest sense, for as
> > long as my current one, which is actually fifteen years old?
> 
> One thing is technical progress, there is 5G coming and at some point you'll
> may need a phone that uses the standard. Another example there are websites
> or services that you would want to use, that only run with hardware and
> software that is newer.

Well, my current phone is only capable of 2G, and although that may go away in 
some places - probably the US with its own peculiar heritage - it will 
probably remain viable for a few more years. But technical progress is not 
really any justification. As I have noted above and elsewhere, many services 
do not need even more bandwidth and even more performance.

It is the same mutually destructive reinforcement seen in the prime of the 
Wintel relationship, where people make elaborate but flawed software to use up 
the extra performance which then drives hardware development to compensate for 
things getting slower and slower. This might be great if we were talking about 
immersive holographic environments or something exotic, but when it comes to 
putting a few things into text fields, pressing a button, and seeing a list of 
things to buy, it just drives obsolescence and waste.

> > where the people trying to make
> > such phones are outsiders and are not part of the manufacturer ecosystem,
> > with its convenient and cheap access to knowledge and technical resources,
> > and so on. And getting access to the right people to solve problems is
> > difficult given the low volumes and outsider status of such initiatives.
> 
> What I've heard from the OpenMoko project and others is that you cannot get
> the top line of SOCs from manufactures in small numbers. Something like
> you'll have to buy 10.000 at least and then put the money down up-front.
> Knownn the right people won't help with that.

True. So as I noted there are numerous obstacles. Independent initiatives are 
locked out in purchasing terms (prohibitive minimum order quantities), 
logistical terms (they do not have convenient, insider access to production 
facilities), and collaborative terms (they do not have in-house access to the 
designers and people who can make up for incomplete documentation and provide 
useful, informal knowledge).

> > I was actually surprised in my review of available phones that Fairphone 2
> > is now no longer available, although factory-refurbished ones can be
> > obtained for a discount.
> 
> This is a recent development (in the last weeks).
> Probably a good one, a Fairphone 3 is needed for a while now.
> 
> > What might have been interesting is if the modular
> > technology had been popularised, shared, standardised, and so on, so that
> > others could have made upgrades and continued the general availability of
> > the product.
> 
> You know that all this would have meant significant efforts and Fairphone is
> a small company (in a growth phase, with all the pain coming with it). At
> least they have shown that it works and there is a market for it (even when
> small). This is a large archievement.

I agree. I hope they can demonstrate collaboration with others with aligned 
interests precisely because addressing all of the different issues is really 
too challenging for a small organisation.

Paul