Open Science SDG project report by Xoeseko Nyomi
1. Introduction :
My involvement in the project started this October when Thomas came to talk to our class about open contributions with structures such as Github and Wikipedia. I personally already knew a little about open source and had without a shadow of a doubt benefitted from it both privately and professionally. I had however never actually contributed personally to an open source project. So I warmly welcomed the idea when Thomas suggested Sofia and I could get involved with the sdg-market he was developing (we were later later on joined by Bayar who provided a great help with design of the website and critical reflection over the financial mechanisms). The aim of the sdg-market is essentially to provide an open and decentralized platform, a market, where projects which work towards sustainable development goals can be “remunerated” through tokens and thus galvanized, be it through the gamification where developers receive tokens for contributions or through financing where investors can purchase tokens of projects they deem interesting. A prototype of the project was already running. Our tasks were thus multiple, we first had to read the whitepaper to understand the goal. We then tested the platform to see if any bug slipped in during development. We could then get involved in the open collaboration process. As testers we were essentially reviewing the project and as contributors we could work directly at issues we found relevant to address.
2. Problem Statement
The main issue addressed by the SDG market is incentivizing involvement in sustainable development goals. When we met with Thomas, we first set 6 objectives we could help implement that could further the project, namely :
- Extensively testing the interface and API of the existing prototype
- Building portfolio visualizations and time series of token generations using Jupyter or d3.js
- Statistical analysis
- General critical reflection
- Reflection on blockchain implementation
- Github documentation page
In my ability, I identified 4 key areas in which I was able to provide help: testing, building visualizations and time series using python and Jupyter, providing critical reflexion, github pages documentation. I was very interested in the blockchain implementation and contributing directly to the project code base but giving our limited time and knowledge it would prove difficult.
In order to address the issues we started by testing the existing solution and understanding the whitepaper so that we could identify what was left to be done. We figured documentation and testing more than coding were where we could provide most value so we set out to develop documentation material using open source tools. The visualizations could also be integrated naturally into the documentation more than the actual codebase. Open science aided us in the sense that we used only open resources in order to develop the documentation website. We decided to use Jekyll as web framework since it was open and extremely easy to understand using simple markdown to write webpages. We also used Git and Github to collaborate. Finally we had our own code and documentation on the market to study the mechanisms from a developer perspective.
4. Research results
We first started by registering on the prototype to see what was possible. It was pretty straight forward since the prototype is essentially a proof of concept relying on Github and that I already had a Github account. That part went down smoothly we then went on to try various transactions and remarked a few UI issues. We notified the developers through their Github repository since the code is openly available. We weren’t able to directly address those for lack of experience in angular web application development but did create issues to notify them. We discovered further issues by trying to use the API to access the data that was supposed to be open. We created an issue for that too and the response was quite prompt within a day the developers commented on the issue and tried to address it although it is still not resolved to this day.
5. Solution Proposal
As contributors, we looked at everything we could do… After applying a checks and balances review process to the prototype we identified areas where we could add something. Those areas were:
- Documenting the platform to make it easier to access
- Building data visualization tools so that users of the platform could easily view the evolution of tokens they owned or projects they were interested in. Similar to how financial platforms do.
- Implementing a blockchain mechanism for storing transactions in a decentralized manner and governing project auditing.
For the documentation part we chose pretty early on to take advantage of open resources available and deploy a Github pages documentation site which hopefully can be integrated to the project at the end of the day. It was an interesting learning curve but truth be told extremely easy to learn considering I am really not used to web development. The result is something we can be proud of. It was also interesting to see the wealth of open templates developed by other users. Maybe one day we can also contribute one. The most interesting aspect is probably learning that Github actually provides a way to host websites totally free of charge and through relatively simple or at least well documented steps.
Data visualizations :
The data visualizations were what I initially thought would be most interesting. Applying what we learn in programming to an actual concrete goal. In all reality it was a pretty straightforward process that didn’t require too much thought. It was however interesting to think about the best ways to visualize what we wanted to with Sofia, leading us to the results you see on my data retrieval or Sofia’s interactive visualization. It also led us to discover other bugs within the API with the data retrieval which I suspect really helped the developers because it would have taken a really long time before anyone thought about testing those features if it wasn’t for the concrete use case we had.
Recursive development :
Everything that we developed we also made openly available through a Github repository that was itself on the sdg-market. We hoped it would lead us to study first hand the token generation process. We also identified an issue through that, which we were able to notify the team of. Essentially, tokens were only generated for the first few projects but not for any of the new ones added. We notified them of that and they are still working on it.
Open source Project Management / Maintaining :
A surprising outcome of the project was how I needed to employ project management skills which I didn’t count on at all but it was a welcome development. As the one with the most experience with open source tools I set up the Github collaboration environment we used for the development of the documentation website. I also retrieved all the data so that the others could work on generating the visualizations and designing the website. In setting up the collaboration environment I was adamant that there shouldn’t be any hierarchy and that we should all go through peer review before posting anything into the documentation which we did. In the end I ended up being the first to review any post to the website. Whenever someone was stuck I would suggest some documentation or online tutorial. Through doing all that I ended up being recognized as somewhat of the project manager or maintainer as it would be called in open source development jargon. I am really grateful for the opportunity this project granted me as it helped me develop my passion for such tasks and also grateful for the trust that my project partners granted me.
Looking at this project critically through an open science scope, the documentation for contributing is partly insufficient since having little to no basics in web development it is difficult to really contribute any code. A better way would have been to add documentation detailing how one can reproduce the build that is available online for testing purposes. I tried on my own to reproduce the results first in Scala and then in python but provided only with somewhat functioning code that is difficult at best and impossible for most. In a truly open science way of functioning documentation would be more complete on how one can really reproduce the project. That might be something we can work on.
Out of the different goals we set off with, we were able to address almost all except one potential contribution. We tested the platform extensively, generated visualizations and added documentation to the platform. We learned a lot through the process but we couldn’t implement a blockchain recording system.
6. Future Plans
- Of course it would have been extremely over ambitious to think we could do everything in the short span of three weeks with everything else we had going on. It can however be left as a potential further development and if we think within the open contribution spirit and the sdg-market spirit which remunerates recommendations. I remembered that I knew a guy who had implemented blockchain solutions before. Leading to one further development where I could learn about blockchain and contribute by getting this person involved demonstrating simultaneously the power and virality of such open mechanisms. A more closed system might have led me to think more selfishly about this development and made me have to learn on my own.
There are also a few other open issues we have to deal with which are viewable on the documentation repository. Some of which will have to wait until the prototype is more stable. In the meantime we decided to pause further development and meet with Thomas one more time to see what we can do. I also tried, on my own since I had some extra time, to program a distributed version of the market first in Scala then in python that would run as a git extension. Since it is only a side project I chose not to open it until I had the chance to talk to Thomas about it.
Finally, a few development questions I would have are :
- An sdg-market is much bigger than only open source software contributions. How does the project move forward by allowing different types of projects to be traded on the platform ? Hardware has a potential since Open hardware repositories exist. But how do more human projects get involved ? Git diff is a good way to measure all sorts of documents but the learning curve might be quite steep for non technical oriented people. A wrapper over git for none technicians might be an interesting way forward.
- Similarly a market based on Github could be a welcome aid to an open source market in dire need of financing opportunities. Is there a way for the platform to open up to all sorts of open source projects while still prioritizing sdg projects ?
- Are the mechanisms for remunerating sufficient ? c.f Bayar’s Report.
- Currently the source code for the project is openly accessible. That still doesn’t make the project open in the sense that free and open source entails. For that we would need to explicitly grant at least some rights through the use of a license. What might be most adapted ? Perhaps some kind of SSPL like MongoDB uses or less controversial an AGPL ?
- An additionally point to reach full openness is working on documentation so that full reproducibility is possible.
- How can I keep further involvement with such a project ?
- As we near the end of the project one important concern comes to mind. As we try to artificially incentivize involvement in such projects through some form of market. It is going to be important to make sure we don’t corrupt the spirit of projects that enjoyed voluntary participation until now. The question is thus how to build such an incentive tool in respect of the open culture that existed before it ?
- http://blog.developers.market/data%20visualisations/dynamicVisu https://github.com/developers-market/developers-market.github.io/issues