Post Mortem - Open Patent Data
Author: Tim Jagodzinski
Six months in hindsight
Open Patent Data is the effort to introduce patents to the mainstream. To move dull boring technical descriptions buried in boring and often confusing official websites into a 21st century web application. When I was introduced to the project I was all excited (and still am) because I believe strongly that if patents would be read by other people than professionals, we wouldn't have all those silly general all-purpose patents polluting our world.
We went through half a year of excitement, blazing progress and crushing setbacks. Through great teamwork and terrible miscoordination. Basically I had the best six months in my career and I want to share this experience, so that you can draw whatever wisdom you like from it.
Kickstarting - like a kick in the groin
Our objective for this project was to build a state-of-the-art web application. It was supposed to be beautiful, efficient and accessible. We wanted to blow all the other "professional" apps out of the water and show the world how a 21st century patent information app is supposed to look like. We wanted to embrace agile development and use the shortest possible iterations to give us the freedom to just try out things or throw away stuff that wouldn't work.
There was only one little problem. None of our team had any notable experience with building web apps.
So if you don't have the experience you get people how do and let the teach you. That's what we did. And we got the talented people of Edenspiekermann(ESPI). We spent the first three weeks of the project with these guys had a great time, produced a complete project bootstrap and learned a lot about our agile workflow which helped us a lot in the long run. With so much good things, there must be bad things to balance out, ain't it? There was :)
We ended up with a great base for our project. We had a working backbone.js / Django tag team. And we didn't know Jack about how the freaking thing worked, let alone all of the little things of backbone.js and CoffeeScript that would blow into our faces every now and then in the upcoming months. Code documentation was sparse at best and virtually incomprehensible for us. It took us 2-3 months to get comfortable with the front end code base. Even today there are still dark corners which we don't fully understand.
Who's to blame? Can't really say. I think the problem was twofold (as so often). We didn't clearly enough convey our demand. Getting taught. And the ESPI guys didn't seem to have much experience in teaching. They were very focused on getting stuff done. An attitude that we adopted for the project (and learned from these guys). So I guess there was still enough benefit for us. However my lesson learned is that if I ever get into a similar situation, I will insist on the "getting taught" bit and not so much on productivity.
Struggling through the Jungle
The mantra we kept hearing from our boss was "Mobile First". We liked that idea, because it forced you to reduce functionality to the absolutely necessary, which leads to a lean and efficient product and a good user experience. We designed everything up from the simple.
When forging our designs in code however we ran into more and more problems which rooted, as we realised half through the project, in the base we got at ESPI. The arrangement of DOM elements got increasingly harder to manage with Backbone views and we started wondering if we did something wrong. A lot of the CSS rules didn't make sense to us either. So what happened?
At the beginning of the project we were kind of overwhelmed with all the new stuff thrown at us at ESPI. So we took a lot of stuff for granted just to keep up the pace. We didn't knew better and I think we didn't clearly enough stated the direction of the app. So the ESPI guys drew their own conclusions - which were perfectly fine at that point in the project - and we ended up with an app structure which wouldn't fit to the way we wanted the app to behave. To fix it we spent a week and basically restructured / redesigned all templates within a week. A good experience since we learned a lot about project structure and arranging stuff with handlebars and backbone.js.
Against all odds
From half way through the project we really took up speed and got good traction. We constantly increased our productivity and quality - despite all the little things we had to struggle against outside the box of our office. A loved one got terribly sick for almost three months, half of the core team only worked half time on the project, half of the Sprint planning / review meetings had to be conducted remotely via Google Hangouts or Skype - just to name a few.
The outsider might think we didn't achieve much in the last six months. But I think we did a damn good job. I think other teams - under the same circumstances - would just have plainly failed. So that's something we can be truly proud of. Working with all my talented and passionate colleagues was just a blast and I love how Open Patent Data turned out. And I am eagerly looking forward to what there's still to come for the future of patent information.
I hope you are too :)












