D3D Thesis

From Open Source Ecology
Jump to: navigation, search

See also Tobben Log


This is the thesis sketching and working document. References, quotes, data are placed at the top of this page, and baked into thesis sketch at bottom of page as time allows.

PDF uploads

Here are unfinished thesis pdf's uploaded whenever Tobben needs to share progress File:Thesis.pdf

Team Charter


Tobben's Available Work Hours, D3D Thesis Project

  • Ca 50 work hours: Feasibility study. Is the task suitable? Write project description. Present tentative project plan for the university examiner.
  • Assure that contact between supervisor (Marcin) and examiner is established.
  • Ca 650 work hours: Engineering work and report writing.
  • Ca 50 work hours: Form opposition.
  • Other administrative work such as self-evaluation report and handling communication: ca 50 work hours.
  • Total 800 work hours

Earlier Work

Printer Designs

Troublemaker 1:

  • Proprietary formats for some of the files, AutoCAD (dwg) - [1]
  • NC license - license incompatibility issues. Need to resolve if we use any of these files.

Troublemaker 2:

  • Blog Post 1 - [2]. Design spec is to be full open source hardware.
  • Blog Post 2 - [3]
  • License not stated at repo- [4]


Lulzbot has an open source bed leveler. http://lulzbot.com



Do-it-yourself reusable construction set for hands-on learning, prototyping cartesian robots. (CC-BY-SA)

Workshops and Printer Designs Developed for Workshops

Prusa Mendel

It was the first printer to be realistically build-able during a weekend workshop[5] (this was 2009-2010). It's printed parts took 10 h to print, compared to the 20 h of its main contester at the time of release, the Sells Mendel[6]. Many Reprappers wanted to print printers for their friends at this time, so it received a lot of initial interest for its short print time[7]. It established a reputation as easily customizable.

Josef Prusa was active in the community[8], listened to feedback and the git repo was updated almost every week. Mean days between commits on master branch: 5.5[9]. He would travel the world giving workshops, and managed to pay airplane tickets by pre-selling the printed parts to workshop participants.

Estimated number of Prusa or remixed Prusa printers by January 15 2014 worldwide, based on retailers' summed estimates[10]: ~70000 to 80000 (300 of these printed by Prusa himself)

He currently sells printers assembled and kit printers[11] through an OSHW enterprise. Most customers by January 2014 was companies[12].

MOST lab workshops

The MOST lab has a concept that's very similar to D3D Fusion. They have

  • Their own kit and kit supplier
  • Detailed online documentation describing both how to build the printer and how to host a workshop
  • Gathered data from workshops
  • Free software source code and design files
  • A separate Train-the-trainer program

Distributive Ideas and Documentation

Seed Factory Dani Eder had a similar initiative in 2013, called Seed Factory. He wrote a Wikibook where he introduces several useful concepts and shares lots of engineering knowledge.

Funding Mechanisms for Open Source Work

The following mechanisms are places online with companies behind them that facilitates/organizes standard bank- or Bitcoin transactions. Some of them give users Internet tools and space for presenting themselves and/or their projects. In general, creating them doesn't seem too hard, and the competition seems rough since there exist more than 40[13] crowdfunding pages in the Internet. Only a very small selection is presented below.

The presented mechanisms don't teach how to do work, they only give technical help with how to ask for money.

Pre-selling/Crowdfunding project-wise

Kickstarter lets any (approved) project start a public fundraising campaign online (crowdfunding). All-or-nothing funding model. USA-centric. Manual approval of campaigns based on "creativeness" criteria. Products need not be open source.

Indiegogo offers less strict approval criteria than Kickstarter. No all-or-nothing model (sometime makes contributions be more like donations that pre-sales).

Bountysource Fundraisers is similar to Indiegogo but focuses solely on developers of free software and OSHW. Bountysource actively helps in connecting with potential sponsors.

Monthly Donations Person-wise

Flattr offers a convenient way to donate small amounts on a monthly basis. After registration and filling the Flatter account, setting up monthly donations to another user takes only one click.

Donations Feature-wise

Bountysource Founties lets vackers initiate and/or (crowd)fund bounties. Developers hunt these bounties by solving the specified issues or feature requests. Backers judge if solutions are good enough.

Freedomsponsors is very similar to Bountysource bounties.

Monthly Donations Team/Project-wise

Bountysource Salt Turns donations into monthly salaries for developers of free software and OSHW. To recieve the money, developers promise to spend some number of hours working on the specific project. Snowdrift is exploring the idea of network effects (organizes donations so that more backers results in more backing per backer). For free open and libre projects only.

Self-replication or Self-reproduction

"Replication seeks to copy an entire system without error, while reproduction
 includes a developmental process that allows for variations."[1]


  1. Making actual potential workshop instructors in a short time. I (Tobben) have never seen a newbie that became a good instructor in < 1 year. Confirmed by MOST lab having separate (and very detailed) train-the-trainer program.

Counting RepRaps

The RepRap Project did their own RepRap count in 2010, which they described in their report RepRap - the replicating rapid prototyper:

Owing to the free distribution of the machine it is difficult to make a worldwide estimate of the number of RepRaps and RepStraps there are, but the sale of electronic kits for the machine (which are also produced commercially) sets a lower limit of 3000 machines. However, some people construct their own electronic kits rather than buying from market. About 4500 machines would seem to be a conservative estimate of the total population at the time of writing this paper (i.e., in 2010).

Josef Prusa used similar methods to estimate that there were ~70k - 80k Prusa machines (including derivatives) in 2014.[14]

3D Hub's numbers

Source of these numbers is 3dhubs Trends page retrieved Jan 5 2016.

The reported percentage of Prusa designs is 10.

The reported percentage of RepRap-manufactured 3D printers is 15.5. That's about twice as high as what other sources report(should ref here). This could be explained by RepRap owners being more proud or their printers (they've built them themselves), or by the sharing-minded culture within the RepRap community.

We should expect the Prusa count to be as skewed/biased as the RepRap count, which leads us to assume a Prusa percentage of ca 5.

Workshop Reproducibility

In this thesis, the term "self-reproducing" will be used to describe a property associated to a set of workshop plans. When a workshop student (former participant of physical workshop or not) organizes a workshop event based on copied (and maybe modified) workshop plans, then the plans have self-reproduced.

We will call it "direct self-reproduction" when a former participant of a physical workshop organizes a workshop based on the set of workshop plans. If the organizer of the workshop using the plans have never attended such a workshop herself, then the WS plans have spread through "indirect self-reproduction".

The WS plans will contain tools and incentives for potential WS organizers. The better the tools and the stronger the incentives, the more WSs will be based on the WS plans. Our metrics should therefore capture the quality of the tools and the strength of the incentives.

Quality of the tools may be measured along for example the axes:

  1. Do they enable fast completion of task? How fast, compared to what?
  2. Do they require additional tools to work, or are they complete/independent?
  3. How much maintenance does it need to keep working over time?

The strength of incentives can be measured along:

  1. Net $-income per invested work hour?
  2. Predictability of $-income?
  3. Possibility of making friends?
  4. Fair cred and respect within community?
  5. Possibility for gaining valuable knowledge?
  6. Possibility for gaining provable knowledge (certificates or similar)?
  7. Sense of ethics and meaning?
  8. Sense of fun?


Line of Reasoning

A line of reasoning I might be able to build my Master's on. After each number comes a sentence that builds the line of reasoning (the rest is just fill-in which will be improved later).

0. Intro: Thomas Piketty has presented firm evidence that profit from ownership is rising faster than profit from labour, resulting in greater economical inequality in Europe and North America. His proposed solution of heavier taxation of the rich has influenced the old and polarized discussion about taxation only slightly. Big companies influence lawmaking and they have the skills to take advantage of any hole in tax legislation worldwide. The new global market has made them fewer and more powerful. A global authority and tax legislation would be needed to control them, but something like that is not on the horizon yet.

1. Economical inequality is bad, but it is not the unequal distribution of money that leads to fundamental social problems, it is unequal distribution of power, so rather than only redistributing money, we could fight inequality by also redistributing ownership of the means of production. This could mean distributing control and ownership of the stock market or it could mean making the production machines themselves distributable, and give smaller groups controll over smaller production machines.

This thesis develops a socially and economically sustainable business model that distributes itself and generally useful machines as part of its core strategy.

2. The GNU project and the Free Software Foundation (FSF) have developed computer programs and licencing schemes that lets anyone take control over their computing machines, making them potential production machines for their owners. Software, and information in general, is a very scalable approach to distributing power since it is so easily copied and transmitted. Their success in empowering the owners of computers running their software is shown clairly by the academical and programmer communities' strong preference for operating systems based on the GNU one, and the wide usage of FSF's licences.

3. However, GNU's and FSF's organizations have relied on donations and have been criticised for not being economically sustainable and not developing economically sustainable business models for the users they seek to empower. The free software community argues that proprietary software is unethical, but has so far avoided the responsibility of being everyones business advisor in order to make them behave ethically.

4. The RepRap Project, initiated by Adrian Bowyer has designed and distributed generally useful production machines by giving the hardware itself the possibility of reproducing a large fraction of it's own parts. The RepRap machines in addition can produce a wide rande of useful goods, and Adrian describes the RepRap/Human relation as a symbiosis comparable to that of Flowers/Bees. A fundamental difference between the Human/RepRap symbiosis and the Flowers/Bees would be that humans are not born to RepRap like bees are born to collect nectar. Humans needs to get motivaton and skills added after we're born to initiate the symbiotic relation.

RepRap workshops are events where RepRappers share and develop their skills. At a ceirtain type of workshops, often called build workshops or assembly workshops, skill transmission happens while participants builds their own machines from parts produced or sourced by other RepRappers.

5. We've observed RepRap assembly workshops for some time, the most successful ones being those of Josef Prusa and friends resulting in > 200.000 Prusa machines built to this date. The first Prusa RepRap design was a simplification of previous designs, shortening its assembly workshops to fit in a weekend. This was an important threshold making assembly workshop attendance possible for people with full time jobs.

6. The Prusa machines have spread, but the number of RepRap assembly workshops has declined. The common way to explain this is purely economical. Assembled RepRaps have gotten so cheap that people can afford skipping the assembly process. This explanation ignores the skills viewpoint and the possibility of social motivation. It also ignores the important time consumption threshold of one weekend.

7. Another way to explain declining assembly workshops is that RepRap machines have self-replicating properties but the workshops themselves does not. A workshop participant will typically not go into hosting workshops him/herself because

  • High quality teaching material source files are not available
  • RepRap teaching requires a different of skills than RepRap assembly
  • There's no social community dedicated to sharing reprap worskhop knowledge. The RepRap forums typically focus on machine design.
  • Participants' willingness to invest time and money into traditional assembly workshops is declining.
  • Arranging workshops takes more than one weekend, so the host can not have a full time job. Hosts must make the workshops their job or arrange the workshop over weekends.

We can quantify the length of a self-replication cycle for a machine or an event by recording time needed for complete self-replication and the number and complexity of human actions needed for self-replication. Both minimum and average times are interesting to record. With recording number and complexity of actions we mean counting clicks, counting screws, unique part count... This will not be absolute metrics, but still useful for comparison.

We can quantify possibility of self-replication by assessing the requirements on the environment, counting the external inputs (vitamins).

We can quantify motivation of self-replication by and by comparing motivational factors, such as honey, money, 3D printed parts or social creds, with required investments.

8. We can combine the above three quantities into an analysis of machine or event viral replicability. A short self-replication cycle along with widely available external inputs and strong motivational factors makes more viral replicable machines and events.

9. To shorten self-replication cycle and lower the number of external inputs to the workshop, we should provide participants with Live USB sticks containing a libre (free as in Free Software Foundation) operating system along with all the information, programs and instructions needed to complete a workshop, host a workshop and contribute to the RepRap Workshop community. This Live USB operating systems will itself be of high viral replicability since it can copy itself onto a USB stick with < 20 clicks of a mouse, and within 10 minutes of time.





While production machinery have gotten more efficient, work hours have increased and wage growth has slowed down for the average Western worker during the past 30 years. Ownership and control over means of production gravitates towards fewer people who don't consume their fortunes fast enough to maintain economic growth of the nations.(Piketty)

While increasingly centralized production has been the norm since the Industrial revolution, one notable university project started in 2007 has now achieved the opposite.(Bowyer et al) The RepRap project created a machine capable of reproducing its own parts, and released all technical details on line with user instructions.(Sells) New versions of the design started appearing in all corners of the world. Research in all related areas accumulated in the online community wiki and in its forums, which established a meritocracy similar to that of FLOSS projects. In addition to being able to self-reproduce, the RepRap machine, a FFF 3D printer, was capable of producing valuable goods.(Pierce) It was shown that a 3D printer would be a worthwhile economical investment for the average household.(Wittbrodt, et.al) It was also shown that the open source model was a viable model for development of physical objects.(de Bruijn) Those who had joined the RepRap community early formed successful enterprises around their new won technical skills. By ca 2013, it was widely believed that means and knowledge of 3D printer would have to centralize again in order to penetrate the household market. (refer to Stratasys/Makerbot aquirement)

But looking at 3D printer design statistics in 2016, it is clear that 7.8 - 20 % of the worlds 3D printer population are still RepRaps. (3Dhubs, Smarttech Publishing report, Ultimachine sales numbers) More so, the majority (better with exact percentage here) of RepRaps are based on designs named after one single RepRap community member, Josef Prusa. There are more Prusa-designed 3D printers in the world than the combined number of printers produced centrally by the two largest brands. (Build under with data!) He learned FFF 3D printing entirely within the community and started the effort that became the first design to be realistically buildable during a weekend workshop. Members of the community called it Prusa Mendel or just "the Prusa printer". Josef Prusa himself travelled across Europe, giving workshops and meeting RepRap community members in person. This formed the printer designs Prusa i2 and Prusa i3 to be workshop-friendly with several known modifications and a large base of workshop-educated users that are still active today. Many former workshop participants base enterprises on remixed designs, but few base their enterprise on workshops.(ref)

This makes the RepRap Prusa the first open source hardware (OSHW) designs to channel (some significant share) of a billion dollar market(ref) towards the spread out, small scale makers, scientists, hackers and dealers costituting of the OSHW community all over the world. Many papers have been written on how RepRap distributes production power and economical opportunity. Build something from them here and list what parts of the RepRap project they mention as distributing production power and economical opportunity.

  • Self-reproduction?
  • Internet?
  • Expired patents?
  • Open source licences?
  • Hacker culture?
  • Others?

The Prusa workshops had the limitation that they continued only as long as Prusa travelled. Others have tried to make workshops based on Prusa designs economically sustainable without the travelling, and failed.(refer to Berlin fablab, Barcelona fablab)

Societal Impact of Digital Technology

"Digital technologies change
 rapidly, but organizations and
 skills aren’t keeping pace.
 As a result, millions of people
 are being left behind. Their
 incomes and jobs are being
 destroyed, leaving them
 worse off … than before the
 digital revolution..."[2]
"Conventional industry has
little use for this idea: why sell a machine to your customers that means that they
never need to come back to you to buy another, never need to buy spares, or even that
allows them to go into production themselves in direct competition with you? But
owning such a machine would have real advantages for people in general: anyone
who had one could use it to make things, and could also make another such machine
and give that to a friend. This is an interesting example of a failure of the market: such
a self-replicating machine is an object that people would value, but that it is in no
one's interest to sell. For these reasons it was decided to make the machine and to
give all its designs away free under the GNU General Public Licence on the web.
This was the start of the RepRap project."[3]



Software tools





  1. ADAMS, Bryant; LIPSON, Hod. A universal framework for analysis of self-replication phenomena. Entropy, 2009, 11.2: 295-325.
  2. BRYNJOLFSSON, Erik; MCAFEE, Andrew. Race against the machine: How the digital revolution is accelerating innovation, driving productivity, and irreversibly transforming employment and the economy. Brynjolfsson and McAfee, 2012.
  3. BRADSHAW, Simon; BOWYER, Adrian; HAUFE, Patrick. The Intellectual Property Implications of Low Cost 3D Printing. ScriptEd, 2010, 7(1), pp. 5-31