It's been awhile since my last check-in, but my thoughts never stray far from the work, nor from this group. At the macro level, my life is an almost-balanced exchange between uncommon freedom -- which has enabled me to put longer stretches of time into my learning here -- and an overwhelming stress, which sometimes keeps me away. One hopes that this balances out over a long period of time, though at times like now it can feel like it definitely doesn't.
I'm interested in any thoughts about how to fix this. I'm considering any option, because it sucks to get pulled away from what I really want to be doing in order to handle repetitive technological tasks in support of work that I understand to lack meaning.
But, tonight (Friday night, when I started putting this together) I have a window of a few hours. I sat down to begin to work, and decided that perhaps the best way to start is to catch up with you all, and put some time into documenting what I've learned recently in the time that I have had to work.
Last weekend Austin generously provided the bulk of his Saturday to walk through his React/Flux code with me, help me understand how and why it's supposed to work, and then to help me pull my own implementation for 'twinput' into form.
My immediate development goal was to explore visualization designs, as per my checkin a few weeks ago regarding the book on data displays, and my process -- mocking up visualizations by building them in React -- had been working really well. But I was beginning to find it difficult to understand the whole of my project given that the data relationships it was structured around had become so convoluted, and it was making it more difficult for me to think about each small piece I needed to be thinking about, with confidence in the layers of abstraction adjacent. Building my visualizations was not as quick an iterative process as I wanted it to be, because I was getting tripped up on the underlying environment, which I felt I had left quite messy.
Implementing Flux seemed like it would be a quick and efficient way to put things in order, so that I could move forward more quickly. But, even though it has a really simple structure and a basic set of design principles that guide it, I struggled for awhile to grok it, and instead of it being a quick design improvement to enable faster iteration of my design, it became a morass, and I got stuck.
Guillaume asked me, at this point, to pick a route: continue to work on learning to code, or work on learning to design visualizations. But, I hadn't intended to leave either behind. The process of iterating the designs felt to me like a veritable font of coding practice exercises, with quick and clear goals. So because I wanted to keep iterating the designs in this way, and faster than I had been able to without Flux, I needed to get my code the rest of the way in shape -- I didn't want to retreat. Thankfully, we have an Austin!
Our 8-or-so hour session was incredibly successful. I now have a solid understanding of how Flux works and have a working implementation of Marty. Until today, it provided only the elements in the story -- the tweets. Today I began implementing a 'distributions' store, which will keep track of the data points that we want to visualize.
Flux is already bearing fruit, I think, in that its implementation helped me think about naming this new store more precisely and more usefully. It forced me to think about what scope data lives on, and in so doing I was reminded of the user story that is driving the timeline/visualization work I'm doing now: revealing the connections between and within stories. (I had been focused on the 'within stories' part, but the 'between stories' part is perhaps even more important. Having distributions as a store that is parallel to the one holding the Stories makes much more sense than storing them within stories. The structural separation of concerns imposed by Flux has led me to a stronger understanding of how the data wants to relate to itself.)
Though I took some notes during our session, Austin as ever went way above and beyond and polished his own up til they got really shiny because he's great. Thanks for all of your help, man.
I have also been plugging away at this graphical display design book, which is very cool and fun to read, though also very dense. I've found that the lessons are substantial -- the stickiness of the information seems to relate closely to the density of my notes -- and they also, for me, have been taking awhile to sink in -- I have to mull over them and parse them a bit. Just today I was finally able to wield some of the things I've learned, which to me means start to feel like I have an intuitive ability to make use of them.
The result is This gist which contains my musings about Twinput and design lessons from the book.
I plan to continue working my way through the book. I'd be particularly interested in hearing feedback about that last document, though, and I'd like to actually sketch some designs out. I think it's been good in some ways, though not what I had hoped would happen, to have been forced to step back here a bit briefly, because it may have helped me see how what I was doing before didn't make a ton of sense.
Currently exploring options but yeah this is the plan right now, at least for now...
OK! <3 y'all!