Image credit: Ratan Abraham VargheseWhen this blog began, it was pretty much for my own amusement. I picked technologies specifically because they seemed fun to play with. Features were limited based on how easy I thought they would be to implement myself. The layout was optimized not for reading experience, but for showing off SVG resizing and CSS transparency. I started writing prose, but switched to listicles and even bullet points simply because it was easy.
Was this all good? Bad? I had no idea. My own opinion reversed itself every few days. As for other people, their feedback was limited to direct messages and a log file that nobody read.
All the while, the wider internet was changing. So over the past few weeks, I decided to enhance this site on several fronts.
The first change was an overhaul of the content management system. The incumbent was a system I wrote myself in Golang:
blom. Golang is a fun
little language for certain tasks, but its unique strengths did
not shine through in blom.
I pretty much rewrote blom such that each task was conducted
using the most suitable language: file management was handled in bash
scripts, templating was handled using
m4 and various other tools
were used as appropriate. This is obviously not something a reader
would notice directly, but it made adding new features and testing new layouts significantly easier.
Another enabling change was switching from a simple NodeJS-based server to nginx. In retrospect this was not necessary, but it made some of the next changes easier to configure.
In the past, every time a file on this site was accessed, a line would be added to a log file which I never, ever checked. However, now I’ve installed a web log analyzer which presents all the data with pretty colors. I used to feel like I was shouting into the void, but now when I write I can be assured that plenty of search engine bots are reading.
Even before this blog existed, the world has been moving to using
HTTPS by default for all websites. ratan.blog has been very late to the party … still, better late than never!
I added a comment system! That merits its own article.
Here is what the blog used to look like.
I actually attempted disable all the browser’s default styling and individually write new styles for every possible HTML element. This was both exhausting and probably a waste of time, but reading the MDN list of HTML elements was very educational. For one thing, I finally know what a description list is (also called a definition list). As for the CSS reset, I ended up just using the budget version:
* {
margin:0;
padding:0;
}
After eliminating the Lufthansa-esque color scheme, the next casualty was the logo.
Image credit: Ratan Abraham Varghese
What was the point of that logo? Was it totally arbitrary? Well, I
don’t actually remember. All I know is, the filename was
calculated_asymmetry.svg, but the logo was symmetrical using an axis
45° from the horizontal. What a scandal!
Image credit: Ratan Abraham Varghese
The meaning of the new logo is that right-angled trapezoids are underrated.
So you know that feeling when some task seems so easy that there would be no reason to test it, but then something goes wrong that could have been easily detected while testing? Yeah, that’s what happened after I registered r3n.me. I tried to make a script that created alternate URLs for every existing article. Unfortunately it created alternate URLs for those alternate URLs every time it was run. Eventually there were 3 or 4 paths to every article and each path had a corresponding entry in the RSS, Atom and JSON Feeds.
If you were subscribed to any of those feeds and saw “re-runs” of old articles, sorry about that.
On the plus side, there are short URLs for every article. They are recorded in the archive.
The third item will shock you!