Start Date: April 22
End Date: April 25
Blog Post
In any exercise, it's always the last few reps that command the bulk of one's willpower, and this month's challenge has been no exception. By now, I've run into enough roadblocks and setbacks that the questions I started out with have been largely answered (more on that at the start of May), and the only thing to do now is finish off the month - ideally without another failure.
I was supposed to have fun with this project. The mock-up for it had been designed since the first week of April, and it was one of the better designs this month. But while this project isn't a failure by the rules of the challenge, it certainly doesn't look like it should. Most of this is just down to me running out of time - I didn't even get the chance to prepare any imagery for it.
I decided to try the Stylus CSS preprocessor as my main technology this time. By no means did I use it well, or even elegantly. However, it might actually be my favorite CSS language out of those I've tried this month, beating out Sass and Less. I used it through ParcelJS, which once again was used as my build tool. Unfortunately, my choice to use Parcel also cost me a bit of time, since (at the time of writing) its HMR has a tendency to crash every 10-20 updates. I also had some trouble using Stylus's import feature consistently, though I'm not entirely sure why. At a certain point, I went ahead and just wrote everything in the top-level file.
I have to admit: I'm getting a little tired at this point. But I have one more project to go.
Remarks
You can tell I was tired by this point, since the quality of this page is definitely a regression from the last few - although unlike the last project, I was at least able to deploy it.
This was the last time I used Parcel.js during the challenge, and I spent most of this time feeling like I probably shouldn't have. As was mentioned above, it was crashing pretty regularly, and I wasn't gaining much by putting up with it. I used it because I wanted to use Stylus, and Parcel promised to facilitate that with no configuration. But as I've since learned, the cost of that abstraction is that when something goes wrong, it's never immediately clear why, or who's to blame.
I did quite like Stylus itself, though. The next time I use it, I'll take care not to use any shortcuts.