Discover the #10YearPhone!

Right to Repair Europe is very excited to share with you our latest campaign: the 10 Year Smartphone, a phone designed to be repaired, with 10 years of software updates, 10 years of spare parts availability, and much more!

Sounds too good to be true, right?! Maybe, but all the more reason to check it out! :wink:

We all know the problem. Over 200 million smartphones are sold every year in the EU. Thatā€™s nearly 7 every second. Each time one of these phones is made, it creates between 40 and 80kgs of CO2 (the same as a 3 hour drive).

And where are the old ones going? The answer is: into the bin. E-waste is the fastest growing waste in the world. Over 50 million tons were discarded in 2019, and only 17% were properly recycled.

How can you help?

The #10YearPhone is an answer to that. If all phones in Europe lasted 10 years, we would save 6 million tons CO2 emissions a year.

Donā€™t worry, we donā€™t think that buying a new phone is the solution. We want Europe to be ambitious about long-lasting, repairable phones! We are asking you to:

  1. Check out the 10 Year Smartphone
  2. Trust us enough to click through to ā€œpre-orderā€, and sign our letter to the European Commission, asking for all phones to be 10 Year phones
  3. Share the 10 Year Smartphone as much as possible on social media. If youā€™re looking for inspiration, here are a Tweet, Facebook and Instagram posts

Andā€¦ in case you are already sold to the idea, you can cut-through to the ā€œpre-orderā€ page. Are you ready to say good bye to throw-away electronics?

Find out more

3 Likes

Related to this speculative campaign rallying cry, hereā€™s an interesting question for more technical or business-oriented people here.

We have a Twitter person who says they are a developer (albeit extreme libertarian without a profile pic) claiming that 10 years of software support will cost upwards of ā‚¬100 per device. (He also has an interesting idea about tax :face_with_hand_over_mouth:).

Fraunhofer estimated a full five years of support for all devices would cost ā‚¬2 - we cite the study in our tweet in the thread.

@Dave @panda @benski and other technical people what do you make of this assertion of ā€œexponentialā€ cost for 10 years of software support? Especially @benski as he works in standards and support for older TVs.

@Jana_Rueckschloss any chance that Fraunhofer researchers could be asked about this?

I remember from our podcast with Ross Anderson that this is an area he was interested in researching. In his words

What you got to do is you got to keep shipping patches for Android, not for 3 years, not for 5 years but for 10 years. Then weā€™ll believe, Mr. Google, that you actually do care about sustainability.

2 Likes

Itā€™s complicated!

It depends a lot on how well the OS has been written in the first place. Itā€™s much easier to maintain a systme based on a micro-kernel with few dependencies. Most OS companies strive to work in that direction, but then new phones often force shortcuts to be delivered on some date. It also depends on (internal) politics. How is the OS team working with the product team? Are they even the same company. When different companies, the phone manufacturer may have its own OS development/modifications making the software stack more complex.

However, it also depends what kind of support weā€™re talking about: feature parity or solely security updates? One model is feature updates for a few years and then security updates for a few more.

You have to consider the quality assurance/testing needed. This is much easier for a company doing everything. However if the OS is integrated by a phone manufacturer and then further modified by a net ops, all this with variants for each countries. It is likely that no one will know exactly what are all the variants and so whether say a security fix may interact with a software change in one of the many variants so the testing may involve many companies and may need to be done in a waterfall model.

And if you donā€™t spend enough on software engineering, it is likely that engineers will work on features for new phones as a higher priority as that where most of the income comes from.

To reduce all this complexity (and I havenā€™t gone into much detail) and ensure new features are added for longer and security updates provided for much longer I believe that the business model has to change. There must be more commercial incentive to support older models. Factoring in the cost to the earth of the different models could be a way forward.

4 Likes

Yes to everything Panda said.

As for Mr. Evers: heā€™s talking gibberish.
Heā€™s taken the ā‚¬2 and the extra 5 years and said itā€™s ā€œmore than likely exponentialā€ (without any evidence) and 2^5 = 64 thereforeā€¦
Who knows where he got the 56.25% tax rate from.

Much of the cost of a phoneā€™s software is not in development, itā€™s in setting up the development tools, homologation, testing (mostly automated), etc. all of which are (mostly) fixed costs amortised over the life of the phone:
10 software updates donā€™t cost 10 times as much as 1 software update.
Thus, yet again, long life = cheaper product.

Another point is that if the software is modular and designed with re-use in mind,
newer models can share code with older models where possible:
much of the code, such as networking, user management etc. should be common (notice the should).

Also, if the software is well-designed and modular, the issue of supporting older hardware with newer software doesnā€™t arise: the software just wonā€™t support the features that the older hardware canā€™t support.

OTOH if the software isnā€™t well designed and modular, is this something that the consumer should pay for?

I may take another break from moving into my new flat to answer him directly if I can be bothered but heā€™s only got 5 followers so Iā€™m not too bothered :smiley:

4 Likes

Thanks @dave and @panda really helpful discussion. Fairphone is now ā€œaimingā€ for 6 years of software support, wondering if they would share some of their calculations around cost. And would be cool to see if we could tempt Fraunhofer researchers to make these estimates for a 10 year smartphone, but I think they need to be paid Ā£Ā£ for this sort of thingā€¦

I was looking at https://shop.fairphone.com/en/?ref=header
they launched the Fair Cobalt Alliance and have an iFixit score: 10 out of 10

2 Likes

Yes, they announced this since we launched our campaign! 5 years warranty and 6 years software support. Just about halfway to the #10YearPhone :slightly_smiling_face: Noting Fairphone are supporting this campaign too!

I asked around which colleagues put the statement in the report.

ā€œAssumption on additional costs per device is based on approximately 1000 different smartphone models being on the EU market, with on average 150.000 sold units, and updates being in the cost range of ā€œseveral hundred thousand US dollars per modelā€ (Clark 2016), i.e. calculating with 2 Euros per device for this option. For comparison: Stated software development costs for the Fairphone 2 are 4,62 ā‚¬ at 140.000 sold phones per year (Fairphone 2015)ā€
Source: https://www.fairphone.com/wp-content/uploads/2015/09/Fairphone2-Cost-Breakdown.pdf

The paragraph was written by Karsten Schischke and he has no more background information. He explained: The software customization costs are almost independent of the number of devices that are updated. Then it becomes significantly more expensive per device for very small numbers of units. For very large numbers of units, it is likely to be significantly less than 2ā‚¬ per device. In any case, the figure is only a rough estimate. The further analysis shows, it almost doesnā€™t matter whether it costs 1ā‚¬ or 10ā‚¬ extra, for the consumer it is worthwhile in each case opposite a new purchase. For the study, they had to make an assumption, but the 2ā‚¬ is not a sound absolute number.

If hardware no longer supports new OS versions, then it may be necessary to extensively change the source code so that all hardware limitations are somehow bypassed. That is certainly expensive, but my colleague could not give a number.

3 Likes

I responded to that twitterer