It's getting close to a year since I released my static site
generator Pile o' Text, and I think it's high time to
evaluate its strenghts and weaknesses:
Pros
* No painpoints in operation; it's still very easy to work
with.
* The text is very readable, especially for the short
bursts of text I wanted to encourage.
* Multiple format outputs haven't been a problem.
Cons
* It's a pain to deal with templates that are embedded in
the script itself.
* I use two different invocations in the typical workflow
of adding an article: the pot script to add the article
to index, and then a makefile to build and push the
resulting artifacts onto the server.
* It doesn't support incremental builds, everything is
rebuilt from scratch.
* The idea of having text -- no more, no less -- isn't a
bad idea _per se_, but it kind of defeats the purpose of
the web by layering this idea on top.
* My primitive markdown-esque markup language is _very_
limited, and I have no interest in improving it.
* It's not very Unix-y at all in that it is a tiny
monolith.
Quite a set of cons, innit? While I stand firm that the idea
of a minimal point-and-shoot solution for a pile of text
files is great, the integrated script that I hacked together
is terrible.
This is no surprise; it has always been terrible. But its
purpose has never been to be great, but to prove a point to
myself that static site generators doesn't need to be
horribly overengineered pieces of shit. It was a
sanity-check. And a reason to play in-depth with bash.
So what will I do now? I came across a post by Vladimir
Sedach where he explained his setup using make and a handful
of tools to manage his site, and I felt that this is the way
I want to evolve my own setup.
