I wrote a web site to learn om. I volunteered to score a six hour running race, the NightOwl Shuffle so the site was to track the runners. The race has a long and a short loop, so each time a runner came through, I would enter the runner’s number and which loop as the runner went past. The site would add up the mileage and adjust the standings.
Figwheel is the coolest!!
I worked with the browser open in the background and an editor open in front. Save a file. The page updates. Change branches in git. Page updates. Spend a day waiting fifteen minutes for a j2ee instance to fire up after every change and that is the greatest thing in the world. For a lot of changes, I didn’t even have to swap over to the browser to see the effects of a code change. It just showed up in the browser in the background.
Back on topic
For my race scoring website, I had runners, courses (the long and short lap
courses), and then when the race starts, laps which are a combination of a
runner and a course. A many to many relationship.
Laps would be an intermediate table with foreign keys to runners and courses.
Of course within an Om application, there is no database. According to the Basic Tutorial:
In Om the application state is held in an atom, the one reference type built into ClojureScript
Within that atom, I had a keyword key of :runners, a keyword key of :courses and a keyword key of :laps Each of those keys has a value that is a vector of maps.
To create the runners page, I needed the list of runners. Each of the first pages built needed the list under it’s key.
If I wanted one particular runner I had to filter those results
This type of code is scattered all over the big ball of mud that became my race scoring site. What I’d like is to be able to treat the app-state atom like a database and query it in a familiar way.
That will be the topic for the next post