Postmortem: Bundesfeeds — an attempt to enter the “prototype fund”

So somewhere in early 2017 I got aware of “prototype fund”, a 50K government based money for an open source idea with the following focus

“The second round of the Prototype Fund calls for projects that show how code and software can strengthen civil society.

These could be tools that enable activists and organizations to mobilize their supporters, organize campaigns, collect information, and share knowledge. Or software that helps people to get new initiatives going, and to organize groups in an easier or more efficient way. And of course, this also includes tools that protect our privacy and enable secure communication. Some examples of these tools can be found in our blog (in German).

Please note: We will continue to support ideas in civic tech, data literacy, data security, and software infrastructure.”

My Intention

Write a tool, that makes the real communication (via blogs) of politicians transparent and comparable, enumerable and this data accessible to users, journalists and the public. The discussion around “fake news” was peaking, while simple, but none the less frightening facts got underreported. It is really interesting to see, how much information really is out there.

I just wanted to have a simple view into what was going on in the “Deutscher Bundestag”. No newspaper or any other tool was giving me a direct view on this set of contents. OpEds, you can get them any day. Even the daily list of contents is overwhelming. There is sure need for something to sort and categorize the content.

So thing is: I can exchange 6–9 Months of time for some bucks and minimize all other expenses as much, so that I can care for a project like this. So why not invest some nights and code it out, apply and just build something.

The topic of the prototype funds was “public infrastructure” and my thinking was: Something that

TL;TR Intentions

• Build something that makes content of politicians viewable.
• Learn ruby/more ruby
• Learn more about grant processes etc.
• Get feedback about the actual idea from pros — aka the fund
• Publish code while doing all of the above
• Get all the RSS feeds
• Optional: Get the money to really build it

Bundesfeeds — The beginning.

So I am collecting the RSS feeds of all “Bundestag” anyway. This is kind of a little obsession. A) I like RSS/ATOM as a format B) I like real politics, as it keeps the tinfoil from flying. So collecting these rss urls is a thing I do for a time. Sadly for my girlfriend, sorting and reading these was part of my holidays in Asia. It was cumbersome: Feed readers, different formats etc. Looked like something could have been built. Especially since everything that reads about 1000 rss feeds (which is my estimate for Bundestag, Bundesrat and other important feeds from official side) costs money. Just try managing 100 triggers on IFTTT and you know what I am pointing out to.

Problems Reading those feeds

• Usability of RSS sucks a bit when its 1000 feeds and RSS feeds are literally HIDDEN from users.
• Some of the best known politicians don’t even bother. Angela Merkels strategy seems to be to “shut the fuck up” anyway. So no wonder, there was no RSS feed
• Very much different platforms out there. Not so sure, why a member of the deutsche Bundestag should use a outdated open source blogsystem.
• Content wise it was funny to see what is going on and that politics is juts more than what we perceive it. Sorting required.
• Data needs normalizing
• Morning, High Noon, Evening, Yesterday (00:01) and The Night Formats are what I want.
• Fuck me, I should have outsourced the collection of the rss feeds.
• Way less anger when reading opposing “sides” with opposing opinions. Maybe it is not good at all to use a Oped for generation of own opinion. Compared to the sole content generated by politicians, the “politics” part of “der Spiegel” for example is small. So we only get reporting on a fraction of contents, added with OpEd to form opinion.
• Parallelism

The first attempt at this was “bundesfeeds”, basically a feed parser that created a new feed, much like a “planet” (see http://planet-php.net/ for an example)

Goals

I tried to recap a bit if I met my own initial criteria for success.

Build something that makes content of politicians viewable

The whole project was a simple and localhost running rails application. It exported a bunch of pages to github pages and et voila. There is a result. In the end I was starting to use Morning and daily Lists, which suited my way of working. Goal80% reached. Could have been a bunch more features, but it did the trick.

Learn ruby/more ruby

I learned a lot out of the ruby and rails world. It was good to really do something. Although, I would do things differently. Probably the experience with IRL projects kicks in now. However, this was the best outcome of the whole thing.

Learn more about grant processes etc.

Pew, I am not even sure I have a email stating there is no grant until now. My expectation was, at least in the beginning, a bit more interaction about the idea and a general feedback, other than the form stating it has received the application. My last contact to the team was a Mail when I asked to re-enter to improve the application it self. My learnings:

a) ask the following things.

1. Is the process completely visible and documented?
2. Are all board members and decision makers publicly stated?
3. Are the criteria of success, as far as they can be set clearly, stated in quality or quantity goals?
4. Is the Organisation adhering to the set up process (aka is feedback given when offered etc)
5. Past projects, what kind of success have they been in terms of show me the code, is it usable.
6. Did people not succeeding (by far the greater number btw) get anything out of it?

TL;TR: Learnings: These “funds” combine more ideas I dearly hate that I realized before: It works like a call of bids, from a government Organisation, but with less oversight and regulations. I have my fair share of experience in my work life, so it a bit funny I was not aware of this. These, Compared to professional projects, these are normally more underfunded than official ones. Then there is the competition and rivalry aspect of the whole “rounds” and winners/losers concept. I could have known this, I did not. Now I know ;)

Get feedback about the actual idea from pros — aka the fund

Since there was no feedback, this was a fail. So the most positive feedback could have been 50K Euro for a project and a bunch of good input and contacts. Given I have Invested 100 hours of my time easily in data, research, building the thing, testing and monitoring and that I could work in the meantime, earning money: A first portion of the 50K was already “spent”. I myself upfront invested it, but did not notice.

Unsuccessful projects are cool for the funds marketing department: 1000 applications …. sounds nice — “You heard, XY has submitted to the fund”. Its just some grants are given and the big rest gets nothing. But has invested the time. So much about value of the feedback (none) compared to the effort (100 hours time hourly rate). Lets put it like this: I had way cheaper and more effective (in terms of there is feedback that I can use) ways of getting a hint if a idea is good or not. ;)

Other feedback was good. Very constructive from users.

I had about 30 pieces of feedback from different users I pinged about the project.

• Needed usability
• Hard to find end Users from “data science”
• The Open Data scene took the data and integrated it in other sites.
• Should have been set up more “seriously”. Although managing a extra server was not m intention
• I liked the download numbers of the bf_multi_rss ruby gem
• Nice people looked at my code and told me it needed fixing — it needed fixing ;)
• The twitter feed is a troll trigger.
• facebook is a dead end. But I don’t want them comments anyway.
• Twitter made me too angry at times. Idiots. next time have someone manage that kind of feedback.

I did not have to fill an application for the user feedback, a about page was enough. It was plenty and more that I could process, but very valid. User beats funds anytime.

Publish code while doing all of the above

I published:

• bundesfeeds rails application
• bf_multi_rss
• a node module containing the feed data

2 of them get unpublished soon. The rails application is incomplete and I don’t want to maintain it. And the node module has no process behind it and lacks the data. None of them is beyond “experimental” and can be unpublished. It goes to backup and does not create more work there.

The bf_multi_rs gets a bit of yak shaving and a kind of a update when I finish my next ruby project. I really liked concurrency in ruby. Since we will most likely see more cores in the future, this is a good thing to maintain.

I published to much code, the rails app could be kept private. The node publish was related to an old project. Not creating the code is the problem, maintaining it is. So this stuff will be cleaned or removed. ;)

Collect all the feeds

Not all, but many. Would be nice to find a collective, where we could find this info as a team. 60% complete and election period. I hope someone sits down post election and collect the data.

My Take?

I have observed, my own expectations did not meet what a organization like “prototype fund” can provide. So next time I want do something good with a pice of software, I just build it and do not invest time in a “funds” promising feedback or money and deliver none of both. This just makes me feel left alone and the only one “investing” in a Project like this, while I want to do a good thing that I simply can do on my own with less drag. Aka for no feedback I wont change the focus of a idea to fit a topic. Maybe a bit smaller would be nice.

The list of projects in round one looks very promising and we see whats coming out. They assembled a list of nice projects. This post mortem has nothing to do with any participant and I wish the best for everyone who applied and everyone who is still grinding on their idea and best of luck to all who made it to round two.