“A Man, a Plan, a Canal – Panama”
The lovely palindrome suggests that the building of the Panama Canal was planned and executed in a rather neat and reversible way. So when I consider my development process which is kinda haphazard, I feel a bit inadequate and envious. Fortunately, a quick glance at the Wikipedia entry on the building of the Panama Canal suggests that its 33 year construction was not without drama, difficulty or incompetence. Ferdinand de Lesseps, the man with the plan, was sentenced to go to jail for his role, but never actually did.
So that makes me feel better. The first Extended Stats site started as a few Python scripts. I spent years running them nightly and uploading them to free hosting sites, which were universally shitty, resulting in my decision to host the site at home. Maybe only a year after that I rewrote the scripts as a proper web site. Between the shitty free hosting and the world’s mistrust of dynamic DNS, the site name kept changing. As web users continue to expect more (e.g. https) and I become more professional in my approach, it continues to evolve.
The same thing is happening with the new site. I spent a good deal of 2015 and 2016 thinking “I should make a new site”, then a good deal of 2017 thinking “hmm, what would be the best way?” And then I spent a few more months trying to get those ideas to work. It was only in June 2018 that the technology started to work for me. About then I realised that Node.js, which I had chosen to write the server-side in, was not one of my talents. So I wrote some fairly bad Node code for a while, and maybe I am still doing so. Though it’s a few months since I’ve found a way I like better.
Similar things happened with the web pages. Although I am competent with Angular, when I started I didn’t know what I wanted on the pages, and I still don’t, really. However they are slowly converging. Whenever I write a new page there are some bits I like and some bits that are horrible. When I’m done, I take the bits that I like and put them into a library on NPM, and use them again later. That makes it easier to re-use those ideas on later pages, or on revisions of existing pages.
So this morning I sat down, determined to put the Plays Of Games Owned table into the new site. The first task was to find the right place for it, which would be on an existing page which already has the correct data. I looked at the Owned Games page, and noticed that it was the first page where I had implemented the pattern of the page loading a set of data, and multiple components rendering their views on it. However some code from that page had been put into a library, so I fixed up the page to use the library and stuck in a loader component while I was there (the coloured bobbles that indicate that the data hasn’t arrived yet).
However the Owned Games page doesn’t have plays data, so it’s not the right page for that table. So I next considered the Favourites table, which does include plays. The selector isn’t quite right though, as the Favourites table defaults to games played and rated, and isn’t anything to do with owning them.
The Favourites page one of the first pages to use Vega charts, so it had a Charts button that you could click on to get to the charts. That preceded the way I do it now where each chart appears on the page as its own thing, and uses the data the page provides. So I moved those charts out onto the main Favourites page, and wired them up to receive the data. Then I noticed that the Favourites page was the one where you can change the selector, but the new selector was injected into the table, rather than into the page. So I went back to the library and changed the way the config component worked, so now you can change the selector on the Favourites page, and the table and all of the charts update. It’s still clunky, but it works. Oh, and the Favourites page had one of the old-style tables without the handy header bar, so I fixed that up as well.
And that’s why it’s now late afternoon and I have not yet achieved the first thing I set out to do today. The site has evolved a little, and I look forward to wiring the selector changer into other pages, and hopefully making it nicer when I do so. I’ll see if I get to try again tomorrow.
“A man, no plan, a clusterfuck, a clean up.”
